Laboratory tests are often ordered inappropriately for patients in whom a rheumatologic illness is suspected; this occurs in both primary and secondary care.¹ Some tests are available both singly and as part of a battery of tests screening healthy people without symptoms.

The problem: negative test results are by no means always reassuring, and false-positive results raise the risks of unnecessary anxiety for patients and clinicians, needless referrals, and potential morbidity due to further unnecessary testing and exposure to wrong treatments.² Clinicians should be aware of the pitfalls of these tests in order to choose them wisely and interpret the results correctly.

This article provides practical guidance on requesting and interpreting some common tests in rheumatology, with the aid of case vignettes.

RHEUMATOID FACTOR AND ANTICITRULLINATED PEPTIDE ANTIBODY

A 41-year-old woman, previously in good health, presents to her primary care practitioner with a 6-week history of pain and swelling in her hands and early morning stiffness lasting about 2 hours. She denies having any extraarticular symptoms. Physical examination reveals synovitis across her right metacarpophalangeal joints, proximal interphalangeal joint of the left middle finger, and left wrist. The primary care physician is concerned that her symptoms might be due to rheumatoid arthritis.

Would testing for rheumatoid factor and anticitrullinated peptide antibody be useful in this patient?

Rheumatoid factor is an antibody (immunoglobulin M, IgG, or IgA) targeted against the Fc fragment of IgG.³ It was so named because it was originally detected in patients with rheumatoid arthritis, but it is neither sensitive nor specific for this condition. A meta-analysis of more than 5,000 patients with rheumatoid arthritis reported that rheumatoid factor testing had a sensitivity of 69% and specificity of 85%.⁴

Anticitrullinated peptide antibody, on the other hand, is much more specific for rheumatoid arthritis (95%), as it is seldom seen in other conditions, but its sensitivity is similar to that of rheumatoid factor (68%).^4–6 A positive result would thus lend strength to the diagnosis of rheumatoid arthritis, but a negative result would not exclude it.

Approach to early arthritis

When faced with a patient with early arthritis, some key questions to ask include^7,8:

Is this an inflammatory or a mechanical problem? Inflammatory arthritis is suggested by joint swelling that is not due to trauma or bony hypertrophy, early morning stiffness lasting longer than 30 minutes, and elevated inflammatory markers (erythrocyte sedimentation rate or C-reactive protein). Involvement of the small joints of the hands and feet may be suggested by pain on compression of the metacarpophalangeal and metatarsophalangeal joints, respectively.

Is there a definite identifiable underlying cause for the inflammatory arthritis? The pattern of development of joint symptoms or the presence of extraarticular symptoms may suggest an underlying problem such as gout, psoriatic arthritis, systemic lupus erythematosus, or sarcoidosis.

If the arthritis is undifferentiated (ie, there is no definite identifiable cause), is it likely to remit or persist? This is perhaps the most important question to ask in order to prognosticate. Patients with risk factors for persistent disease, ie, for development of rheumatoid arthritis, should be referred to a rheumatologist early for timely institution of disease-modifying antirheumatic drug therapy.⁹ Multiple studies have shown that patients in whom this therapy is started early have much better clinical, functional, and radiologic outcomes than those in whom it is delayed.^10–12

The revised American College of Rheumatology and European League Against Rheumatism criteria¹³ include the following factors as predictors of persistence:

• Number of involved joints (with greater weight given to involvement of small joints)
• Duration of symptoms 6 weeks or longer
• Elevated acute-phase response (erythrocyte sedimentation rate or C-reactive protein level)
• A positive serologic test (either rheumatoid factor or anticitrullinated peptide antibody).

If both rheumatoid factor and anticitrullinated peptide antibody are positive in a patient with early undifferentiated arthritis, the risk of progression to rheumatoid arthritis is almost 100%, thus underscoring the importance of testing for these antibodies.^5,6 Referral to a rheumatologist should, however, not be delayed in patients with negative test results (more than one-third of patients with rheumatoid arthritis may be negative for both), and should be considered in those with inflammatory joint symptoms persisting longer than 6 weeks, especially with involvement of the small joints (sparing the distal interphalangeals) and elevated acute-phase response.

Rheumatoid factor in healthy people without symptoms

In some countries, testing for rheumatoid factor is offered as part of a battery of screening tests in healthy people who have no symptoms, a practice that should be strongly discouraged.

Multiple studies, both prospective and retrospective, have demonstrated that both rheumatoid factor and anticitrullinated peptide antibody may be present several years before the clinical diagnosis of rheumatoid arthritis.^6,14–16 But the risk of developing rheumatoid arthritis for asymptomatic individuals who are rheumatoid factor-positive depends on the rheumatoid factor titer, positive family history of rheumatoid arthritis in first-degree relatives, and copresence of anticitrullinated peptide antibody. The absolute risk, nevertheless, is still very small. In some, there might be an alternative explanation such as undiagnosed Sjögren syndrome or hepatitis C.

In any event, no strategy is currently available that is proven to prevent the development of rheumatoid arthritis, and there is no role for disease-modifying therapy during the preclinical phase.¹⁶

Back to our patient

Blood testing in our patient reveals normal complete blood cell counts, aminotransferase levels, and serum creatinine concentration; findings on urinalysis are normal. Her erythrocyte sedimentation rate is 56 mm/hour (reference range 0–15), and her C-reactive protein level is 26 mg/dL (normal < 3). Testing is negative for rheumatoid factor and anticitrullinated peptide antibody.

Although her rheumatoid factor and anticitrullinated peptide antibody tests are negative, she is referred to a rheumatologist because she has predictors of persistent disease, ie, symptom duration of 6 weeks, involvement of the small joints of the hands, and elevated erythrocyte sedimentation rate and C-reactive protein. The rheumatologist checks her parvovirus serology, which is negative.

The patient is given parenteral depot corticosteroid therapy, to which she responds briefly. Because her symptoms persist and continue to worsen, methotrexate treatment is started after an additional 6 weeks.

 

 

ANTINUCLEAR ANTIBODY

A 37-year-old woman presents to her primary care physician with the complaint of tiredness. She has a family history of systemic lupus erythematosus in her sister and maternal aunt. She is understandably worried about lupus because of the family history and is asking to be tested for it.

Would testing for antinuclear antibody be reasonable?

Antinuclear antibody is not a single antibody but rather a family of autoantibodies that are directed against nuclear constituents such as single- or double-stranded deoxyribonucleic acid (dsDNA), histones, centromeres, proteins complexed with ribonucleic acid (RNA), and enzymes such as topoisomerase.^17,18

Protein antigens complexed with RNA and some enzymes in the nucleus are also known as extractable nuclear antigens (ENAs). They include Ro, La, Sm, Jo-1, RNP, and ScL-70 and are named after the patient in whom they were first discovered (Robert, Lavine, Smith, and John), the antigen that is targeted (ribonucleoprotein or RNP), and the disease with which they are associated (anti-ScL-70 or antitopoisomerase in diffuse cutaneous scleroderma).

Antinuclear antibody testing is commonly requested to exclude connective tissue diseases such as lupus, but the clinician needs to be aware of the following points:

Antinuclear antibody may be encountered in conditions other than lupus

These include¹⁹:

• Other autoimmune diseases such as rheumatoid arthritis, primary Sjögren syndrome, systemic sclerosis, autoimmune thyroid disease, and myasthenia gravis
• Infection with organisms that share the epitope with self-antigens (molecular mimicry)
• Cancers
• Drugs such as hydralazine, procainamide, and minocycline.

Antinuclear antibody might also be produced by the healthy immune system from time to time to clear the nuclear debris that is extruded from aging cells.

A study in healthy individuals²⁰ reported a prevalence of positive antinuclear antibody of 32% at a titer of 1/40, 15% at a titer of 1/80, 7% at a titer of 1/160, and 3% at a titer of 1/320. Importantly, a positive result was more common among family members of patients with autoimmune connective tissue diseases.²¹ Hence, a positive antinuclear antibody result does not always mean lupus.

Antinuclear antibody testing is highly sensitive for lupus

With current laboratory methods, antinuclear antibody testing has a sensitivity close to 100%. Hence, a negative result virtually rules out lupus.

Two methods are commonly used to test for antinuclear antibody: indirect immunofluorescence and enzyme-linked immunosorbent assay (ELISA).²² While human epithelial (Hep2) cells are used as the source of antigen in immunofluorescence, purified nuclear antigens coated on multiple-well plates are used in ELISA.

Although ELISA is simpler to perform, immunofluorescence has a slightly better sensitivity (because the Hep2 cells express a wide range of antigens) and is still considered the gold standard. As expected, the higher sensitivity occurs at the cost of reduced specificity (about 60%), so antinuclear antibody will also be detected in all the other conditions listed above.²³

To improve the specificity of antinuclear antibody testing, laboratories report titers (the highest dilution of the test serum that tested positive); a cutoff of greater than 1/80 is generally considered significant.

Do not order antinuclear antibody testing indiscriminately

To sum up, the antinuclear antibody test should be requested only in patients with involvement of multiple organ systems. Although a negative result would make it extremely unlikely that the clinical presentation is due to lupus, a positive result is insufficient on its own to make a diagnosis of lupus.

Diagnosing lupus is straightforward when patients present with a specific manifestation such as inflammatory arthritis, photosensitive skin rash, hemolytic anemia, thrombocytopenia, or nephritis, or with specific antibodies such as those against dsDNA or Sm. Patients who present with nonspecific symptoms such as arthralgia or tiredness with a positive antinuclear antibody and negative anti-dsDNA and anti-Sm may present difficulties even for the specialist.^25–27

Back to our patient

Our patient denies arthralgia. She has no extraarticular symptoms such as skin rashes, oral ulcers, sicca symptoms, muscle weakness, Raynaud phenomenon, pleuritic chest pain, or breathlessness. Findings on physical examination and urinalysis are unremarkable.

Her primary care physician decides to check her complete blood cell count, erythrocyte sedimentation rate, and thyroid-stimulating hormone level. Although she is reassured that her tiredness is not due to lupus, she insists on getting an antinuclear antibody test.

Her complete blood cell counts are normal. Her erythrocyte sedimentation rate is 6 mm/hour. However, her thyroid-stimulating hormone level is elevated, and subsequent testing shows low free thyroxine and positive thyroid peroxidase antibodies. The antinuclear antibody is positive in a titer of 1/80 and negative for anti-dsDNA and anti-ENA.

We explain to her that the positive antinuclear antibody is most likely related to her autoimmune thyroid disease. She is referred to an endocrinologist.

 

 

ANTIPHOSPHOLIPID ANTIBODIES

A 24-year-old woman presents to the emergency department with acute unprovoked deep vein thrombosis in her right leg, confirmed by ultrasonography. She has no history of previous thrombosis, and the relevant family history is unremarkable. She has never been pregnant. Her platelet count is 84 × 10⁹/L (reference range 150–400), and her baseline activated partial thromboplastin time is prolonged at 62 seconds (reference range 23.0–32.4). The rest of her blood counts and her prothrombin time, liver enzyme levels, and serum creatinine level are normal.

Should this patient be tested for antiphospholipid antibodies?

Antiphospholipid antibodies are important because of their association with thrombotic risk (both venous and arterial) and pregnancy morbidity. The name is a misnomer, as these antibodies are targeted against some proteins that are bound to phospholipids and not only to the phospholipids themselves.

According to the modified Sapporo criteria for the classification of antiphospholipid syndrome,²⁸ antiphospholipid antibodies should remain persistently positive on at least 2 separate occasions at least 12 weeks apart for the result to be considered significant because some infections and drugs may be associated with the transient presence of antiphospholipid antibodies.

Screening for antiphospholipid antibodies should include testing for IgM and IgG anticardiolipin antibodies, lupus anticoagulant, and IgM and IgG beta-2 glycoprotein I antibodies.^29,30

Anticardiolipin antibodies

Anticardiolipin (aCL) antibodies may be targeted either against beta-2 glycoprotein I (beta-2GPI) that is bound to cardiolipin (a phospholipid) or against cardiolipin alone; the former is more specific. Antibodies directed against cardiolipin alone are usually transient and are associated with infections and drugs. The result is considered significant only when anticardiolipin antibodies are present in a medium to high titer (> 40 IgG phospholipid units or IgM phospholipid units, or > 99th percentile).

Lupus anticoagulant

The antibody with “lupus anticoagulant activity” is targeted against prothrombin plus phospholipid or beta-2GPI plus phospholipid. The test for it is a functional assay involving 3 steps:

Demonstrating the prolongation of a phospholipid-dependent coagulation assay like the activated partial thromboplastin time (aPTT). (This may explain the prolongation of aPTT in the patient described in the vignette.) Although the presence of lupus anticoagulant is associated with thrombosis, it is called an “anticoagulant” because of this in vitro prolongation of phospholipid-dependent coagulation assays.

Mixing study. The phospholipid-dependent coagulation assay could be prolonged because of either the deficiency of a coagulation factor or the presence of the antiphospholipid antibodies. This can be differentiated by mixing the patient’s plasma with normal plasma (which will have all the clotting factors) in a 1:1 ratio. If the coagulation assay remains prolonged after the addition of normal plasma, clotting factor deficiency can be excluded.

Addition of a phospholipid. If the prolongation of the coagulation assay is due to the presence of an antiphospholipid antibody, addition of extra phospholipid will correct this.

Beta-2 glycoprotein I antibody (anti-beta-2GPI)

The beta-2GPI that is not bound to the cardiolipin can be detected by separately testing for beta-2GPI (the anticardiolipin test only detects the beta-2GPI that is bound to the cardiolipin). The result is considered significant if beta-2GPI is present in a medium to high titer (> 99th percentile).

Studies have shown that antiphospholipid antibodies may be present in 1% to 5% of apparently healthy people in the general population.³¹ These are usually low-titer anticardiolipin or anti-beta-GPI IgM antibodies that are not associated with thrombosis or adverse pregnancy outcomes. Hence, the term antiphospholipid syndrome should be reserved for those who have had at least 1 episode of thrombosis or pregnancy morbidity and persistent antiphospholipid antibodies, and not those who have asymptomatic or transient antiphospholipid antibodies.

Triple positivity (positive anticardiolipin, lupus anticoagulant, and anti-beta-2GPI) seems to be associated with the highest risk of thrombosis, with a 10-year cumulative incidence of 37.1% (95% confidence interval [CI] 19.9–54.3) for a first thrombotic event,³² and 44.2% (95% CI 38.6–49.8) for recurrent thrombosis.³³

The association with thrombosis is stronger for lupus anticoagulant than with the other 2 antibodies, with different studies³⁴ finding an odds ratio ranging from 5 to 16. A positive lupus anticoagulant test with or without a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a high-risk profile, while a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a moderate-risk profile. A low titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a low-risk profile that may not be associated with thrombosis.³⁵

Antiphospholipid syndrome is important to recognize because of the need for long-term anticoagulation to prevent recurrence.³⁶ It may be primary, when it occurs on its own, or secondary, when it occurs in association with another autoimmune disease such as lupus.

Venous events in antiphospholipid syndrome most commonly manifest as lower-limb deep vein thrombosis or pulmonary embolism, while arterial events most commonly manifest as stroke or transient ischemic attack.³⁷ Obstetric manifestations may include not only miscarriage and stillbirth, but also preterm delivery, intrauterine growth retardation, and preeclampsia, all occurring due to placental insufficiency.

The frequency of antiphospholipid antibodies has been estimated as 13.5% in patients with stroke, 11% with myocardial infarction, 9.5% with deep vein thrombosis, and 6% for those with pregnancy morbidity.³⁸

Some noncriteria manifestations have also been recognized in antiphospholipid syndrome, such as thrombocytopenia, cardiac vegetations (Libman-Sachs endocarditis), livedo reticularis, and nephropathy.

Back to our patient

Our patient’s anticardiolipin IgG test is negative, while her lupus anticoagulant and beta-2GPI IgG are positive. She has no clinical or laboratory features suggesting lupus.

She is started on warfarin. After 3 months, the warfarin is interrupted for several days, and she is retested for all 3 antiphospholipid antibodies. Her beta-2GPI I IgG and lupus anticoagulant tests are again positive. Because of the persistent antiphospholipid antibody positivity and clinical history of deep vein thrombosis, her condition is diagnosed as primary antiphospholipid syndrome. She is advised to continue anticoagulant therapy indefinitely.

 

 

ANTINEUTROPHIL CYTOPLASMIC ANTIBODY

A 34-year-old man who is an injecting drug user presents with a 2-week history of fever, malaise, and generalized arthralgia. There are no localizing symptoms of infection. Notable findings on examination include a temperature of 38.0°C (100.4°F), needle track marks in his arms, nonblanching vasculitic rash in his legs, and a systolic murmur over the precordium.

His white blood cell count is 15.3 × 10⁹/L (reference range 3.7–11.0), and his C-reactive protein level is 234 mg/dL (normal < 3). Otherwise, results of blood cell counts, liver enzyme tests, renal function tests, urinalysis, and chest radiography are normal.

Two sets of blood cultures are drawn. Transthoracic echocardiography and the antineutrophil cytoplasmic antibody (ANCA) test are requested, as are screening tests for human immunodeficiency virus, hepatitis B, and hepatitis C.

Was the ANCA test indicated in this patient?

ANCAs are autoantibodies against antigens located in the cytoplasmic granules of neutrophils and monocytes. They are associated with small-vessel vasculitides such as granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), eosinophilic granulomatosis with polyangiitis (EGPA), and isolated pauciimmune crescentic glomerulonephritis, all collectively known as ANCA-associated vasculitis (AAV).³⁹

Laboratory methods to detect ANCA include indirect immunofluorescence and antigen-specific enzyme immunoassays. Indirect immunofluorescence only tells us whether or not an antibody that is targeting a cytoplasmic antigen is present. Based on the indirect immunofluorescent pattern, ANCA can be classified as follows:

• Perinuclear or p-ANCA (if the targeted antigen is located just around the nucleus and extends into it)
• Cytoplasmic or c-ANCA (if the targeted antigen is located farther away from the nucleus)
• Atypical ANCA (if the indirect immunofluorescent pattern does not fit with either p-ANCA or c-ANCA).

Indirect immunofluorescence does not give information about the exact antigen that is targeted; this can only be obtained by performing 1 of the antigen-specific immunoassays. The target antigen for c-ANCA is usually proteinase-3 (PR3), while that for p-ANCA could be myeloperoxidase (MPO), cathepsin, lysozyme, lactoferrin, or bactericidal permeability inhibitor. Anti-PR3 is highly specific for GPA, while anti-MPO is usually associated with MPA and EGPA. Less commonly, anti-PR3 may be seen in patients with MPA and anti-MPO in those with GPA. Hence, there is an increasing trend toward classifying ANCA-associated vasculitis into PR3-associated or MPO-associated vasculitis rather than as GPA, MPA, EGPA, or renal-limited vasculitis.⁴⁰

Several audits have shown that the ANCA test is widely misused and requested indiscriminately to rule out vasculitis. This results in a lower positive predictive value, possible harm to patients due to increased false-positive rates, and increased burden on the laboratory.^41–43 At least 2 separate groups have demonstrated that a gating policy that refuses ANCA testing in patients without clinical evidence of systemic vasculitis can reduce the number of inappropriate requests, improve the diagnostic yield, and make it more clinically relevant and cost-effective.^44,45

The clinician should bear in mind that:

Current guidelines recommend using one of the antigen-specific assays for PR3 and MPO as the primary screening method.⁴⁸ Until recently, indirect immunofluorescence was used to screen for ANCA-associated vasculitis, and positive results were confirmed by ELISA to detect ANCAs specific for PR3 and MPO,⁴⁹ but this is no longer recommended because of recent evidence suggesting a large variability between the different indirect immunofluorescent methods and improved diagnostic performance of the antigen-specific assays.

In a large multicenter study by Damoiseaux et al, the specificity with the different antigen-specific immunoassays was 98% to 99% for PR3-ANCA and 96% to 99% for MPO-ANCA.⁵⁰

ANCA-associated vasculitis should not be considered excluded if the PR3 and MPO-ANCA are negative. In the Damoiseaux study, about 11% to 15% of patients with GPA and 8% to 24% of patients with MPA tested negative for both PR3 and MPO-ANCA.⁵⁰

If the ANCA result is negative and clinical suspicion for ANCA-associated vasculitis is high, the clinician may wish to consider requesting another immunoassay method or indirect immunofluorescence. Results of indirect immunofluorescent testing results may be positive in those with a negative immunoassay, and vice versa.

Thus, the ANCA result should always be interpreted in the context of the whole clinical picture.⁵¹ Biopsy should still be considered the gold standard for the diagnosis of ANCA-associated vasculitis. The ANCA titer can help to improve clinical interpretation, because the likelihood of ANCA-associated vasculitis increases with higher levels of PR3 and MPO-ANCA.⁵²

Back to our patient

Our patient’s blood cultures grow methicillin-sensitive Staphylococcus aureus in both sets after 48 hours. Transthoracic echocardiography reveals vegetations around the tricuspid valve, with no evidence of valvular regurgitation. The diagnosis is right-sided infective endocarditis. He is started on appropriate antibiotics.

Tests for human immunodeficiency virus, hepatitis B, and hepatitis C are negative. The ANCA test is positive for MPO-ANCA at 28 IU/mL (normal < 10).

The positive ANCA is thought to be related to the infective endocarditis. His vasculitis is most likely secondary to infective endocarditis and not ANCA-associated vasculitis. The ANCA test need not have been requested in the first place.

 

 

HUMAN LEUKOCYTE ANTIGEN-B27

A 22-year-old man presents to his primary care physician with a 4-month history of gradually worsening low back pain associated with early morning stiffness lasting more than 2 hours. He has no peripheral joint symptoms.

In the last 2 years, he has had 2 separate episodes of uveitis. There is a family history of ankylosing spondylitis in his father. Examination reveals global restriction of lumbar movements but is otherwise unremarkable. Magnetic resonance imaging (MRI) of the lumbar spine and sacroiliac joints is normal.

Should this patient be tested for human leukocyte antigen-B27 (HLA-B27)?

The major histocompatibility complex (MHC) is a gene complex that is present in all animals. It encodes proteins that help with immunologic tolerance. HLA simply refers to the human version of the MHC.⁵³ The HLA gene complex, located on chromosome 6, is categorized into class I, class II, and class III. HLA-B is one of the 3 class I genes. Thus, a positive HLA-B27 result simply means that the particular gene is present in that person.

HLA-B27 is strongly associated with ankylosing spondylitis, also known as axial spondyloarthropathy.⁵⁴ Other genes also contribute to the pathogenesis of ankylosing spondylitis, but HLA-B27 is present in more than 90% of patients with this disease and is by far considered the most important. The association is not as strong for peripheral spondyloarthropathy, with studies reporting a frequency of up to 75% for reactive arthritis and inflammatory bowel disease-associated arthritis, and up to 50% for psoriatic arthritis and uveitis.⁵⁵

About 9% of healthy, asymptomatic individuals may have HLA-B27, so the mere presence of this gene is not evidence of disease.⁵⁶ There may be up to a 20-fold increased risk of ankylosing spondylitis among those who are HLA-B27-positive.⁵⁷

Some HLA genes have many different alleles, each of which is given a number (explaining the number 27 that follows the B). Closely related alleles that differ from one another by only a few amino-acid substitutions are then categorized together, thus accounting for more than 100 subtypes of HLA-B27 (designated from HLA-B*2701 to HLA-B*27106). These subtypes vary in frequency among different racial groups, and the population prevalence of ankylosing spondylitis parallels the frequency of HLA-B27.⁵⁸ The most common subtype seen in white people and American Indians is B*2705. HLA-B27 is rare in blacks, explaining the rarity of ankylosing spondylitis in this population. Further examples include HLA-B*2704, which is seen in Asians, and HLA-B*2702, seen in Mediterranean populations. Not all subtypes of HLA-B27 are associated with disease, and some, like HLA-B*2706, may also be protective.

When should the clinician consider testing for HLA-B27?

Peripheral spondyloarthropathy may present with arthritis, enthesitis (eg, heel pain due to inflammation at the site of insertion of the Achilles tendon or plantar fascia), or dactylitis (“sausage” swelling of the whole finger or toe due to extension of inflammation beyond the margins of the joint). Other clues may include psoriasis, inflammatory bowel disease, history of preceding gastrointestinal or genitourinary infection, family history of similar conditions, and history of recurrent uveitis.

For the initial assessment of patients who have inflammatory back pain, plain radiography of the sacroiliac joints is considered the gold standard.⁵⁹ If plain radiography does not show evidence of sacroiliitis, MRI of the sacroiliac joints should be considered. While plain radiography can reveal only structural changes such as sclerosis, erosions, and ankylosis, MRI is useful to evaluate for early inflammatory changes such as bone marrow edema. Imaging the lumbar spine is not necessary, as the sacroiliac joints are almost invariably involved in axial spondyloarthropathy, and lesions seldom occur in the lumbar spine in isolation.⁶⁰

The diagnosis of ankylosing spondylitis previously relied on confirmatory imaging features, but based on the new International Society classification criteria,^61–63 which can be applied to patients with more than 3 months of back pain and age of onset of symptoms before age 45, patients can be classified as having 1 of the following:

• Radiographic axial spondyloarthropathy, if they have evidence of sacroiliitis on imaging plus 1 other feature of spondyloarthropathy
• Nonradiographic axial spondyloarthropathy, if they have a positive HLA-B27 plus 2 other features of spondyloarthropathy (Table 7).

These new criteria have a sensitivity of 82.9% and specificity of 84.4%.^62,63 The disease burden of radiographic and nonradiographic axial spondyloarthropathy has been shown to be similar, suggesting that they are part of the same disease spectrum. Thus, the HLA-B27 test is useful to make a diagnosis of axial spondyloarthropathy even in the absence of imaging features and could be requested in patients with 2 or more features of spondyloarthropathy. In the absence of imaging features and a negative HLA-B27 result, however, the patient cannot be classified as having axial spondyloarthropathy.

Back to our patient

The absence of radiographic evidence would not exclude axial spondyloarthropathy in our patient. The HLA-B27 test is requested because of the inflammatory back pain and the presence of 2 spondyloarthropathy features (uveitis and the family history) and is reported to be positive. His disease is classified as nonradiographic axial spondyloarthropathy.

He is started on regular naproxen and is referred to a physiotherapist. After 1 month, he reports significant symptomatic improvement. He asks if he can be retested for HLA-B27 to see if it has become negative. We tell him that there is no point in repeating it, as it is a gene and will not disappear.

SUMMARY: CONSIDER THE CLINICAL PICTURE

When approaching a patient suspected of having a rheumatologic disease, a clinician should first consider the clinical presentation and the intended purpose of each test. The tests, in general, might serve several purposes. They might help to:

Increase the likelihood of the diagnosis in question. For example, a positive rheumatoid factor or anticitrullinated peptide antibody can help diagnose rheumatoid arthritis in a patient with early polyarthritis, a positive HLA-B27 can help diagnose ankylosing spondylitis in patients with inflammatory back pain and normal imaging, and a positive ANCA can help diagnose ANCA-associated vasculitis in a patient with glomerulonephritis.

Reduce the likelihood of the diagnosis in question. For example, a negative antinuclear antibody test reduces the likelihood of lupus in a patient with joint pains.

Monitor the condition. For example DNA antibodies can be used to monitor the activity of lupus.

Plan the treatment strategy. For example, one might consider lifelong anticoagulation if antiphospholipid antibodies are persistently positive in a patient with thrombosis.

Prognosticate. For example, positive rheumatoid factor and anticitrullinated peptide antibody increase the risk of erosive rheumatoid arthritis.

If the test was requested in the absence of a clear indication and the result is positive, it is important to bear in mind the potential pitfalls associated with that test and not attach a diagnostic label prematurely. None of the tests can confirm or exclude a condition, so the results should always be interpreted in the context of the whole clinical picture.   

Laboratory tests are often ordered inappropriately for patients in whom a rheumatologic illness is suspected; this occurs in both primary and secondary care.¹ Some tests are available both singly and as part of a battery of tests screening healthy people without symptoms.

The problem: negative test results are by no means always reassuring, and false-positive results raise the risks of unnecessary anxiety for patients and clinicians, needless referrals, and potential morbidity due to further unnecessary testing and exposure to wrong treatments.² Clinicians should be aware of the pitfalls of these tests in order to choose them wisely and interpret the results correctly.

This article provides practical guidance on requesting and interpreting some common tests in rheumatology, with the aid of case vignettes.

RHEUMATOID FACTOR AND ANTICITRULLINATED PEPTIDE ANTIBODY

A 41-year-old woman, previously in good health, presents to her primary care practitioner with a 6-week history of pain and swelling in her hands and early morning stiffness lasting about 2 hours. She denies having any extraarticular symptoms. Physical examination reveals synovitis across her right metacarpophalangeal joints, proximal interphalangeal joint of the left middle finger, and left wrist. The primary care physician is concerned that her symptoms might be due to rheumatoid arthritis.

Would testing for rheumatoid factor and anticitrullinated peptide antibody be useful in this patient?

Rheumatoid factor is an antibody (immunoglobulin M, IgG, or IgA) targeted against the Fc fragment of IgG.³ It was so named because it was originally detected in patients with rheumatoid arthritis, but it is neither sensitive nor specific for this condition. A meta-analysis of more than 5,000 patients with rheumatoid arthritis reported that rheumatoid factor testing had a sensitivity of 69% and specificity of 85%.⁴

Anticitrullinated peptide antibody, on the other hand, is much more specific for rheumatoid arthritis (95%), as it is seldom seen in other conditions, but its sensitivity is similar to that of rheumatoid factor (68%).^4–6 A positive result would thus lend strength to the diagnosis of rheumatoid arthritis, but a negative result would not exclude it.

Approach to early arthritis

When faced with a patient with early arthritis, some key questions to ask include^7,8:

Is this an inflammatory or a mechanical problem? Inflammatory arthritis is suggested by joint swelling that is not due to trauma or bony hypertrophy, early morning stiffness lasting longer than 30 minutes, and elevated inflammatory markers (erythrocyte sedimentation rate or C-reactive protein). Involvement of the small joints of the hands and feet may be suggested by pain on compression of the metacarpophalangeal and metatarsophalangeal joints, respectively.

Is there a definite identifiable underlying cause for the inflammatory arthritis? The pattern of development of joint symptoms or the presence of extraarticular symptoms may suggest an underlying problem such as gout, psoriatic arthritis, systemic lupus erythematosus, or sarcoidosis.

If the arthritis is undifferentiated (ie, there is no definite identifiable cause), is it likely to remit or persist? This is perhaps the most important question to ask in order to prognosticate. Patients with risk factors for persistent disease, ie, for development of rheumatoid arthritis, should be referred to a rheumatologist early for timely institution of disease-modifying antirheumatic drug therapy.⁹ Multiple studies have shown that patients in whom this therapy is started early have much better clinical, functional, and radiologic outcomes than those in whom it is delayed.^10–12

The revised American College of Rheumatology and European League Against Rheumatism criteria¹³ include the following factors as predictors of persistence:

• Number of involved joints (with greater weight given to involvement of small joints)
• Duration of symptoms 6 weeks or longer
• Elevated acute-phase response (erythrocyte sedimentation rate or C-reactive protein level)
• A positive serologic test (either rheumatoid factor or anticitrullinated peptide antibody).

If both rheumatoid factor and anticitrullinated peptide antibody are positive in a patient with early undifferentiated arthritis, the risk of progression to rheumatoid arthritis is almost 100%, thus underscoring the importance of testing for these antibodies.^5,6 Referral to a rheumatologist should, however, not be delayed in patients with negative test results (more than one-third of patients with rheumatoid arthritis may be negative for both), and should be considered in those with inflammatory joint symptoms persisting longer than 6 weeks, especially with involvement of the small joints (sparing the distal interphalangeals) and elevated acute-phase response.

Rheumatoid factor in healthy people without symptoms

In some countries, testing for rheumatoid factor is offered as part of a battery of screening tests in healthy people who have no symptoms, a practice that should be strongly discouraged.

Multiple studies, both prospective and retrospective, have demonstrated that both rheumatoid factor and anticitrullinated peptide antibody may be present several years before the clinical diagnosis of rheumatoid arthritis.^6,14–16 But the risk of developing rheumatoid arthritis for asymptomatic individuals who are rheumatoid factor-positive depends on the rheumatoid factor titer, positive family history of rheumatoid arthritis in first-degree relatives, and copresence of anticitrullinated peptide antibody. The absolute risk, nevertheless, is still very small. In some, there might be an alternative explanation such as undiagnosed Sjögren syndrome or hepatitis C.

In any event, no strategy is currently available that is proven to prevent the development of rheumatoid arthritis, and there is no role for disease-modifying therapy during the preclinical phase.¹⁶

Back to our patient

Blood testing in our patient reveals normal complete blood cell counts, aminotransferase levels, and serum creatinine concentration; findings on urinalysis are normal. Her erythrocyte sedimentation rate is 56 mm/hour (reference range 0–15), and her C-reactive protein level is 26 mg/dL (normal < 3). Testing is negative for rheumatoid factor and anticitrullinated peptide antibody.

Although her rheumatoid factor and anticitrullinated peptide antibody tests are negative, she is referred to a rheumatologist because she has predictors of persistent disease, ie, symptom duration of 6 weeks, involvement of the small joints of the hands, and elevated erythrocyte sedimentation rate and C-reactive protein. The rheumatologist checks her parvovirus serology, which is negative.

The patient is given parenteral depot corticosteroid therapy, to which she responds briefly. Because her symptoms persist and continue to worsen, methotrexate treatment is started after an additional 6 weeks.

 

 

ANTINUCLEAR ANTIBODY

A 37-year-old woman presents to her primary care physician with the complaint of tiredness. She has a family history of systemic lupus erythematosus in her sister and maternal aunt. She is understandably worried about lupus because of the family history and is asking to be tested for it.

Would testing for antinuclear antibody be reasonable?

Antinuclear antibody is not a single antibody but rather a family of autoantibodies that are directed against nuclear constituents such as single- or double-stranded deoxyribonucleic acid (dsDNA), histones, centromeres, proteins complexed with ribonucleic acid (RNA), and enzymes such as topoisomerase.^17,18

Protein antigens complexed with RNA and some enzymes in the nucleus are also known as extractable nuclear antigens (ENAs). They include Ro, La, Sm, Jo-1, RNP, and ScL-70 and are named after the patient in whom they were first discovered (Robert, Lavine, Smith, and John), the antigen that is targeted (ribonucleoprotein or RNP), and the disease with which they are associated (anti-ScL-70 or antitopoisomerase in diffuse cutaneous scleroderma).

Antinuclear antibody testing is commonly requested to exclude connective tissue diseases such as lupus, but the clinician needs to be aware of the following points:

Antinuclear antibody may be encountered in conditions other than lupus

These include¹⁹:

• Other autoimmune diseases such as rheumatoid arthritis, primary Sjögren syndrome, systemic sclerosis, autoimmune thyroid disease, and myasthenia gravis
• Infection with organisms that share the epitope with self-antigens (molecular mimicry)
• Cancers
• Drugs such as hydralazine, procainamide, and minocycline.

Antinuclear antibody might also be produced by the healthy immune system from time to time to clear the nuclear debris that is extruded from aging cells.

A study in healthy individuals²⁰ reported a prevalence of positive antinuclear antibody of 32% at a titer of 1/40, 15% at a titer of 1/80, 7% at a titer of 1/160, and 3% at a titer of 1/320. Importantly, a positive result was more common among family members of patients with autoimmune connective tissue diseases.²¹ Hence, a positive antinuclear antibody result does not always mean lupus.

Antinuclear antibody testing is highly sensitive for lupus

With current laboratory methods, antinuclear antibody testing has a sensitivity close to 100%. Hence, a negative result virtually rules out lupus.

Two methods are commonly used to test for antinuclear antibody: indirect immunofluorescence and enzyme-linked immunosorbent assay (ELISA).²² While human epithelial (Hep2) cells are used as the source of antigen in immunofluorescence, purified nuclear antigens coated on multiple-well plates are used in ELISA.

Although ELISA is simpler to perform, immunofluorescence has a slightly better sensitivity (because the Hep2 cells express a wide range of antigens) and is still considered the gold standard. As expected, the higher sensitivity occurs at the cost of reduced specificity (about 60%), so antinuclear antibody will also be detected in all the other conditions listed above.²³

To improve the specificity of antinuclear antibody testing, laboratories report titers (the highest dilution of the test serum that tested positive); a cutoff of greater than 1/80 is generally considered significant.

Do not order antinuclear antibody testing indiscriminately

To sum up, the antinuclear antibody test should be requested only in patients with involvement of multiple organ systems. Although a negative result would make it extremely unlikely that the clinical presentation is due to lupus, a positive result is insufficient on its own to make a diagnosis of lupus.

Diagnosing lupus is straightforward when patients present with a specific manifestation such as inflammatory arthritis, photosensitive skin rash, hemolytic anemia, thrombocytopenia, or nephritis, or with specific antibodies such as those against dsDNA or Sm. Patients who present with nonspecific symptoms such as arthralgia or tiredness with a positive antinuclear antibody and negative anti-dsDNA and anti-Sm may present difficulties even for the specialist.^25–27

Back to our patient

Our patient denies arthralgia. She has no extraarticular symptoms such as skin rashes, oral ulcers, sicca symptoms, muscle weakness, Raynaud phenomenon, pleuritic chest pain, or breathlessness. Findings on physical examination and urinalysis are unremarkable.

Her primary care physician decides to check her complete blood cell count, erythrocyte sedimentation rate, and thyroid-stimulating hormone level. Although she is reassured that her tiredness is not due to lupus, she insists on getting an antinuclear antibody test.

Her complete blood cell counts are normal. Her erythrocyte sedimentation rate is 6 mm/hour. However, her thyroid-stimulating hormone level is elevated, and subsequent testing shows low free thyroxine and positive thyroid peroxidase antibodies. The antinuclear antibody is positive in a titer of 1/80 and negative for anti-dsDNA and anti-ENA.

We explain to her that the positive antinuclear antibody is most likely related to her autoimmune thyroid disease. She is referred to an endocrinologist.

 

 

ANTIPHOSPHOLIPID ANTIBODIES

A 24-year-old woman presents to the emergency department with acute unprovoked deep vein thrombosis in her right leg, confirmed by ultrasonography. She has no history of previous thrombosis, and the relevant family history is unremarkable. She has never been pregnant. Her platelet count is 84 × 10⁹/L (reference range 150–400), and her baseline activated partial thromboplastin time is prolonged at 62 seconds (reference range 23.0–32.4). The rest of her blood counts and her prothrombin time, liver enzyme levels, and serum creatinine level are normal.

Should this patient be tested for antiphospholipid antibodies?

Antiphospholipid antibodies are important because of their association with thrombotic risk (both venous and arterial) and pregnancy morbidity. The name is a misnomer, as these antibodies are targeted against some proteins that are bound to phospholipids and not only to the phospholipids themselves.

According to the modified Sapporo criteria for the classification of antiphospholipid syndrome,²⁸ antiphospholipid antibodies should remain persistently positive on at least 2 separate occasions at least 12 weeks apart for the result to be considered significant because some infections and drugs may be associated with the transient presence of antiphospholipid antibodies.

Screening for antiphospholipid antibodies should include testing for IgM and IgG anticardiolipin antibodies, lupus anticoagulant, and IgM and IgG beta-2 glycoprotein I antibodies.^29,30

Anticardiolipin antibodies

Anticardiolipin (aCL) antibodies may be targeted either against beta-2 glycoprotein I (beta-2GPI) that is bound to cardiolipin (a phospholipid) or against cardiolipin alone; the former is more specific. Antibodies directed against cardiolipin alone are usually transient and are associated with infections and drugs. The result is considered significant only when anticardiolipin antibodies are present in a medium to high titer (> 40 IgG phospholipid units or IgM phospholipid units, or > 99th percentile).

Lupus anticoagulant

The antibody with “lupus anticoagulant activity” is targeted against prothrombin plus phospholipid or beta-2GPI plus phospholipid. The test for it is a functional assay involving 3 steps:

Demonstrating the prolongation of a phospholipid-dependent coagulation assay like the activated partial thromboplastin time (aPTT). (This may explain the prolongation of aPTT in the patient described in the vignette.) Although the presence of lupus anticoagulant is associated with thrombosis, it is called an “anticoagulant” because of this in vitro prolongation of phospholipid-dependent coagulation assays.

Mixing study. The phospholipid-dependent coagulation assay could be prolonged because of either the deficiency of a coagulation factor or the presence of the antiphospholipid antibodies. This can be differentiated by mixing the patient’s plasma with normal plasma (which will have all the clotting factors) in a 1:1 ratio. If the coagulation assay remains prolonged after the addition of normal plasma, clotting factor deficiency can be excluded.

Addition of a phospholipid. If the prolongation of the coagulation assay is due to the presence of an antiphospholipid antibody, addition of extra phospholipid will correct this.

Beta-2 glycoprotein I antibody (anti-beta-2GPI)

The beta-2GPI that is not bound to the cardiolipin can be detected by separately testing for beta-2GPI (the anticardiolipin test only detects the beta-2GPI that is bound to the cardiolipin). The result is considered significant if beta-2GPI is present in a medium to high titer (> 99th percentile).

Studies have shown that antiphospholipid antibodies may be present in 1% to 5% of apparently healthy people in the general population.³¹ These are usually low-titer anticardiolipin or anti-beta-GPI IgM antibodies that are not associated with thrombosis or adverse pregnancy outcomes. Hence, the term antiphospholipid syndrome should be reserved for those who have had at least 1 episode of thrombosis or pregnancy morbidity and persistent antiphospholipid antibodies, and not those who have asymptomatic or transient antiphospholipid antibodies.

Triple positivity (positive anticardiolipin, lupus anticoagulant, and anti-beta-2GPI) seems to be associated with the highest risk of thrombosis, with a 10-year cumulative incidence of 37.1% (95% confidence interval [CI] 19.9–54.3) for a first thrombotic event,³² and 44.2% (95% CI 38.6–49.8) for recurrent thrombosis.³³

The association with thrombosis is stronger for lupus anticoagulant than with the other 2 antibodies, with different studies³⁴ finding an odds ratio ranging from 5 to 16. A positive lupus anticoagulant test with or without a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a high-risk profile, while a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a moderate-risk profile. A low titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a low-risk profile that may not be associated with thrombosis.³⁵

Antiphospholipid syndrome is important to recognize because of the need for long-term anticoagulation to prevent recurrence.³⁶ It may be primary, when it occurs on its own, or secondary, when it occurs in association with another autoimmune disease such as lupus.

Venous events in antiphospholipid syndrome most commonly manifest as lower-limb deep vein thrombosis or pulmonary embolism, while arterial events most commonly manifest as stroke or transient ischemic attack.³⁷ Obstetric manifestations may include not only miscarriage and stillbirth, but also preterm delivery, intrauterine growth retardation, and preeclampsia, all occurring due to placental insufficiency.

The frequency of antiphospholipid antibodies has been estimated as 13.5% in patients with stroke, 11% with myocardial infarction, 9.5% with deep vein thrombosis, and 6% for those with pregnancy morbidity.³⁸

Some noncriteria manifestations have also been recognized in antiphospholipid syndrome, such as thrombocytopenia, cardiac vegetations (Libman-Sachs endocarditis), livedo reticularis, and nephropathy.

Back to our patient

Our patient’s anticardiolipin IgG test is negative, while her lupus anticoagulant and beta-2GPI IgG are positive. She has no clinical or laboratory features suggesting lupus.

She is started on warfarin. After 3 months, the warfarin is interrupted for several days, and she is retested for all 3 antiphospholipid antibodies. Her beta-2GPI I IgG and lupus anticoagulant tests are again positive. Because of the persistent antiphospholipid antibody positivity and clinical history of deep vein thrombosis, her condition is diagnosed as primary antiphospholipid syndrome. She is advised to continue anticoagulant therapy indefinitely.

 

 

ANTINEUTROPHIL CYTOPLASMIC ANTIBODY

A 34-year-old man who is an injecting drug user presents with a 2-week history of fever, malaise, and generalized arthralgia. There are no localizing symptoms of infection. Notable findings on examination include a temperature of 38.0°C (100.4°F), needle track marks in his arms, nonblanching vasculitic rash in his legs, and a systolic murmur over the precordium.

His white blood cell count is 15.3 × 10⁹/L (reference range 3.7–11.0), and his C-reactive protein level is 234 mg/dL (normal < 3). Otherwise, results of blood cell counts, liver enzyme tests, renal function tests, urinalysis, and chest radiography are normal.

Two sets of blood cultures are drawn. Transthoracic echocardiography and the antineutrophil cytoplasmic antibody (ANCA) test are requested, as are screening tests for human immunodeficiency virus, hepatitis B, and hepatitis C.

Was the ANCA test indicated in this patient?

ANCAs are autoantibodies against antigens located in the cytoplasmic granules of neutrophils and monocytes. They are associated with small-vessel vasculitides such as granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), eosinophilic granulomatosis with polyangiitis (EGPA), and isolated pauciimmune crescentic glomerulonephritis, all collectively known as ANCA-associated vasculitis (AAV).³⁹

Laboratory methods to detect ANCA include indirect immunofluorescence and antigen-specific enzyme immunoassays. Indirect immunofluorescence only tells us whether or not an antibody that is targeting a cytoplasmic antigen is present. Based on the indirect immunofluorescent pattern, ANCA can be classified as follows:

• Perinuclear or p-ANCA (if the targeted antigen is located just around the nucleus and extends into it)
• Cytoplasmic or c-ANCA (if the targeted antigen is located farther away from the nucleus)
• Atypical ANCA (if the indirect immunofluorescent pattern does not fit with either p-ANCA or c-ANCA).

Indirect immunofluorescence does not give information about the exact antigen that is targeted; this can only be obtained by performing 1 of the antigen-specific immunoassays. The target antigen for c-ANCA is usually proteinase-3 (PR3), while that for p-ANCA could be myeloperoxidase (MPO), cathepsin, lysozyme, lactoferrin, or bactericidal permeability inhibitor. Anti-PR3 is highly specific for GPA, while anti-MPO is usually associated with MPA and EGPA. Less commonly, anti-PR3 may be seen in patients with MPA and anti-MPO in those with GPA. Hence, there is an increasing trend toward classifying ANCA-associated vasculitis into PR3-associated or MPO-associated vasculitis rather than as GPA, MPA, EGPA, or renal-limited vasculitis.⁴⁰

Several audits have shown that the ANCA test is widely misused and requested indiscriminately to rule out vasculitis. This results in a lower positive predictive value, possible harm to patients due to increased false-positive rates, and increased burden on the laboratory.^41–43 At least 2 separate groups have demonstrated that a gating policy that refuses ANCA testing in patients without clinical evidence of systemic vasculitis can reduce the number of inappropriate requests, improve the diagnostic yield, and make it more clinically relevant and cost-effective.^44,45

The clinician should bear in mind that:

Current guidelines recommend using one of the antigen-specific assays for PR3 and MPO as the primary screening method.⁴⁸ Until recently, indirect immunofluorescence was used to screen for ANCA-associated vasculitis, and positive results were confirmed by ELISA to detect ANCAs specific for PR3 and MPO,⁴⁹ but this is no longer recommended because of recent evidence suggesting a large variability between the different indirect immunofluorescent methods and improved diagnostic performance of the antigen-specific assays.

In a large multicenter study by Damoiseaux et al, the specificity with the different antigen-specific immunoassays was 98% to 99% for PR3-ANCA and 96% to 99% for MPO-ANCA.⁵⁰

ANCA-associated vasculitis should not be considered excluded if the PR3 and MPO-ANCA are negative. In the Damoiseaux study, about 11% to 15% of patients with GPA and 8% to 24% of patients with MPA tested negative for both PR3 and MPO-ANCA.⁵⁰

If the ANCA result is negative and clinical suspicion for ANCA-associated vasculitis is high, the clinician may wish to consider requesting another immunoassay method or indirect immunofluorescence. Results of indirect immunofluorescent testing results may be positive in those with a negative immunoassay, and vice versa.

Thus, the ANCA result should always be interpreted in the context of the whole clinical picture.⁵¹ Biopsy should still be considered the gold standard for the diagnosis of ANCA-associated vasculitis. The ANCA titer can help to improve clinical interpretation, because the likelihood of ANCA-associated vasculitis increases with higher levels of PR3 and MPO-ANCA.⁵²

Back to our patient

Our patient’s blood cultures grow methicillin-sensitive Staphylococcus aureus in both sets after 48 hours. Transthoracic echocardiography reveals vegetations around the tricuspid valve, with no evidence of valvular regurgitation. The diagnosis is right-sided infective endocarditis. He is started on appropriate antibiotics.

Tests for human immunodeficiency virus, hepatitis B, and hepatitis C are negative. The ANCA test is positive for MPO-ANCA at 28 IU/mL (normal < 10).

The positive ANCA is thought to be related to the infective endocarditis. His vasculitis is most likely secondary to infective endocarditis and not ANCA-associated vasculitis. The ANCA test need not have been requested in the first place.

 

 

HUMAN LEUKOCYTE ANTIGEN-B27

A 22-year-old man presents to his primary care physician with a 4-month history of gradually worsening low back pain associated with early morning stiffness lasting more than 2 hours. He has no peripheral joint symptoms.

In the last 2 years, he has had 2 separate episodes of uveitis. There is a family history of ankylosing spondylitis in his father. Examination reveals global restriction of lumbar movements but is otherwise unremarkable. Magnetic resonance imaging (MRI) of the lumbar spine and sacroiliac joints is normal.

Should this patient be tested for human leukocyte antigen-B27 (HLA-B27)?

The major histocompatibility complex (MHC) is a gene complex that is present in all animals. It encodes proteins that help with immunologic tolerance. HLA simply refers to the human version of the MHC.⁵³ The HLA gene complex, located on chromosome 6, is categorized into class I, class II, and class III. HLA-B is one of the 3 class I genes. Thus, a positive HLA-B27 result simply means that the particular gene is present in that person.

HLA-B27 is strongly associated with ankylosing spondylitis, also known as axial spondyloarthropathy.⁵⁴ Other genes also contribute to the pathogenesis of ankylosing spondylitis, but HLA-B27 is present in more than 90% of patients with this disease and is by far considered the most important. The association is not as strong for peripheral spondyloarthropathy, with studies reporting a frequency of up to 75% for reactive arthritis and inflammatory bowel disease-associated arthritis, and up to 50% for psoriatic arthritis and uveitis.⁵⁵

About 9% of healthy, asymptomatic individuals may have HLA-B27, so the mere presence of this gene is not evidence of disease.⁵⁶ There may be up to a 20-fold increased risk of ankylosing spondylitis among those who are HLA-B27-positive.⁵⁷

Some HLA genes have many different alleles, each of which is given a number (explaining the number 27 that follows the B). Closely related alleles that differ from one another by only a few amino-acid substitutions are then categorized together, thus accounting for more than 100 subtypes of HLA-B27 (designated from HLA-B*2701 to HLA-B*27106). These subtypes vary in frequency among different racial groups, and the population prevalence of ankylosing spondylitis parallels the frequency of HLA-B27.⁵⁸ The most common subtype seen in white people and American Indians is B*2705. HLA-B27 is rare in blacks, explaining the rarity of ankylosing spondylitis in this population. Further examples include HLA-B*2704, which is seen in Asians, and HLA-B*2702, seen in Mediterranean populations. Not all subtypes of HLA-B27 are associated with disease, and some, like HLA-B*2706, may also be protective.

When should the clinician consider testing for HLA-B27?

Peripheral spondyloarthropathy may present with arthritis, enthesitis (eg, heel pain due to inflammation at the site of insertion of the Achilles tendon or plantar fascia), or dactylitis (“sausage” swelling of the whole finger or toe due to extension of inflammation beyond the margins of the joint). Other clues may include psoriasis, inflammatory bowel disease, history of preceding gastrointestinal or genitourinary infection, family history of similar conditions, and history of recurrent uveitis.

For the initial assessment of patients who have inflammatory back pain, plain radiography of the sacroiliac joints is considered the gold standard.⁵⁹ If plain radiography does not show evidence of sacroiliitis, MRI of the sacroiliac joints should be considered. While plain radiography can reveal only structural changes such as sclerosis, erosions, and ankylosis, MRI is useful to evaluate for early inflammatory changes such as bone marrow edema. Imaging the lumbar spine is not necessary, as the sacroiliac joints are almost invariably involved in axial spondyloarthropathy, and lesions seldom occur in the lumbar spine in isolation.⁶⁰

The diagnosis of ankylosing spondylitis previously relied on confirmatory imaging features, but based on the new International Society classification criteria,^61–63 which can be applied to patients with more than 3 months of back pain and age of onset of symptoms before age 45, patients can be classified as having 1 of the following:

• Radiographic axial spondyloarthropathy, if they have evidence of sacroiliitis on imaging plus 1 other feature of spondyloarthropathy
• Nonradiographic axial spondyloarthropathy, if they have a positive HLA-B27 plus 2 other features of spondyloarthropathy (Table 7).

These new criteria have a sensitivity of 82.9% and specificity of 84.4%.^62,63 The disease burden of radiographic and nonradiographic axial spondyloarthropathy has been shown to be similar, suggesting that they are part of the same disease spectrum. Thus, the HLA-B27 test is useful to make a diagnosis of axial spondyloarthropathy even in the absence of imaging features and could be requested in patients with 2 or more features of spondyloarthropathy. In the absence of imaging features and a negative HLA-B27 result, however, the patient cannot be classified as having axial spondyloarthropathy.

Back to our patient

The absence of radiographic evidence would not exclude axial spondyloarthropathy in our patient. The HLA-B27 test is requested because of the inflammatory back pain and the presence of 2 spondyloarthropathy features (uveitis and the family history) and is reported to be positive. His disease is classified as nonradiographic axial spondyloarthropathy.

He is started on regular naproxen and is referred to a physiotherapist. After 1 month, he reports significant symptomatic improvement. He asks if he can be retested for HLA-B27 to see if it has become negative. We tell him that there is no point in repeating it, as it is a gene and will not disappear.

SUMMARY: CONSIDER THE CLINICAL PICTURE

When approaching a patient suspected of having a rheumatologic disease, a clinician should first consider the clinical presentation and the intended purpose of each test. The tests, in general, might serve several purposes. They might help to:

Increase the likelihood of the diagnosis in question. For example, a positive rheumatoid factor or anticitrullinated peptide antibody can help diagnose rheumatoid arthritis in a patient with early polyarthritis, a positive HLA-B27 can help diagnose ankylosing spondylitis in patients with inflammatory back pain and normal imaging, and a positive ANCA can help diagnose ANCA-associated vasculitis in a patient with glomerulonephritis.

Reduce the likelihood of the diagnosis in question. For example, a negative antinuclear antibody test reduces the likelihood of lupus in a patient with joint pains.

Monitor the condition. For example DNA antibodies can be used to monitor the activity of lupus.

Plan the treatment strategy. For example, one might consider lifelong anticoagulation if antiphospholipid antibodies are persistently positive in a patient with thrombosis.

Prognosticate. For example, positive rheumatoid factor and anticitrullinated peptide antibody increase the risk of erosive rheumatoid arthritis.

If the test was requested in the absence of a clear indication and the result is positive, it is important to bear in mind the potential pitfalls associated with that test and not attach a diagnostic label prematurely. None of the tests can confirm or exclude a condition, so the results should always be interpreted in the context of the whole clinical picture.   

Laboratory tests are often ordered inappropriately for patients in whom a rheumatologic illness is suspected; this occurs in both primary and secondary care.¹ Some tests are available both singly and as part of a battery of tests screening healthy people without symptoms.

The problem: negative test results are by no means always reassuring, and false-positive results raise the risks of unnecessary anxiety for patients and clinicians, needless referrals, and potential morbidity due to further unnecessary testing and exposure to wrong treatments.² Clinicians should be aware of the pitfalls of these tests in order to choose them wisely and interpret the results correctly.

This article provides practical guidance on requesting and interpreting some common tests in rheumatology, with the aid of case vignettes.

RHEUMATOID FACTOR AND ANTICITRULLINATED PEPTIDE ANTIBODY

A 41-year-old woman, previously in good health, presents to her primary care practitioner with a 6-week history of pain and swelling in her hands and early morning stiffness lasting about 2 hours. She denies having any extraarticular symptoms. Physical examination reveals synovitis across her right metacarpophalangeal joints, proximal interphalangeal joint of the left middle finger, and left wrist. The primary care physician is concerned that her symptoms might be due to rheumatoid arthritis.

Would testing for rheumatoid factor and anticitrullinated peptide antibody be useful in this patient?

Rheumatoid factor is an antibody (immunoglobulin M, IgG, or IgA) targeted against the Fc fragment of IgG.³ It was so named because it was originally detected in patients with rheumatoid arthritis, but it is neither sensitive nor specific for this condition. A meta-analysis of more than 5,000 patients with rheumatoid arthritis reported that rheumatoid factor testing had a sensitivity of 69% and specificity of 85%.⁴

Anticitrullinated peptide antibody, on the other hand, is much more specific for rheumatoid arthritis (95%), as it is seldom seen in other conditions, but its sensitivity is similar to that of rheumatoid factor (68%).^4–6 A positive result would thus lend strength to the diagnosis of rheumatoid arthritis, but a negative result would not exclude it.

Approach to early arthritis

When faced with a patient with early arthritis, some key questions to ask include^7,8:

Is this an inflammatory or a mechanical problem? Inflammatory arthritis is suggested by joint swelling that is not due to trauma or bony hypertrophy, early morning stiffness lasting longer than 30 minutes, and elevated inflammatory markers (erythrocyte sedimentation rate or C-reactive protein). Involvement of the small joints of the hands and feet may be suggested by pain on compression of the metacarpophalangeal and metatarsophalangeal joints, respectively.

Is there a definite identifiable underlying cause for the inflammatory arthritis? The pattern of development of joint symptoms or the presence of extraarticular symptoms may suggest an underlying problem such as gout, psoriatic arthritis, systemic lupus erythematosus, or sarcoidosis.

If the arthritis is undifferentiated (ie, there is no definite identifiable cause), is it likely to remit or persist? This is perhaps the most important question to ask in order to prognosticate. Patients with risk factors for persistent disease, ie, for development of rheumatoid arthritis, should be referred to a rheumatologist early for timely institution of disease-modifying antirheumatic drug therapy.⁹ Multiple studies have shown that patients in whom this therapy is started early have much better clinical, functional, and radiologic outcomes than those in whom it is delayed.^10–12

The revised American College of Rheumatology and European League Against Rheumatism criteria¹³ include the following factors as predictors of persistence:

• Number of involved joints (with greater weight given to involvement of small joints)
• Duration of symptoms 6 weeks or longer
• Elevated acute-phase response (erythrocyte sedimentation rate or C-reactive protein level)
• A positive serologic test (either rheumatoid factor or anticitrullinated peptide antibody).

If both rheumatoid factor and anticitrullinated peptide antibody are positive in a patient with early undifferentiated arthritis, the risk of progression to rheumatoid arthritis is almost 100%, thus underscoring the importance of testing for these antibodies.^5,6 Referral to a rheumatologist should, however, not be delayed in patients with negative test results (more than one-third of patients with rheumatoid arthritis may be negative for both), and should be considered in those with inflammatory joint symptoms persisting longer than 6 weeks, especially with involvement of the small joints (sparing the distal interphalangeals) and elevated acute-phase response.

Rheumatoid factor in healthy people without symptoms

In some countries, testing for rheumatoid factor is offered as part of a battery of screening tests in healthy people who have no symptoms, a practice that should be strongly discouraged.

Multiple studies, both prospective and retrospective, have demonstrated that both rheumatoid factor and anticitrullinated peptide antibody may be present several years before the clinical diagnosis of rheumatoid arthritis.^6,14–16 But the risk of developing rheumatoid arthritis for asymptomatic individuals who are rheumatoid factor-positive depends on the rheumatoid factor titer, positive family history of rheumatoid arthritis in first-degree relatives, and copresence of anticitrullinated peptide antibody. The absolute risk, nevertheless, is still very small. In some, there might be an alternative explanation such as undiagnosed Sjögren syndrome or hepatitis C.

In any event, no strategy is currently available that is proven to prevent the development of rheumatoid arthritis, and there is no role for disease-modifying therapy during the preclinical phase.¹⁶

Back to our patient

Blood testing in our patient reveals normal complete blood cell counts, aminotransferase levels, and serum creatinine concentration; findings on urinalysis are normal. Her erythrocyte sedimentation rate is 56 mm/hour (reference range 0–15), and her C-reactive protein level is 26 mg/dL (normal < 3). Testing is negative for rheumatoid factor and anticitrullinated peptide antibody.

Although her rheumatoid factor and anticitrullinated peptide antibody tests are negative, she is referred to a rheumatologist because she has predictors of persistent disease, ie, symptom duration of 6 weeks, involvement of the small joints of the hands, and elevated erythrocyte sedimentation rate and C-reactive protein. The rheumatologist checks her parvovirus serology, which is negative.

The patient is given parenteral depot corticosteroid therapy, to which she responds briefly. Because her symptoms persist and continue to worsen, methotrexate treatment is started after an additional 6 weeks.

 

 

ANTINUCLEAR ANTIBODY

A 37-year-old woman presents to her primary care physician with the complaint of tiredness. She has a family history of systemic lupus erythematosus in her sister and maternal aunt. She is understandably worried about lupus because of the family history and is asking to be tested for it.

Would testing for antinuclear antibody be reasonable?

Antinuclear antibody is not a single antibody but rather a family of autoantibodies that are directed against nuclear constituents such as single- or double-stranded deoxyribonucleic acid (dsDNA), histones, centromeres, proteins complexed with ribonucleic acid (RNA), and enzymes such as topoisomerase.^17,18

Protein antigens complexed with RNA and some enzymes in the nucleus are also known as extractable nuclear antigens (ENAs). They include Ro, La, Sm, Jo-1, RNP, and ScL-70 and are named after the patient in whom they were first discovered (Robert, Lavine, Smith, and John), the antigen that is targeted (ribonucleoprotein or RNP), and the disease with which they are associated (anti-ScL-70 or antitopoisomerase in diffuse cutaneous scleroderma).

Antinuclear antibody testing is commonly requested to exclude connective tissue diseases such as lupus, but the clinician needs to be aware of the following points:

Antinuclear antibody may be encountered in conditions other than lupus

These include¹⁹:

• Other autoimmune diseases such as rheumatoid arthritis, primary Sjögren syndrome, systemic sclerosis, autoimmune thyroid disease, and myasthenia gravis
• Infection with organisms that share the epitope with self-antigens (molecular mimicry)
• Cancers
• Drugs such as hydralazine, procainamide, and minocycline.

Antinuclear antibody might also be produced by the healthy immune system from time to time to clear the nuclear debris that is extruded from aging cells.

A study in healthy individuals²⁰ reported a prevalence of positive antinuclear antibody of 32% at a titer of 1/40, 15% at a titer of 1/80, 7% at a titer of 1/160, and 3% at a titer of 1/320. Importantly, a positive result was more common among family members of patients with autoimmune connective tissue diseases.²¹ Hence, a positive antinuclear antibody result does not always mean lupus.

Antinuclear antibody testing is highly sensitive for lupus

With current laboratory methods, antinuclear antibody testing has a sensitivity close to 100%. Hence, a negative result virtually rules out lupus.

Two methods are commonly used to test for antinuclear antibody: indirect immunofluorescence and enzyme-linked immunosorbent assay (ELISA).²² While human epithelial (Hep2) cells are used as the source of antigen in immunofluorescence, purified nuclear antigens coated on multiple-well plates are used in ELISA.

Although ELISA is simpler to perform, immunofluorescence has a slightly better sensitivity (because the Hep2 cells express a wide range of antigens) and is still considered the gold standard. As expected, the higher sensitivity occurs at the cost of reduced specificity (about 60%), so antinuclear antibody will also be detected in all the other conditions listed above.²³

To improve the specificity of antinuclear antibody testing, laboratories report titers (the highest dilution of the test serum that tested positive); a cutoff of greater than 1/80 is generally considered significant.

Do not order antinuclear antibody testing indiscriminately

To sum up, the antinuclear antibody test should be requested only in patients with involvement of multiple organ systems. Although a negative result would make it extremely unlikely that the clinical presentation is due to lupus, a positive result is insufficient on its own to make a diagnosis of lupus.

Diagnosing lupus is straightforward when patients present with a specific manifestation such as inflammatory arthritis, photosensitive skin rash, hemolytic anemia, thrombocytopenia, or nephritis, or with specific antibodies such as those against dsDNA or Sm. Patients who present with nonspecific symptoms such as arthralgia or tiredness with a positive antinuclear antibody and negative anti-dsDNA and anti-Sm may present difficulties even for the specialist.^25–27

Back to our patient

Our patient denies arthralgia. She has no extraarticular symptoms such as skin rashes, oral ulcers, sicca symptoms, muscle weakness, Raynaud phenomenon, pleuritic chest pain, or breathlessness. Findings on physical examination and urinalysis are unremarkable.

Her primary care physician decides to check her complete blood cell count, erythrocyte sedimentation rate, and thyroid-stimulating hormone level. Although she is reassured that her tiredness is not due to lupus, she insists on getting an antinuclear antibody test.

Her complete blood cell counts are normal. Her erythrocyte sedimentation rate is 6 mm/hour. However, her thyroid-stimulating hormone level is elevated, and subsequent testing shows low free thyroxine and positive thyroid peroxidase antibodies. The antinuclear antibody is positive in a titer of 1/80 and negative for anti-dsDNA and anti-ENA.

We explain to her that the positive antinuclear antibody is most likely related to her autoimmune thyroid disease. She is referred to an endocrinologist.

 

 

ANTIPHOSPHOLIPID ANTIBODIES

A 24-year-old woman presents to the emergency department with acute unprovoked deep vein thrombosis in her right leg, confirmed by ultrasonography. She has no history of previous thrombosis, and the relevant family history is unremarkable. She has never been pregnant. Her platelet count is 84 × 10⁹/L (reference range 150–400), and her baseline activated partial thromboplastin time is prolonged at 62 seconds (reference range 23.0–32.4). The rest of her blood counts and her prothrombin time, liver enzyme levels, and serum creatinine level are normal.

Should this patient be tested for antiphospholipid antibodies?

Antiphospholipid antibodies are important because of their association with thrombotic risk (both venous and arterial) and pregnancy morbidity. The name is a misnomer, as these antibodies are targeted against some proteins that are bound to phospholipids and not only to the phospholipids themselves.

According to the modified Sapporo criteria for the classification of antiphospholipid syndrome,²⁸ antiphospholipid antibodies should remain persistently positive on at least 2 separate occasions at least 12 weeks apart for the result to be considered significant because some infections and drugs may be associated with the transient presence of antiphospholipid antibodies.

Screening for antiphospholipid antibodies should include testing for IgM and IgG anticardiolipin antibodies, lupus anticoagulant, and IgM and IgG beta-2 glycoprotein I antibodies.^29,30

Anticardiolipin antibodies

Anticardiolipin (aCL) antibodies may be targeted either against beta-2 glycoprotein I (beta-2GPI) that is bound to cardiolipin (a phospholipid) or against cardiolipin alone; the former is more specific. Antibodies directed against cardiolipin alone are usually transient and are associated with infections and drugs. The result is considered significant only when anticardiolipin antibodies are present in a medium to high titer (> 40 IgG phospholipid units or IgM phospholipid units, or > 99th percentile).

Lupus anticoagulant

The antibody with “lupus anticoagulant activity” is targeted against prothrombin plus phospholipid or beta-2GPI plus phospholipid. The test for it is a functional assay involving 3 steps:

Demonstrating the prolongation of a phospholipid-dependent coagulation assay like the activated partial thromboplastin time (aPTT). (This may explain the prolongation of aPTT in the patient described in the vignette.) Although the presence of lupus anticoagulant is associated with thrombosis, it is called an “anticoagulant” because of this in vitro prolongation of phospholipid-dependent coagulation assays.

Mixing study. The phospholipid-dependent coagulation assay could be prolonged because of either the deficiency of a coagulation factor or the presence of the antiphospholipid antibodies. This can be differentiated by mixing the patient’s plasma with normal plasma (which will have all the clotting factors) in a 1:1 ratio. If the coagulation assay remains prolonged after the addition of normal plasma, clotting factor deficiency can be excluded.

Addition of a phospholipid. If the prolongation of the coagulation assay is due to the presence of an antiphospholipid antibody, addition of extra phospholipid will correct this.

Beta-2 glycoprotein I antibody (anti-beta-2GPI)

The beta-2GPI that is not bound to the cardiolipin can be detected by separately testing for beta-2GPI (the anticardiolipin test only detects the beta-2GPI that is bound to the cardiolipin). The result is considered significant if beta-2GPI is present in a medium to high titer (> 99th percentile).

Studies have shown that antiphospholipid antibodies may be present in 1% to 5% of apparently healthy people in the general population.³¹ These are usually low-titer anticardiolipin or anti-beta-GPI IgM antibodies that are not associated with thrombosis or adverse pregnancy outcomes. Hence, the term antiphospholipid syndrome should be reserved for those who have had at least 1 episode of thrombosis or pregnancy morbidity and persistent antiphospholipid antibodies, and not those who have asymptomatic or transient antiphospholipid antibodies.

Triple positivity (positive anticardiolipin, lupus anticoagulant, and anti-beta-2GPI) seems to be associated with the highest risk of thrombosis, with a 10-year cumulative incidence of 37.1% (95% confidence interval [CI] 19.9–54.3) for a first thrombotic event,³² and 44.2% (95% CI 38.6–49.8) for recurrent thrombosis.³³

The association with thrombosis is stronger for lupus anticoagulant than with the other 2 antibodies, with different studies³⁴ finding an odds ratio ranging from 5 to 16. A positive lupus anticoagulant test with or without a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a high-risk profile, while a moderate to high titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a moderate-risk profile. A low titer of anticardiolipin or anti-beta-2GPI IgM or IgG constitutes a low-risk profile that may not be associated with thrombosis.³⁵

Antiphospholipid syndrome is important to recognize because of the need for long-term anticoagulation to prevent recurrence.³⁶ It may be primary, when it occurs on its own, or secondary, when it occurs in association with another autoimmune disease such as lupus.

Venous events in antiphospholipid syndrome most commonly manifest as lower-limb deep vein thrombosis or pulmonary embolism, while arterial events most commonly manifest as stroke or transient ischemic attack.³⁷ Obstetric manifestations may include not only miscarriage and stillbirth, but also preterm delivery, intrauterine growth retardation, and preeclampsia, all occurring due to placental insufficiency.

The frequency of antiphospholipid antibodies has been estimated as 13.5% in patients with stroke, 11% with myocardial infarction, 9.5% with deep vein thrombosis, and 6% for those with pregnancy morbidity.³⁸

Some noncriteria manifestations have also been recognized in antiphospholipid syndrome, such as thrombocytopenia, cardiac vegetations (Libman-Sachs endocarditis), livedo reticularis, and nephropathy.

Back to our patient

Our patient’s anticardiolipin IgG test is negative, while her lupus anticoagulant and beta-2GPI IgG are positive. She has no clinical or laboratory features suggesting lupus.

She is started on warfarin. After 3 months, the warfarin is interrupted for several days, and she is retested for all 3 antiphospholipid antibodies. Her beta-2GPI I IgG and lupus anticoagulant tests are again positive. Because of the persistent antiphospholipid antibody positivity and clinical history of deep vein thrombosis, her condition is diagnosed as primary antiphospholipid syndrome. She is advised to continue anticoagulant therapy indefinitely.

 

 

ANTINEUTROPHIL CYTOPLASMIC ANTIBODY

A 34-year-old man who is an injecting drug user presents with a 2-week history of fever, malaise, and generalized arthralgia. There are no localizing symptoms of infection. Notable findings on examination include a temperature of 38.0°C (100.4°F), needle track marks in his arms, nonblanching vasculitic rash in his legs, and a systolic murmur over the precordium.

His white blood cell count is 15.3 × 10⁹/L (reference range 3.7–11.0), and his C-reactive protein level is 234 mg/dL (normal < 3). Otherwise, results of blood cell counts, liver enzyme tests, renal function tests, urinalysis, and chest radiography are normal.

Two sets of blood cultures are drawn. Transthoracic echocardiography and the antineutrophil cytoplasmic antibody (ANCA) test are requested, as are screening tests for human immunodeficiency virus, hepatitis B, and hepatitis C.

Was the ANCA test indicated in this patient?

ANCAs are autoantibodies against antigens located in the cytoplasmic granules of neutrophils and monocytes. They are associated with small-vessel vasculitides such as granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), eosinophilic granulomatosis with polyangiitis (EGPA), and isolated pauciimmune crescentic glomerulonephritis, all collectively known as ANCA-associated vasculitis (AAV).³⁹

Laboratory methods to detect ANCA include indirect immunofluorescence and antigen-specific enzyme immunoassays. Indirect immunofluorescence only tells us whether or not an antibody that is targeting a cytoplasmic antigen is present. Based on the indirect immunofluorescent pattern, ANCA can be classified as follows:

• Perinuclear or p-ANCA (if the targeted antigen is located just around the nucleus and extends into it)
• Cytoplasmic or c-ANCA (if the targeted antigen is located farther away from the nucleus)
• Atypical ANCA (if the indirect immunofluorescent pattern does not fit with either p-ANCA or c-ANCA).

Indirect immunofluorescence does not give information about the exact antigen that is targeted; this can only be obtained by performing 1 of the antigen-specific immunoassays. The target antigen for c-ANCA is usually proteinase-3 (PR3), while that for p-ANCA could be myeloperoxidase (MPO), cathepsin, lysozyme, lactoferrin, or bactericidal permeability inhibitor. Anti-PR3 is highly specific for GPA, while anti-MPO is usually associated with MPA and EGPA. Less commonly, anti-PR3 may be seen in patients with MPA and anti-MPO in those with GPA. Hence, there is an increasing trend toward classifying ANCA-associated vasculitis into PR3-associated or MPO-associated vasculitis rather than as GPA, MPA, EGPA, or renal-limited vasculitis.⁴⁰

Several audits have shown that the ANCA test is widely misused and requested indiscriminately to rule out vasculitis. This results in a lower positive predictive value, possible harm to patients due to increased false-positive rates, and increased burden on the laboratory.^41–43 At least 2 separate groups have demonstrated that a gating policy that refuses ANCA testing in patients without clinical evidence of systemic vasculitis can reduce the number of inappropriate requests, improve the diagnostic yield, and make it more clinically relevant and cost-effective.^44,45

The clinician should bear in mind that:

Current guidelines recommend using one of the antigen-specific assays for PR3 and MPO as the primary screening method.⁴⁸ Until recently, indirect immunofluorescence was used to screen for ANCA-associated vasculitis, and positive results were confirmed by ELISA to detect ANCAs specific for PR3 and MPO,⁴⁹ but this is no longer recommended because of recent evidence suggesting a large variability between the different indirect immunofluorescent methods and improved diagnostic performance of the antigen-specific assays.

In a large multicenter study by Damoiseaux et al, the specificity with the different antigen-specific immunoassays was 98% to 99% for PR3-ANCA and 96% to 99% for MPO-ANCA.⁵⁰

ANCA-associated vasculitis should not be considered excluded if the PR3 and MPO-ANCA are negative. In the Damoiseaux study, about 11% to 15% of patients with GPA and 8% to 24% of patients with MPA tested negative for both PR3 and MPO-ANCA.⁵⁰

If the ANCA result is negative and clinical suspicion for ANCA-associated vasculitis is high, the clinician may wish to consider requesting another immunoassay method or indirect immunofluorescence. Results of indirect immunofluorescent testing results may be positive in those with a negative immunoassay, and vice versa.

Thus, the ANCA result should always be interpreted in the context of the whole clinical picture.⁵¹ Biopsy should still be considered the gold standard for the diagnosis of ANCA-associated vasculitis. The ANCA titer can help to improve clinical interpretation, because the likelihood of ANCA-associated vasculitis increases with higher levels of PR3 and MPO-ANCA.⁵²

Back to our patient

Our patient’s blood cultures grow methicillin-sensitive Staphylococcus aureus in both sets after 48 hours. Transthoracic echocardiography reveals vegetations around the tricuspid valve, with no evidence of valvular regurgitation. The diagnosis is right-sided infective endocarditis. He is started on appropriate antibiotics.

Tests for human immunodeficiency virus, hepatitis B, and hepatitis C are negative. The ANCA test is positive for MPO-ANCA at 28 IU/mL (normal < 10).

The positive ANCA is thought to be related to the infective endocarditis. His vasculitis is most likely secondary to infective endocarditis and not ANCA-associated vasculitis. The ANCA test need not have been requested in the first place.

 

 

HUMAN LEUKOCYTE ANTIGEN-B27

A 22-year-old man presents to his primary care physician with a 4-month history of gradually worsening low back pain associated with early morning stiffness lasting more than 2 hours. He has no peripheral joint symptoms.

In the last 2 years, he has had 2 separate episodes of uveitis. There is a family history of ankylosing spondylitis in his father. Examination reveals global restriction of lumbar movements but is otherwise unremarkable. Magnetic resonance imaging (MRI) of the lumbar spine and sacroiliac joints is normal.

Should this patient be tested for human leukocyte antigen-B27 (HLA-B27)?

The major histocompatibility complex (MHC) is a gene complex that is present in all animals. It encodes proteins that help with immunologic tolerance. HLA simply refers to the human version of the MHC.⁵³ The HLA gene complex, located on chromosome 6, is categorized into class I, class II, and class III. HLA-B is one of the 3 class I genes. Thus, a positive HLA-B27 result simply means that the particular gene is present in that person.

HLA-B27 is strongly associated with ankylosing spondylitis, also known as axial spondyloarthropathy.⁵⁴ Other genes also contribute to the pathogenesis of ankylosing spondylitis, but HLA-B27 is present in more than 90% of patients with this disease and is by far considered the most important. The association is not as strong for peripheral spondyloarthropathy, with studies reporting a frequency of up to 75% for reactive arthritis and inflammatory bowel disease-associated arthritis, and up to 50% for psoriatic arthritis and uveitis.⁵⁵

About 9% of healthy, asymptomatic individuals may have HLA-B27, so the mere presence of this gene is not evidence of disease.⁵⁶ There may be up to a 20-fold increased risk of ankylosing spondylitis among those who are HLA-B27-positive.⁵⁷

Some HLA genes have many different alleles, each of which is given a number (explaining the number 27 that follows the B). Closely related alleles that differ from one another by only a few amino-acid substitutions are then categorized together, thus accounting for more than 100 subtypes of HLA-B27 (designated from HLA-B*2701 to HLA-B*27106). These subtypes vary in frequency among different racial groups, and the population prevalence of ankylosing spondylitis parallels the frequency of HLA-B27.⁵⁸ The most common subtype seen in white people and American Indians is B*2705. HLA-B27 is rare in blacks, explaining the rarity of ankylosing spondylitis in this population. Further examples include HLA-B*2704, which is seen in Asians, and HLA-B*2702, seen in Mediterranean populations. Not all subtypes of HLA-B27 are associated with disease, and some, like HLA-B*2706, may also be protective.

When should the clinician consider testing for HLA-B27?

Peripheral spondyloarthropathy may present with arthritis, enthesitis (eg, heel pain due to inflammation at the site of insertion of the Achilles tendon or plantar fascia), or dactylitis (“sausage” swelling of the whole finger or toe due to extension of inflammation beyond the margins of the joint). Other clues may include psoriasis, inflammatory bowel disease, history of preceding gastrointestinal or genitourinary infection, family history of similar conditions, and history of recurrent uveitis.

For the initial assessment of patients who have inflammatory back pain, plain radiography of the sacroiliac joints is considered the gold standard.⁵⁹ If plain radiography does not show evidence of sacroiliitis, MRI of the sacroiliac joints should be considered. While plain radiography can reveal only structural changes such as sclerosis, erosions, and ankylosis, MRI is useful to evaluate for early inflammatory changes such as bone marrow edema. Imaging the lumbar spine is not necessary, as the sacroiliac joints are almost invariably involved in axial spondyloarthropathy, and lesions seldom occur in the lumbar spine in isolation.⁶⁰

The diagnosis of ankylosing spondylitis previously relied on confirmatory imaging features, but based on the new International Society classification criteria,^61–63 which can be applied to patients with more than 3 months of back pain and age of onset of symptoms before age 45, patients can be classified as having 1 of the following:

• Radiographic axial spondyloarthropathy, if they have evidence of sacroiliitis on imaging plus 1 other feature of spondyloarthropathy
• Nonradiographic axial spondyloarthropathy, if they have a positive HLA-B27 plus 2 other features of spondyloarthropathy (Table 7).

These new criteria have a sensitivity of 82.9% and specificity of 84.4%.^62,63 The disease burden of radiographic and nonradiographic axial spondyloarthropathy has been shown to be similar, suggesting that they are part of the same disease spectrum. Thus, the HLA-B27 test is useful to make a diagnosis of axial spondyloarthropathy even in the absence of imaging features and could be requested in patients with 2 or more features of spondyloarthropathy. In the absence of imaging features and a negative HLA-B27 result, however, the patient cannot be classified as having axial spondyloarthropathy.

Back to our patient

The absence of radiographic evidence would not exclude axial spondyloarthropathy in our patient. The HLA-B27 test is requested because of the inflammatory back pain and the presence of 2 spondyloarthropathy features (uveitis and the family history) and is reported to be positive. His disease is classified as nonradiographic axial spondyloarthropathy.

He is started on regular naproxen and is referred to a physiotherapist. After 1 month, he reports significant symptomatic improvement. He asks if he can be retested for HLA-B27 to see if it has become negative. We tell him that there is no point in repeating it, as it is a gene and will not disappear.

SUMMARY: CONSIDER THE CLINICAL PICTURE

When approaching a patient suspected of having a rheumatologic disease, a clinician should first consider the clinical presentation and the intended purpose of each test. The tests, in general, might serve several purposes. They might help to:

Increase the likelihood of the diagnosis in question. For example, a positive rheumatoid factor or anticitrullinated peptide antibody can help diagnose rheumatoid arthritis in a patient with early polyarthritis, a positive HLA-B27 can help diagnose ankylosing spondylitis in patients with inflammatory back pain and normal imaging, and a positive ANCA can help diagnose ANCA-associated vasculitis in a patient with glomerulonephritis.

Reduce the likelihood of the diagnosis in question. For example, a negative antinuclear antibody test reduces the likelihood of lupus in a patient with joint pains.

Monitor the condition. For example DNA antibodies can be used to monitor the activity of lupus.

Plan the treatment strategy. For example, one might consider lifelong anticoagulation if antiphospholipid antibodies are persistently positive in a patient with thrombosis.

Prognosticate. For example, positive rheumatoid factor and anticitrullinated peptide antibody increase the risk of erosive rheumatoid arthritis.

If the test was requested in the absence of a clear indication and the result is positive, it is important to bear in mind the potential pitfalls associated with that test and not attach a diagnostic label prematurely. None of the tests can confirm or exclude a condition, so the results should always be interpreted in the context of the whole clinical picture.   

May et al discuss one of the most common laboratory tests we order, the complete blood cell count, and how to interpret and unlock additional information that we often overlook.

Singh et al explain the utility and limitations of assessing hepatic fibrosis in patients with known liver disease using specialized and increasingly available imaging techniques in patients with common diseases that may progress to liver failure.

Using several clinical scenarios, Suresh explores the limitations of serologic testing in patients with a potential “autoimmune” or systemic inflammatory syndrome (which, based on new consultations I see in my rheumatology clinic, seems to be virtually everyone who has experienced pain or fatigue).

The Journal also continues our ongoing series on Smart Testing that has focused on tests and testing strategies that have a strong evidence basis to support or discourage their utilization in specific settings. But in most real-life clinical scenarios, relatively little directly applicable evidence can be brought to bear on our decision process with a specific patient. Hence the ongoing need for each of us to refine our clinical reasoning skills, and to recognize the continuing challenges facing the incorporation of artificial intelligence and algorithmic practice into the management of the individual patient sitting or lying in front of us.

The challenge is to balance input from Watson, “Dr. Google,” our accumulated anecdotal and group experience, and specific data from the patient’s physical examination and provided history. All these sources are valuable, and I believe that how we thoughtfully and purposefully weigh and incorporate this information into practice defines us as the clinicians we are.

May et al discuss one of the most common laboratory tests we order, the complete blood cell count, and how to interpret and unlock additional information that we often overlook.

Singh et al explain the utility and limitations of assessing hepatic fibrosis in patients with known liver disease using specialized and increasingly available imaging techniques in patients with common diseases that may progress to liver failure.

Using several clinical scenarios, Suresh explores the limitations of serologic testing in patients with a potential “autoimmune” or systemic inflammatory syndrome (which, based on new consultations I see in my rheumatology clinic, seems to be virtually everyone who has experienced pain or fatigue).

The Journal also continues our ongoing series on Smart Testing that has focused on tests and testing strategies that have a strong evidence basis to support or discourage their utilization in specific settings. But in most real-life clinical scenarios, relatively little directly applicable evidence can be brought to bear on our decision process with a specific patient. Hence the ongoing need for each of us to refine our clinical reasoning skills, and to recognize the continuing challenges facing the incorporation of artificial intelligence and algorithmic practice into the management of the individual patient sitting or lying in front of us.

The challenge is to balance input from Watson, “Dr. Google,” our accumulated anecdotal and group experience, and specific data from the patient’s physical examination and provided history. All these sources are valuable, and I believe that how we thoughtfully and purposefully weigh and incorporate this information into practice defines us as the clinicians we are.

May et al discuss one of the most common laboratory tests we order, the complete blood cell count, and how to interpret and unlock additional information that we often overlook.

Singh et al explain the utility and limitations of assessing hepatic fibrosis in patients with known liver disease using specialized and increasingly available imaging techniques in patients with common diseases that may progress to liver failure.

Using several clinical scenarios, Suresh explores the limitations of serologic testing in patients with a potential “autoimmune” or systemic inflammatory syndrome (which, based on new consultations I see in my rheumatology clinic, seems to be virtually everyone who has experienced pain or fatigue).

The Journal also continues our ongoing series on Smart Testing that has focused on tests and testing strategies that have a strong evidence basis to support or discourage their utilization in specific settings. But in most real-life clinical scenarios, relatively little directly applicable evidence can be brought to bear on our decision process with a specific patient. Hence the ongoing need for each of us to refine our clinical reasoning skills, and to recognize the continuing challenges facing the incorporation of artificial intelligence and algorithmic practice into the management of the individual patient sitting or lying in front of us.

The challenge is to balance input from Watson, “Dr. Google,” our accumulated anecdotal and group experience, and specific data from the patient’s physical examination and provided history. All these sources are valuable, and I believe that how we thoughtfully and purposefully weigh and incorporate this information into practice defines us as the clinicians we are.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

DIAGNOSIS, TREATMENT, CONTROL

The differential diagnosis of Norwegian scabies includes psoriasis, eczema, contact dermatitis, insect bites, seborrheic dermatitis, lichen planus, systemic infection, palmoplantar keratoderma, and cutaneous lymphoma.²

Treatment involves eradicating the infestation with a topical ointment consisting of permethrin, crotamiton, lindane, benzyl benzoate, and sulfur, applied directly to the skin. However, topical treatments often cannot penetrate the crusted and thickened skin, leading to treatment failure. A dose of oral ivermectin 200 µg/kg on days 1, 2, and 8 is a safe, effective, first-line treatment for Norwegian scabies, rapidly reducing scabies symptoms.³ Adverse effects of oral ivermectin are rare and usually minor.

Norwegian scabies is extremely contagious, spread by close physical contact and sharing of contaminated items such as clothing, bedding, towels, and furniture. Scabies mites can survive off the skin for 48 to 72 hours at room temperature.⁴ Potentially contaminated items should be decontaminated by washing in hot water and drying in a drying machine or by dry cleaning. Body contact with other contaminated items should be avoided for at least 72 hours.

Outbreaks can spread among patients, visitors, and medical staff in institutions such as nursing homes, day care centers, long-term-care facilities, and hospitals.⁵ Early identification facilitates appropriate management and treatment, thereby preventing infection and community-wide scabies outbreaks.          

Acknowledgment: The authors would like to sincerely thank Paul Williams for his editing of the article.

DIAGNOSIS, TREATMENT, CONTROL

The differential diagnosis of Norwegian scabies includes psoriasis, eczema, contact dermatitis, insect bites, seborrheic dermatitis, lichen planus, systemic infection, palmoplantar keratoderma, and cutaneous lymphoma.²

Treatment involves eradicating the infestation with a topical ointment consisting of permethrin, crotamiton, lindane, benzyl benzoate, and sulfur, applied directly to the skin. However, topical treatments often cannot penetrate the crusted and thickened skin, leading to treatment failure. A dose of oral ivermectin 200 µg/kg on days 1, 2, and 8 is a safe, effective, first-line treatment for Norwegian scabies, rapidly reducing scabies symptoms.³ Adverse effects of oral ivermectin are rare and usually minor.

Norwegian scabies is extremely contagious, spread by close physical contact and sharing of contaminated items such as clothing, bedding, towels, and furniture. Scabies mites can survive off the skin for 48 to 72 hours at room temperature.⁴ Potentially contaminated items should be decontaminated by washing in hot water and drying in a drying machine or by dry cleaning. Body contact with other contaminated items should be avoided for at least 72 hours.

Outbreaks can spread among patients, visitors, and medical staff in institutions such as nursing homes, day care centers, long-term-care facilities, and hospitals.⁵ Early identification facilitates appropriate management and treatment, thereby preventing infection and community-wide scabies outbreaks.          

Acknowledgment: The authors would like to sincerely thank Paul Williams for his editing of the article.

DIAGNOSIS, TREATMENT, CONTROL

The differential diagnosis of Norwegian scabies includes psoriasis, eczema, contact dermatitis, insect bites, seborrheic dermatitis, lichen planus, systemic infection, palmoplantar keratoderma, and cutaneous lymphoma.²

Treatment involves eradicating the infestation with a topical ointment consisting of permethrin, crotamiton, lindane, benzyl benzoate, and sulfur, applied directly to the skin. However, topical treatments often cannot penetrate the crusted and thickened skin, leading to treatment failure. A dose of oral ivermectin 200 µg/kg on days 1, 2, and 8 is a safe, effective, first-line treatment for Norwegian scabies, rapidly reducing scabies symptoms.³ Adverse effects of oral ivermectin are rare and usually minor.

Norwegian scabies is extremely contagious, spread by close physical contact and sharing of contaminated items such as clothing, bedding, towels, and furniture. Scabies mites can survive off the skin for 48 to 72 hours at room temperature.⁴ Potentially contaminated items should be decontaminated by washing in hot water and drying in a drying machine or by dry cleaning. Body contact with other contaminated items should be avoided for at least 72 hours.

Outbreaks can spread among patients, visitors, and medical staff in institutions such as nursing homes, day care centers, long-term-care facilities, and hospitals.⁵ Early identification facilitates appropriate management and treatment, thereby preventing infection and community-wide scabies outbreaks.          

Acknowledgment: The authors would like to sincerely thank Paul Williams for his editing of the article.

No. The early repolarization pattern on electrocardiography (ECG) in asymp­tomatic patients is nearly always a benign incidental finding. However, in a patient with a history of idiopathic ventricular fibrillation or a family history of sudden cardiac death, the finding warrants further evaluation.

DEFINING EARLY REPOLARIZATION

The early repolarization pattern may mimic patterns seen in myocardial infarction, pericarditis, ventricular aneurysm, hyperkalemia, and hypothermia,^1,3 and misinterpreting the pattern can lead to unnecessary laboratory testing, imaging, medication use, and hospital admissions. On the other hand, misinterpreting it as benign in the presence of certain features of the history or clinical presentation can delay the diagnosis and treatment of a potentially critical condition.

PREVALENCE AND MECHANISMS

The prevalence of the early repolarization pattern in the general population ranges from 5% to 15%; the wide range reflects differences in the definition, as well as variability in the pattern of early repolarization over time.⁴

The early repolarization pattern is more commonly seen in African American men and in young, physically active individuals.³ In one study, it was observed in 15% of cases of idiopathic ventricular fibrillation and sudden cardiac death, especially in people ages 35 to 45.⁴ While there is evidence of a heritable basis in the general population, a family history of early repolarization is not known to increase the risk of sudden cardiac death.

A proposed mechanism for the early repolarization pattern is an imbalance in the ion channel system, resulting in variable refractoriness of multiple myocardial regions and varying excitability in the myocardium. This can produce a voltage gradient between myocardial regions, which is believed to cause the major hallmarks of the early repolarization pattern, ie, ST-segment elevation and QRS notching or slurring.³

MANAGEMENT

The early repolarization pattern is nearly always a benign incidental finding on ECG, with no specific signs or symptoms attributed to it. High-risk features on ECG are associated with a modest increase in absolute risk of sudden cardiac death and warrant clinical correlation.

In the absence of syncope or family history of sudden cardiac death, early repolarization does not merit further workup.²

In patients with a history of unexplained syncope and a family history of sudden cardiac death, early repolarization should be considered in overall risk stratification.¹ Early repolarization in a patient with previous idiopathic ventricular fibrillation warrants referral for electrophysiologic study and, if indicated, insertion of an implantable cardiac defibrillator for secondary prevention.⁵

No. The early repolarization pattern on electrocardiography (ECG) in asymp­tomatic patients is nearly always a benign incidental finding. However, in a patient with a history of idiopathic ventricular fibrillation or a family history of sudden cardiac death, the finding warrants further evaluation.

DEFINING EARLY REPOLARIZATION

The early repolarization pattern may mimic patterns seen in myocardial infarction, pericarditis, ventricular aneurysm, hyperkalemia, and hypothermia,^1,3 and misinterpreting the pattern can lead to unnecessary laboratory testing, imaging, medication use, and hospital admissions. On the other hand, misinterpreting it as benign in the presence of certain features of the history or clinical presentation can delay the diagnosis and treatment of a potentially critical condition.

PREVALENCE AND MECHANISMS

The prevalence of the early repolarization pattern in the general population ranges from 5% to 15%; the wide range reflects differences in the definition, as well as variability in the pattern of early repolarization over time.⁴

The early repolarization pattern is more commonly seen in African American men and in young, physically active individuals.³ In one study, it was observed in 15% of cases of idiopathic ventricular fibrillation and sudden cardiac death, especially in people ages 35 to 45.⁴ While there is evidence of a heritable basis in the general population, a family history of early repolarization is not known to increase the risk of sudden cardiac death.

A proposed mechanism for the early repolarization pattern is an imbalance in the ion channel system, resulting in variable refractoriness of multiple myocardial regions and varying excitability in the myocardium. This can produce a voltage gradient between myocardial regions, which is believed to cause the major hallmarks of the early repolarization pattern, ie, ST-segment elevation and QRS notching or slurring.³

MANAGEMENT

The early repolarization pattern is nearly always a benign incidental finding on ECG, with no specific signs or symptoms attributed to it. High-risk features on ECG are associated with a modest increase in absolute risk of sudden cardiac death and warrant clinical correlation.

In the absence of syncope or family history of sudden cardiac death, early repolarization does not merit further workup.²

In patients with a history of unexplained syncope and a family history of sudden cardiac death, early repolarization should be considered in overall risk stratification.¹ Early repolarization in a patient with previous idiopathic ventricular fibrillation warrants referral for electrophysiologic study and, if indicated, insertion of an implantable cardiac defibrillator for secondary prevention.⁵

No. The early repolarization pattern on electrocardiography (ECG) in asymp­tomatic patients is nearly always a benign incidental finding. However, in a patient with a history of idiopathic ventricular fibrillation or a family history of sudden cardiac death, the finding warrants further evaluation.

DEFINING EARLY REPOLARIZATION

The early repolarization pattern may mimic patterns seen in myocardial infarction, pericarditis, ventricular aneurysm, hyperkalemia, and hypothermia,^1,3 and misinterpreting the pattern can lead to unnecessary laboratory testing, imaging, medication use, and hospital admissions. On the other hand, misinterpreting it as benign in the presence of certain features of the history or clinical presentation can delay the diagnosis and treatment of a potentially critical condition.

PREVALENCE AND MECHANISMS

The prevalence of the early repolarization pattern in the general population ranges from 5% to 15%; the wide range reflects differences in the definition, as well as variability in the pattern of early repolarization over time.⁴

The early repolarization pattern is more commonly seen in African American men and in young, physically active individuals.³ In one study, it was observed in 15% of cases of idiopathic ventricular fibrillation and sudden cardiac death, especially in people ages 35 to 45.⁴ While there is evidence of a heritable basis in the general population, a family history of early repolarization is not known to increase the risk of sudden cardiac death.

A proposed mechanism for the early repolarization pattern is an imbalance in the ion channel system, resulting in variable refractoriness of multiple myocardial regions and varying excitability in the myocardium. This can produce a voltage gradient between myocardial regions, which is believed to cause the major hallmarks of the early repolarization pattern, ie, ST-segment elevation and QRS notching or slurring.³

MANAGEMENT

The early repolarization pattern is nearly always a benign incidental finding on ECG, with no specific signs or symptoms attributed to it. High-risk features on ECG are associated with a modest increase in absolute risk of sudden cardiac death and warrant clinical correlation.

In the absence of syncope or family history of sudden cardiac death, early repolarization does not merit further workup.²

In patients with a history of unexplained syncope and a family history of sudden cardiac death, early repolarization should be considered in overall risk stratification.¹ Early repolarization in a patient with previous idiopathic ventricular fibrillation warrants referral for electrophysiologic study and, if indicated, insertion of an implantable cardiac defibrillator for secondary prevention.⁵

NOTE: The scenario presented here is partly based on cases reported elsewhere by Martínez-Valles et al¹ and Fernández-Rodríguez et al.²

A 55-year-old man is admitted to the hospital with generalized malaise, paresthesias, and severe hypertension. He says he had experienced agitation along with weakness on exertion 24 hours before presentation to the emergency department, with subsequent onset of paresthesias in his lower extremities and perioral area.

He is already known to have mild chronic obstructive pulmonary disease, with a ratio of forced expiratory volume in 1 second (FEV₁)to forced vital capacity (FVC) of less than 70% and an FEV₁ 85% of predicted. In addition, he was recently diagnosed with diabetes, resistant hypertension requiring maximum doses of 3 agents (a calcium channel blocker, an angiotensin-converting enzyme inhibitor, and a loop diuretic), and hyperlipidemia.

He is a current smoker with a 30-pack-year smoking history. He does not use alcohol. His family history is noncontributory.

ASSESSING ACID-BASE DISORDERS

1. What type of acid-base disorder does this patient have?

• Metabolic acidosis
• Respiratory acidosis
• Metabolic alkalosis
• Respiratory alkalosis

The patient has metabolic alkalosis.

A 5-step approach

1. Acidosis or alkalosis? The patient’s arterial pH is 7.5, which is alkalemic because it is higher than 7.44.

2. Metabolic or respiratory? The primary process in our patient is overwhelmingly metabolic, as his partial pressure of carbon dioxide (Pco₂) is slightly elevated, a direction that would cause acidosis, not alkalosis.

3. The anion gap (the serum sodium concentration minus the sum of the chloride and bicarbonate concentrations) is normal at 8 mmol/L (DRG:HYBRiD-XL Immunoassay and Clinical Chemistry Analyzer, reference range 8–16).

4. Is the disturbance compensated? We have determined that this patient has a metabolic alkalemia; the question now is whether there is any compensation for the primary disturbance.

In metabolic alkalosis, the Pco₂ may increase by approximately 0.6 mm Hg (range 0.5–0.8) above the nominal normal level of 40 mm Hg for each 1-mmol/L increase in bicarbonate above the nominal normal level of 25 mmol/L.⁴ If the patient requires oxygen, the calculation may be unreliable, however, as hypoxemia may have an overriding influence on respiratory drive.

Patients with chronically high Pco₂ levels such as those with chronic obstructive pulmonary disease can become accustomed to high carbon dioxide levels and lose their hyper-
capnic respiratory drive. Giving oxygen supplementation is thought to decrease respiratory drive in these patients, so that they will breathe slower and retain more carbon dioxide. There is some degree of respiratory compensation for metabolic alkalosis that occurs by breathing less, though it is limited overall—even in very alkalotic patients, breathing less results in CO₂ retention, which, by displacing O₂ molecules in the alveoli, will in turn result in hypoxia. The brain then senses the hypoxia and makes one breathe faster, thereby limiting this compensation. 

This patient’s serum bicarbonate level is 40 mmol/L, or 15 mmol/L higher than the nominal normal level. If he is compensating, his Pco₂ should be 40 + (15 × 0.6) = 49 mm Hg, and in fact it is 51 mm Hg, which is within the normal range of expected compensation (47.5–52 mm Hg). Therefore, yes, he is compensating for the primary disturbance.  

5. In metabolic acidosis, is there a delta gap? As our patient has metabolic alkalosis, not acidosis, this question does not apply in this case.

 

 

WHICH TEST TO FIND THE CAUSE?

2. Which is the best test to order next to determine the cause of this patient’s hypokalemic metabolic alkalosis?

• Serum magnesium level
• Spot urine chloride
• Renal ultrasonography
• 24-hour urine collection for sodium, potassium, and chloride

The patient’s loop diuretic is withheld for 12 hours and a spot urine chloride is obtained, which is reported as 44 mmol/L. This high value suggests that a volume-independent hypo­kalemic metabolic alkalosis is present with potassium depletion.

As for the other answer choices:

Serum magnesium. Though hypomagnesemia can cause hypokalemia due to lack of inhibition of renal outer medullary potassium channels and subsequent increased excretion of potassium in the apical tubular membrane, it is not independently associated with acid-base disturbances.⁵

Renal ultrasonography gives information about structural kidney disease but is of limited utility in identifying the cause of hypokalemic metabolic alkalosis.

A 24-hour urine collection is unnecessary in this setting and would ultimately result in delay in diagnosis, as spot urine chloride is a more efficient means of rapidly distinguishing volume-responsive vs volume-independent causes of hypokalemic metabolic alkalosis.⁶

IS HIS HYPERTENSION SECONDARY? IF SO, WHAT IS THE CAUSE?

Several features of this case suggest that the patient’s hypertension is secondary rather than primary. It is of recent onset. The patient’s family history is noncontributory, and his hypertension is resistant to the use of maximum doses of 3 antihypertensive agents.

3. Which of the following causes of secondary hypertension is not commonly associated with hypokalemia and metabolic alkalosis?

• Hyperaldosteronism
• Liddle syndrome
• Cushing syndrome
• Renal parenchymal disease
• Chronic licorice ingestion

Renal parenchymal disease is a cause of resistant hypertension, but it is not characterized by metabolic alkalosis, hypokalemia, and  elevated urine chloride,⁷ while the others listed here—hyperaldosteronism, Liddle syndrome, Cushing syndrome, and chronic licorice ingestion­—are. Other common causes of resistant hypertension without these metabolic abnormalities include obstructive sleep apnea, alcohol abuse, and nonadherence to treatment.

While treatment of hypertension with loop diuretics can result in hypokalemia and metabolic alkalosis due to the effect of these drugs on potassium reabsorption in the loop of Henle, the patient’s hypokalemia persisted after this agent was withdrawn.⁸

Causes of hypokalemic metabolic alkalosis with and without hypertension are further delineated in Figure 1.

Additional diagnostic testing: Plasma renin and plasma aldosterone

At this juncture, the differential diagnosis for this patient’s potassium depletion, metabolic alkalosis, high urine chloride, and hypertension has been narrowed to primary or secondary hyperaldosteronism, surreptitious mineralocorticoid ingestion, Cushing syndrome, licorice ingestion, Liddle syndrome, or one of the 3 hydroxylase deficiencies (11-, 17-, and 21-) (Figure 1).

Although clues in the history, physical examination, and imaging may suggest a specific cause of his abnormal laboratory values, the next step in the diagnostic workup is to measure the plasma renin and aldosterone levels (Table 3).

HYPERALDOSTERONISM

4. Hyperaldosteronism is associated with which of the following patterns of renin and aldosterone values?

• High renin, high aldosterone, normal ratio of plasma aldosterone concentration (PAC) to plasma renin activity (PRA)
• Low renin, low aldosterone, normal PAC–PRA ratio
• Low renin, high aldosterone, high PAC–PRA ratio
• High renin, low aldosterone, low PAC–PRA ratio

The pattern of low renin, high aldosterone, and high PAC–PRA ratio is associated with hyperaldosteronism.

Primary hyperaldosteronism

Primary hyperaldosteronism is one of the most common causes of resistant hypertension and is underappreciated, being diagnosed in up to 20% of patients referred to hypertension specialty clinics.⁷ Potassium levels may be normal, likely contributing to its lack of recognition in this target population.

Primary hyperaldosteronism should be suspected in patients who have a plasma aldosterone PAC–PRA ratio greater than 20 with elevated plasma aldosterone concentrations
(> 15 ng/dL).

Persistently elevated aldosterone levels in the setting of elevated plasma volume is proof that aldosterone secretion is independent of the renin-angiotensin-aldosterone axis, and therefore is autonomous (secondary to adrenal tumor or hyperplasia). Further testing in the form of oral salt loading, saline infusion, or fludrocortisone (a sodium-retaining steroid) administration is thus required to confirm inappropriate, autonomous aldosterone secretion.⁹

After establishing the diagnosis of primary hyperaldosteronism, one should determine the subtype (ie, due to an adrenal carcinoma, unilateral hypersecreting adenoma, or unilateral or bilateral hyperplasia). Further testing includes adrenal computed tomography (CT) to rule out adrenal carcinomas, which are suspected with adenomas larger than 4 cm. Though part of the diagnostic workup, CT as a means of confirmational testing alone does not preclude the possibility of bilateral adrenal hyperplasia in some patients, even in the presence of an adrenal adenoma. For this reason, adrenal venous sampling is required to definitively determine whether the condition is due to a hypersecreting adrenal adenoma or unilateral or bilateral hyperplasia.^9,10

Treatment of primary hyperaldosteronism depends on the subtype of the disease and involves salt restriction in addition to an aldosterone antagonist (spironolactone or eplerenone in the case of bilateral disease) or surgery (unilateral disease).^9,11,12

Secondary hyperaldosteronism

Secondary hyperaldosteronism should be suspected when plasma renin and aldosterone levels are both elevated with a PAC–PRA ratio less than 10.

This pattern is most commonly seen with diuretic use but can also be a consequence of renal artery stenosis or, rarely, a renin-secreting tumor.¹³ Renal artery stenosis is a common finding in patients with hypertension undergoing cardiac catheterization, which is not surprising as more than 90% of such stenoses are atherosclerotic.⁷ Renin-secreting tumors are exceedingly rare, with fewer than 100 cases reported in the literature, and are more common in younger individuals.¹³

Our patient has low-normal aldosterone and plasma renin

On further testing, this patient’s plasma aldosterone level is 2.55 ng/dL (normal < 15 ng/dL), his plasma renin activity is 0.53 ng/mL/hour (normal 0.2–2.8 ng/mL/hour), and his PAC–PRA ratio is therefore 4.81.

The categories discussed thus far have included primary and secondary hyperaldosteronism, which typically do not present with low to normal levels of both renin and aldosterone. Surreptitious mineralocorticoid use could present in this manner, but is unlikely in this patient, whose medications do not include fludrocortisone.

The low-normal values thus lead to consideration of a third category: apparent mineralocorticoid excess. Diseases in this category such as Cushing disease or adrenocorticotropic hormone (ACTH) excess are characterized by increases in corticosteroids so that the potassium depletion, metabolic alkalosis, and hypertension are not a consequence of renin and aldosterone but rather the excess corticosteroids.¹⁴

Causes of apparent mineralocorticoid excess

There are several possible causes of mineralocorticoid excess associated with hypertension and hypokalemic metabolic alkalosis not due to renin and aldosterone.

Chronic licorice ingestion in high volumes is one such cause and is thought to result in inhibition of 11B-hydroxysteroid dehydrogenase or possibly cortisol oxidase by licorice’s active component, glycyrrhetinic acid. This inhibition results in an inability to convert cortisol to cortisone. The cortisol excess binds to mineralocorticoid receptors, and acting like aldosterone, results in hypertension and hypokalemic metabolic alkalosis as well as feedback inhibition of renin and aldosterone levels.¹⁵

Partial hydroxylase deficiencies, though rare, should also be considered as a cause of hypokalemic metabolic alkalosis, hypertension, and, potentially, hirsutism and clitoromegaly in women. They can be diagnosed with elevated levels of 17-ketosteroids and dehydroepiandrosterone sulfate, both of which, in excess, may act on aldosterone receptors in a manner similar to cortisol.¹⁶

Liddle syndrome, a rare autosomal dominant condition, may also present with suppressed levels of both renin and aldosterone. In contrast to the disorders of nonaldosterone mineralocorticoid excess, however, the sodium channel defect in Liddle syndrome is characterized by a primary increase in sodium reabsorption in the collecting tubule and potassium wasting. The resultant volume expansion leads to suppressed renin and aldosterone levels and hypertension with low potassium and elevated bicarbonate concentrations.¹⁷

Liddle syndrome is commonly diagnosed in childhood but may go unrecognized due to occasional absence of hypokalemia at presentation. Potassium-sparing diuretics such as amiloride or triamterene are the mainstays of treatment.¹⁸

Rates of cardiovascular and all-cause mortality are increased in patients with long-term hypercortisolism, even after plasma concentrations of cortisol are normalized.²¹

Figure 2 shows the cascade of the hypothalamic-pituitary-adrenal axis.

 

 

TESTING FOR HYPERCORTISOLISM IN OUR PATIENT

Given the patient’s clinical presentation and laboratory and imaging findings with normal plasma renin and aldosterone levels, a workup for suspected hypercortisolism is initiated.

Initial diagnostic testing for hypercortisolism depends on the degree of clinical suspicion. In those with low probability of the disease, testing should consist of 1 of the following, as a single negative test may be sufficient to rule out the disease:

• 24-hour urinary cortisol levels
• Overnight dexamethasone suppression testing
• Late-night salivary cortisol measurements.

In those with a high index of suspicion, 2 of the aforementioned tests should be performed, as 1 normal result may not be sufficient to exclude the diagnosis.^22,23

A 24-hour urinary cortisol collection and overnight dexamethasone suppression test are obtained. His 24-hour urinary free cortisol level is elevated at 6,600 µg (normal 4–100), and suppression testing with 8 mg of dexamethasone (a form of “high-dose” testing)demonstrates only an 8% decline in serum cortisol levels. Cortisol should generally drop more than 90%.

Morning serum cortisol concentration is less than 5 µg/dL (140 nmol/L) in most patients with Cushing disease (ie, a pituitary tumor), and is usually undetectable in normal subjects. Only about 50% of neuroendocrine ACTH-secreting tumors will suppress with this test.

The patient’s clinical presentation, in conjunction with his diagnostic testing, are thus consistent with Cushing syndrome.

CUSHING SYNDROME

Cushing syndrome is most often exogenous or iatrogenic, ie, a result of supraphysiologic doses of glucocorticoids used to treat a variety of inflammatory, autoimmune, and neoplastic conditions.

Endogenous Cushing syndrome, on the other hand, is rare, with an estimated prevalence of 0.7 to 2.4 cases per million per year. ACTH-dependent causes account for 80% of endogenous Cushing syndrome cases, with ACTH-secreting pituitary adenomas (Cushing disease) accounting for 75% to 80% and ectopic ACTH secretion accounting for 15% to 20%. Less than 1% of cases are due to tumors that produce corticotropin-releasing hormone (CRH).

ACTH-independent Cushing syndrome is diagnosed in 20% of endogenous cases and is most commonly caused by a unilateral adrenal tumor. Rare causes of ACTH-independent disease include adrenal carcinoma, McCune-Albright syndrome, and adrenal hyperplasia.²⁴

The patient’s ACTH is high

To determine whether this is an ACTH-dependent or independent process, the next step is to order an ACTH level. His ACTH level is high at 107 pg/mL (normal < 46 pg/mL), confirming the diagnosis of ACTH-dependent Cushing syndrome.

To find out if this ACTH-dependent process is due to a pituitary adenoma, magnetic resonance imaging (MRI) of the pituitary is obtained but is normal.

Large masses (> 6 mm) strongly suggest Cushing disease, but these tumors are often small and may be missed even with more advanced imaging techniques. Corticotropin-secreting adenomas arising from normal cells in the pituitary retain some sensitivity to glucocorticoid negative feedback and CRH stimulation, and thus high-dose dexamethasone suppression testing in conjunction with CRH stimulation testing can be used to differentiate Cushing disease from ectopic ACTH secretion.^24,25 Both of these tests have poor diagnostic accuracy, however, and thus inferior petrosal sampling remains the gold standard for the diagnosis of Cushing disease.

ACTH-SECRETING TUMORS

5. Cushing syndrome due to ectopic ACTH secretion is most commonly attributed to which of the following tumors?

• Small-cell lung carcinoma
• Pancreatic carcinoma
• Medullary thyroid carcinoma
• Gastrinoma

Severe cases of Cushing syndrome are often attributable to ectopic ACTH secretion due to an underlying malignancy, most commonly small-cell lung carcinoma or neuroendocrine tumors of pulmonary origin. Other causes include pancreatic and thymic neuroendocrine tumors, gastrinomas, and medullary thyroid carcinoma.^25,26

Because most ACTH-producing tumors are intrathoracic, initial imaging in cases of suspected ectopic ACTH secretion should focus on the chest, with CT the usual first choice. Octreotide scintigraphy can also be useful in localizing disease, as many neuroendocrine tumors express somatostatin receptors. Specialized positron-emission tomography scans may also be helpful in tumor identification.²⁴

 

 

TREATMENT OF CUSHING SYNDROME DUE TO ECTOPIC ACTH SECRETION

6. Which of the following is most appropriate medical therapy for suppression of cortisol secretion in Cushing syndrome due to ectopic ACTH secretion?

• Spironolactone
• Dexamethasone
• Somatostatin
• Estrogen
• Ketoconazole

Hyperglycemia, hypokalemia, hypertension, psychiatric disturbances, venous thromboembolism, and systemic infections appear to be common in ectopic ACTH syndrome and often correlate with the degree of hypercortisolemia. Severe Cushing syndrome due to ectopic ACTH secretion is an emergency requiring prompt control of cortisol secretion.

First-line treatments include steroidogenesis inhibitors (ketoconazole, metyrapone, etomidate, mitotane) and glucocorticoid receptor antagonists (mifepristone). High-dose spironolactone and eplerenone can also be used to treat the hypertension and hypokalemia associated with mineralocorticoid receptor stimulation. Definitive treatment involves surgical resection, chemotherapy, or radiotherapy when applicable.^24,25

After confirmation of the diagnosis, the patient is prescribed ketoconazole and spironolactone, with substantial improvement. He subsequently is started on combination chemotherapy and radiation therapy for his small-cell lung carcinoma.

DISCUSSION

The differential diagnosis for hypokalemia is broad and relies on information obtained during the history and physical examination, followed by interpretation of selected laboratory results. Myriad pathologies in diverse organ systems, eg, diarrhea, renal tubular acidosis, and adrenal disease, may be responsible for a low serum potassium. Further categorizing potassium depletion on the basis of an associated acid-base disturbance, such as metabolic alkalosis, allows one to use an algorithmic approach that can identify specific etiologies responsible for both the potassium and the acid-base disturbances.

Using the spot urine chloride in the setting of hypokalemic metabolic alkalosis with or without hypertension can narrow the differential diagnosis and allow additional clinical findings to guide clinical problem-solving and decision-making, even for conditions not commonly encountered in routine medical practice.

Obtaining renin and aldosterone measurements in patients with potassium depletion, metabolic alkalosis, high urine chloride excretion, and hypertension permits further categorization into 3 clinical groups: elevated aldosterone and renin (secondary hyperaldosteronism), elevated aldosterone and low renin (primary hyperaldosteronism), or apparent mineralocorticoid excess wherein neither renin nor aldosterone are responsible for the syndrome.

The patient in our case had apparent mineralocorticoid excess as a consequence of an ACTH-producing small-cell carcinoma.

NOTE: The scenario presented here is partly based on cases reported elsewhere by Martínez-Valles et al¹ and Fernández-Rodríguez et al.²

A 55-year-old man is admitted to the hospital with generalized malaise, paresthesias, and severe hypertension. He says he had experienced agitation along with weakness on exertion 24 hours before presentation to the emergency department, with subsequent onset of paresthesias in his lower extremities and perioral area.

He is already known to have mild chronic obstructive pulmonary disease, with a ratio of forced expiratory volume in 1 second (FEV₁)to forced vital capacity (FVC) of less than 70% and an FEV₁ 85% of predicted. In addition, he was recently diagnosed with diabetes, resistant hypertension requiring maximum doses of 3 agents (a calcium channel blocker, an angiotensin-converting enzyme inhibitor, and a loop diuretic), and hyperlipidemia.

He is a current smoker with a 30-pack-year smoking history. He does not use alcohol. His family history is noncontributory.

ASSESSING ACID-BASE DISORDERS

1. What type of acid-base disorder does this patient have?

• Metabolic acidosis
• Respiratory acidosis
• Metabolic alkalosis
• Respiratory alkalosis

The patient has metabolic alkalosis.

A 5-step approach

1. Acidosis or alkalosis? The patient’s arterial pH is 7.5, which is alkalemic because it is higher than 7.44.

2. Metabolic or respiratory? The primary process in our patient is overwhelmingly metabolic, as his partial pressure of carbon dioxide (Pco₂) is slightly elevated, a direction that would cause acidosis, not alkalosis.

3. The anion gap (the serum sodium concentration minus the sum of the chloride and bicarbonate concentrations) is normal at 8 mmol/L (DRG:HYBRiD-XL Immunoassay and Clinical Chemistry Analyzer, reference range 8–16).

4. Is the disturbance compensated? We have determined that this patient has a metabolic alkalemia; the question now is whether there is any compensation for the primary disturbance.

In metabolic alkalosis, the Pco₂ may increase by approximately 0.6 mm Hg (range 0.5–0.8) above the nominal normal level of 40 mm Hg for each 1-mmol/L increase in bicarbonate above the nominal normal level of 25 mmol/L.⁴ If the patient requires oxygen, the calculation may be unreliable, however, as hypoxemia may have an overriding influence on respiratory drive.

Patients with chronically high Pco₂ levels such as those with chronic obstructive pulmonary disease can become accustomed to high carbon dioxide levels and lose their hyper-
capnic respiratory drive. Giving oxygen supplementation is thought to decrease respiratory drive in these patients, so that they will breathe slower and retain more carbon dioxide. There is some degree of respiratory compensation for metabolic alkalosis that occurs by breathing less, though it is limited overall—even in very alkalotic patients, breathing less results in CO₂ retention, which, by displacing O₂ molecules in the alveoli, will in turn result in hypoxia. The brain then senses the hypoxia and makes one breathe faster, thereby limiting this compensation. 

This patient’s serum bicarbonate level is 40 mmol/L, or 15 mmol/L higher than the nominal normal level. If he is compensating, his Pco₂ should be 40 + (15 × 0.6) = 49 mm Hg, and in fact it is 51 mm Hg, which is within the normal range of expected compensation (47.5–52 mm Hg). Therefore, yes, he is compensating for the primary disturbance.  

5. In metabolic acidosis, is there a delta gap? As our patient has metabolic alkalosis, not acidosis, this question does not apply in this case.

 

 

WHICH TEST TO FIND THE CAUSE?

2. Which is the best test to order next to determine the cause of this patient’s hypokalemic metabolic alkalosis?

• Serum magnesium level
• Spot urine chloride
• Renal ultrasonography
• 24-hour urine collection for sodium, potassium, and chloride

The patient’s loop diuretic is withheld for 12 hours and a spot urine chloride is obtained, which is reported as 44 mmol/L. This high value suggests that a volume-independent hypo­kalemic metabolic alkalosis is present with potassium depletion.

As for the other answer choices:

Serum magnesium. Though hypomagnesemia can cause hypokalemia due to lack of inhibition of renal outer medullary potassium channels and subsequent increased excretion of potassium in the apical tubular membrane, it is not independently associated with acid-base disturbances.⁵

Renal ultrasonography gives information about structural kidney disease but is of limited utility in identifying the cause of hypokalemic metabolic alkalosis.

A 24-hour urine collection is unnecessary in this setting and would ultimately result in delay in diagnosis, as spot urine chloride is a more efficient means of rapidly distinguishing volume-responsive vs volume-independent causes of hypokalemic metabolic alkalosis.⁶

IS HIS HYPERTENSION SECONDARY? IF SO, WHAT IS THE CAUSE?

Several features of this case suggest that the patient’s hypertension is secondary rather than primary. It is of recent onset. The patient’s family history is noncontributory, and his hypertension is resistant to the use of maximum doses of 3 antihypertensive agents.

3. Which of the following causes of secondary hypertension is not commonly associated with hypokalemia and metabolic alkalosis?

• Hyperaldosteronism
• Liddle syndrome
• Cushing syndrome
• Renal parenchymal disease
• Chronic licorice ingestion

Renal parenchymal disease is a cause of resistant hypertension, but it is not characterized by metabolic alkalosis, hypokalemia, and  elevated urine chloride,⁷ while the others listed here—hyperaldosteronism, Liddle syndrome, Cushing syndrome, and chronic licorice ingestion­—are. Other common causes of resistant hypertension without these metabolic abnormalities include obstructive sleep apnea, alcohol abuse, and nonadherence to treatment.

While treatment of hypertension with loop diuretics can result in hypokalemia and metabolic alkalosis due to the effect of these drugs on potassium reabsorption in the loop of Henle, the patient’s hypokalemia persisted after this agent was withdrawn.⁸

Causes of hypokalemic metabolic alkalosis with and without hypertension are further delineated in Figure 1.

Additional diagnostic testing: Plasma renin and plasma aldosterone

At this juncture, the differential diagnosis for this patient’s potassium depletion, metabolic alkalosis, high urine chloride, and hypertension has been narrowed to primary or secondary hyperaldosteronism, surreptitious mineralocorticoid ingestion, Cushing syndrome, licorice ingestion, Liddle syndrome, or one of the 3 hydroxylase deficiencies (11-, 17-, and 21-) (Figure 1).

Although clues in the history, physical examination, and imaging may suggest a specific cause of his abnormal laboratory values, the next step in the diagnostic workup is to measure the plasma renin and aldosterone levels (Table 3).

HYPERALDOSTERONISM

4. Hyperaldosteronism is associated with which of the following patterns of renin and aldosterone values?

• High renin, high aldosterone, normal ratio of plasma aldosterone concentration (PAC) to plasma renin activity (PRA)
• Low renin, low aldosterone, normal PAC–PRA ratio
• Low renin, high aldosterone, high PAC–PRA ratio
• High renin, low aldosterone, low PAC–PRA ratio

The pattern of low renin, high aldosterone, and high PAC–PRA ratio is associated with hyperaldosteronism.

Primary hyperaldosteronism

Primary hyperaldosteronism is one of the most common causes of resistant hypertension and is underappreciated, being diagnosed in up to 20% of patients referred to hypertension specialty clinics.⁷ Potassium levels may be normal, likely contributing to its lack of recognition in this target population.

Primary hyperaldosteronism should be suspected in patients who have a plasma aldosterone PAC–PRA ratio greater than 20 with elevated plasma aldosterone concentrations
(> 15 ng/dL).

Persistently elevated aldosterone levels in the setting of elevated plasma volume is proof that aldosterone secretion is independent of the renin-angiotensin-aldosterone axis, and therefore is autonomous (secondary to adrenal tumor or hyperplasia). Further testing in the form of oral salt loading, saline infusion, or fludrocortisone (a sodium-retaining steroid) administration is thus required to confirm inappropriate, autonomous aldosterone secretion.⁹

After establishing the diagnosis of primary hyperaldosteronism, one should determine the subtype (ie, due to an adrenal carcinoma, unilateral hypersecreting adenoma, or unilateral or bilateral hyperplasia). Further testing includes adrenal computed tomography (CT) to rule out adrenal carcinomas, which are suspected with adenomas larger than 4 cm. Though part of the diagnostic workup, CT as a means of confirmational testing alone does not preclude the possibility of bilateral adrenal hyperplasia in some patients, even in the presence of an adrenal adenoma. For this reason, adrenal venous sampling is required to definitively determine whether the condition is due to a hypersecreting adrenal adenoma or unilateral or bilateral hyperplasia.^9,10

Treatment of primary hyperaldosteronism depends on the subtype of the disease and involves salt restriction in addition to an aldosterone antagonist (spironolactone or eplerenone in the case of bilateral disease) or surgery (unilateral disease).^9,11,12

Secondary hyperaldosteronism

Secondary hyperaldosteronism should be suspected when plasma renin and aldosterone levels are both elevated with a PAC–PRA ratio less than 10.

This pattern is most commonly seen with diuretic use but can also be a consequence of renal artery stenosis or, rarely, a renin-secreting tumor.¹³ Renal artery stenosis is a common finding in patients with hypertension undergoing cardiac catheterization, which is not surprising as more than 90% of such stenoses are atherosclerotic.⁷ Renin-secreting tumors are exceedingly rare, with fewer than 100 cases reported in the literature, and are more common in younger individuals.¹³

Our patient has low-normal aldosterone and plasma renin

On further testing, this patient’s plasma aldosterone level is 2.55 ng/dL (normal < 15 ng/dL), his plasma renin activity is 0.53 ng/mL/hour (normal 0.2–2.8 ng/mL/hour), and his PAC–PRA ratio is therefore 4.81.

The categories discussed thus far have included primary and secondary hyperaldosteronism, which typically do not present with low to normal levels of both renin and aldosterone. Surreptitious mineralocorticoid use could present in this manner, but is unlikely in this patient, whose medications do not include fludrocortisone.

The low-normal values thus lead to consideration of a third category: apparent mineralocorticoid excess. Diseases in this category such as Cushing disease or adrenocorticotropic hormone (ACTH) excess are characterized by increases in corticosteroids so that the potassium depletion, metabolic alkalosis, and hypertension are not a consequence of renin and aldosterone but rather the excess corticosteroids.¹⁴

Causes of apparent mineralocorticoid excess

There are several possible causes of mineralocorticoid excess associated with hypertension and hypokalemic metabolic alkalosis not due to renin and aldosterone.

Chronic licorice ingestion in high volumes is one such cause and is thought to result in inhibition of 11B-hydroxysteroid dehydrogenase or possibly cortisol oxidase by licorice’s active component, glycyrrhetinic acid. This inhibition results in an inability to convert cortisol to cortisone. The cortisol excess binds to mineralocorticoid receptors, and acting like aldosterone, results in hypertension and hypokalemic metabolic alkalosis as well as feedback inhibition of renin and aldosterone levels.¹⁵

Partial hydroxylase deficiencies, though rare, should also be considered as a cause of hypokalemic metabolic alkalosis, hypertension, and, potentially, hirsutism and clitoromegaly in women. They can be diagnosed with elevated levels of 17-ketosteroids and dehydroepiandrosterone sulfate, both of which, in excess, may act on aldosterone receptors in a manner similar to cortisol.¹⁶

Liddle syndrome, a rare autosomal dominant condition, may also present with suppressed levels of both renin and aldosterone. In contrast to the disorders of nonaldosterone mineralocorticoid excess, however, the sodium channel defect in Liddle syndrome is characterized by a primary increase in sodium reabsorption in the collecting tubule and potassium wasting. The resultant volume expansion leads to suppressed renin and aldosterone levels and hypertension with low potassium and elevated bicarbonate concentrations.¹⁷

Liddle syndrome is commonly diagnosed in childhood but may go unrecognized due to occasional absence of hypokalemia at presentation. Potassium-sparing diuretics such as amiloride or triamterene are the mainstays of treatment.¹⁸

Rates of cardiovascular and all-cause mortality are increased in patients with long-term hypercortisolism, even after plasma concentrations of cortisol are normalized.²¹

Figure 2 shows the cascade of the hypothalamic-pituitary-adrenal axis.

 

 

TESTING FOR HYPERCORTISOLISM IN OUR PATIENT

Given the patient’s clinical presentation and laboratory and imaging findings with normal plasma renin and aldosterone levels, a workup for suspected hypercortisolism is initiated.

Initial diagnostic testing for hypercortisolism depends on the degree of clinical suspicion. In those with low probability of the disease, testing should consist of 1 of the following, as a single negative test may be sufficient to rule out the disease:

• 24-hour urinary cortisol levels
• Overnight dexamethasone suppression testing
• Late-night salivary cortisol measurements.

In those with a high index of suspicion, 2 of the aforementioned tests should be performed, as 1 normal result may not be sufficient to exclude the diagnosis.^22,23

A 24-hour urinary cortisol collection and overnight dexamethasone suppression test are obtained. His 24-hour urinary free cortisol level is elevated at 6,600 µg (normal 4–100), and suppression testing with 8 mg of dexamethasone (a form of “high-dose” testing)demonstrates only an 8% decline in serum cortisol levels. Cortisol should generally drop more than 90%.

Morning serum cortisol concentration is less than 5 µg/dL (140 nmol/L) in most patients with Cushing disease (ie, a pituitary tumor), and is usually undetectable in normal subjects. Only about 50% of neuroendocrine ACTH-secreting tumors will suppress with this test.

The patient’s clinical presentation, in conjunction with his diagnostic testing, are thus consistent with Cushing syndrome.

CUSHING SYNDROME

Cushing syndrome is most often exogenous or iatrogenic, ie, a result of supraphysiologic doses of glucocorticoids used to treat a variety of inflammatory, autoimmune, and neoplastic conditions.

Endogenous Cushing syndrome, on the other hand, is rare, with an estimated prevalence of 0.7 to 2.4 cases per million per year. ACTH-dependent causes account for 80% of endogenous Cushing syndrome cases, with ACTH-secreting pituitary adenomas (Cushing disease) accounting for 75% to 80% and ectopic ACTH secretion accounting for 15% to 20%. Less than 1% of cases are due to tumors that produce corticotropin-releasing hormone (CRH).

ACTH-independent Cushing syndrome is diagnosed in 20% of endogenous cases and is most commonly caused by a unilateral adrenal tumor. Rare causes of ACTH-independent disease include adrenal carcinoma, McCune-Albright syndrome, and adrenal hyperplasia.²⁴

The patient’s ACTH is high

To determine whether this is an ACTH-dependent or independent process, the next step is to order an ACTH level. His ACTH level is high at 107 pg/mL (normal < 46 pg/mL), confirming the diagnosis of ACTH-dependent Cushing syndrome.

To find out if this ACTH-dependent process is due to a pituitary adenoma, magnetic resonance imaging (MRI) of the pituitary is obtained but is normal.

Large masses (> 6 mm) strongly suggest Cushing disease, but these tumors are often small and may be missed even with more advanced imaging techniques. Corticotropin-secreting adenomas arising from normal cells in the pituitary retain some sensitivity to glucocorticoid negative feedback and CRH stimulation, and thus high-dose dexamethasone suppression testing in conjunction with CRH stimulation testing can be used to differentiate Cushing disease from ectopic ACTH secretion.^24,25 Both of these tests have poor diagnostic accuracy, however, and thus inferior petrosal sampling remains the gold standard for the diagnosis of Cushing disease.

ACTH-SECRETING TUMORS

5. Cushing syndrome due to ectopic ACTH secretion is most commonly attributed to which of the following tumors?

• Small-cell lung carcinoma
• Pancreatic carcinoma
• Medullary thyroid carcinoma
• Gastrinoma

Severe cases of Cushing syndrome are often attributable to ectopic ACTH secretion due to an underlying malignancy, most commonly small-cell lung carcinoma or neuroendocrine tumors of pulmonary origin. Other causes include pancreatic and thymic neuroendocrine tumors, gastrinomas, and medullary thyroid carcinoma.^25,26

Because most ACTH-producing tumors are intrathoracic, initial imaging in cases of suspected ectopic ACTH secretion should focus on the chest, with CT the usual first choice. Octreotide scintigraphy can also be useful in localizing disease, as many neuroendocrine tumors express somatostatin receptors. Specialized positron-emission tomography scans may also be helpful in tumor identification.²⁴

 

 

TREATMENT OF CUSHING SYNDROME DUE TO ECTOPIC ACTH SECRETION

6. Which of the following is most appropriate medical therapy for suppression of cortisol secretion in Cushing syndrome due to ectopic ACTH secretion?

• Spironolactone
• Dexamethasone
• Somatostatin
• Estrogen
• Ketoconazole

Hyperglycemia, hypokalemia, hypertension, psychiatric disturbances, venous thromboembolism, and systemic infections appear to be common in ectopic ACTH syndrome and often correlate with the degree of hypercortisolemia. Severe Cushing syndrome due to ectopic ACTH secretion is an emergency requiring prompt control of cortisol secretion.

First-line treatments include steroidogenesis inhibitors (ketoconazole, metyrapone, etomidate, mitotane) and glucocorticoid receptor antagonists (mifepristone). High-dose spironolactone and eplerenone can also be used to treat the hypertension and hypokalemia associated with mineralocorticoid receptor stimulation. Definitive treatment involves surgical resection, chemotherapy, or radiotherapy when applicable.^24,25

After confirmation of the diagnosis, the patient is prescribed ketoconazole and spironolactone, with substantial improvement. He subsequently is started on combination chemotherapy and radiation therapy for his small-cell lung carcinoma.

DISCUSSION

The differential diagnosis for hypokalemia is broad and relies on information obtained during the history and physical examination, followed by interpretation of selected laboratory results. Myriad pathologies in diverse organ systems, eg, diarrhea, renal tubular acidosis, and adrenal disease, may be responsible for a low serum potassium. Further categorizing potassium depletion on the basis of an associated acid-base disturbance, such as metabolic alkalosis, allows one to use an algorithmic approach that can identify specific etiologies responsible for both the potassium and the acid-base disturbances.

Using the spot urine chloride in the setting of hypokalemic metabolic alkalosis with or without hypertension can narrow the differential diagnosis and allow additional clinical findings to guide clinical problem-solving and decision-making, even for conditions not commonly encountered in routine medical practice.

Obtaining renin and aldosterone measurements in patients with potassium depletion, metabolic alkalosis, high urine chloride excretion, and hypertension permits further categorization into 3 clinical groups: elevated aldosterone and renin (secondary hyperaldosteronism), elevated aldosterone and low renin (primary hyperaldosteronism), or apparent mineralocorticoid excess wherein neither renin nor aldosterone are responsible for the syndrome.

The patient in our case had apparent mineralocorticoid excess as a consequence of an ACTH-producing small-cell carcinoma.

NOTE: The scenario presented here is partly based on cases reported elsewhere by Martínez-Valles et al¹ and Fernández-Rodríguez et al.²

A 55-year-old man is admitted to the hospital with generalized malaise, paresthesias, and severe hypertension. He says he had experienced agitation along with weakness on exertion 24 hours before presentation to the emergency department, with subsequent onset of paresthesias in his lower extremities and perioral area.

He is already known to have mild chronic obstructive pulmonary disease, with a ratio of forced expiratory volume in 1 second (FEV₁)to forced vital capacity (FVC) of less than 70% and an FEV₁ 85% of predicted. In addition, he was recently diagnosed with diabetes, resistant hypertension requiring maximum doses of 3 agents (a calcium channel blocker, an angiotensin-converting enzyme inhibitor, and a loop diuretic), and hyperlipidemia.

He is a current smoker with a 30-pack-year smoking history. He does not use alcohol. His family history is noncontributory.

ASSESSING ACID-BASE DISORDERS

1. What type of acid-base disorder does this patient have?

• Metabolic acidosis
• Respiratory acidosis
• Metabolic alkalosis
• Respiratory alkalosis

The patient has metabolic alkalosis.

A 5-step approach

1. Acidosis or alkalosis? The patient’s arterial pH is 7.5, which is alkalemic because it is higher than 7.44.

2. Metabolic or respiratory? The primary process in our patient is overwhelmingly metabolic, as his partial pressure of carbon dioxide (Pco₂) is slightly elevated, a direction that would cause acidosis, not alkalosis.

3. The anion gap (the serum sodium concentration minus the sum of the chloride and bicarbonate concentrations) is normal at 8 mmol/L (DRG:HYBRiD-XL Immunoassay and Clinical Chemistry Analyzer, reference range 8–16).

4. Is the disturbance compensated? We have determined that this patient has a metabolic alkalemia; the question now is whether there is any compensation for the primary disturbance.

In metabolic alkalosis, the Pco₂ may increase by approximately 0.6 mm Hg (range 0.5–0.8) above the nominal normal level of 40 mm Hg for each 1-mmol/L increase in bicarbonate above the nominal normal level of 25 mmol/L.⁴ If the patient requires oxygen, the calculation may be unreliable, however, as hypoxemia may have an overriding influence on respiratory drive.

Patients with chronically high Pco₂ levels such as those with chronic obstructive pulmonary disease can become accustomed to high carbon dioxide levels and lose their hyper-
capnic respiratory drive. Giving oxygen supplementation is thought to decrease respiratory drive in these patients, so that they will breathe slower and retain more carbon dioxide. There is some degree of respiratory compensation for metabolic alkalosis that occurs by breathing less, though it is limited overall—even in very alkalotic patients, breathing less results in CO₂ retention, which, by displacing O₂ molecules in the alveoli, will in turn result in hypoxia. The brain then senses the hypoxia and makes one breathe faster, thereby limiting this compensation. 

This patient’s serum bicarbonate level is 40 mmol/L, or 15 mmol/L higher than the nominal normal level. If he is compensating, his Pco₂ should be 40 + (15 × 0.6) = 49 mm Hg, and in fact it is 51 mm Hg, which is within the normal range of expected compensation (47.5–52 mm Hg). Therefore, yes, he is compensating for the primary disturbance.  

5. In metabolic acidosis, is there a delta gap? As our patient has metabolic alkalosis, not acidosis, this question does not apply in this case.

 

 

WHICH TEST TO FIND THE CAUSE?

2. Which is the best test to order next to determine the cause of this patient’s hypokalemic metabolic alkalosis?

• Serum magnesium level
• Spot urine chloride
• Renal ultrasonography
• 24-hour urine collection for sodium, potassium, and chloride

The patient’s loop diuretic is withheld for 12 hours and a spot urine chloride is obtained, which is reported as 44 mmol/L. This high value suggests that a volume-independent hypo­kalemic metabolic alkalosis is present with potassium depletion.

As for the other answer choices:

Serum magnesium. Though hypomagnesemia can cause hypokalemia due to lack of inhibition of renal outer medullary potassium channels and subsequent increased excretion of potassium in the apical tubular membrane, it is not independently associated with acid-base disturbances.⁵

Renal ultrasonography gives information about structural kidney disease but is of limited utility in identifying the cause of hypokalemic metabolic alkalosis.

A 24-hour urine collection is unnecessary in this setting and would ultimately result in delay in diagnosis, as spot urine chloride is a more efficient means of rapidly distinguishing volume-responsive vs volume-independent causes of hypokalemic metabolic alkalosis.⁶

IS HIS HYPERTENSION SECONDARY? IF SO, WHAT IS THE CAUSE?

Several features of this case suggest that the patient’s hypertension is secondary rather than primary. It is of recent onset. The patient’s family history is noncontributory, and his hypertension is resistant to the use of maximum doses of 3 antihypertensive agents.

3. Which of the following causes of secondary hypertension is not commonly associated with hypokalemia and metabolic alkalosis?

• Hyperaldosteronism
• Liddle syndrome
• Cushing syndrome
• Renal parenchymal disease
• Chronic licorice ingestion

Renal parenchymal disease is a cause of resistant hypertension, but it is not characterized by metabolic alkalosis, hypokalemia, and  elevated urine chloride,⁷ while the others listed here—hyperaldosteronism, Liddle syndrome, Cushing syndrome, and chronic licorice ingestion­—are. Other common causes of resistant hypertension without these metabolic abnormalities include obstructive sleep apnea, alcohol abuse, and nonadherence to treatment.

While treatment of hypertension with loop diuretics can result in hypokalemia and metabolic alkalosis due to the effect of these drugs on potassium reabsorption in the loop of Henle, the patient’s hypokalemia persisted after this agent was withdrawn.⁸

Causes of hypokalemic metabolic alkalosis with and without hypertension are further delineated in Figure 1.

Additional diagnostic testing: Plasma renin and plasma aldosterone

At this juncture, the differential diagnosis for this patient’s potassium depletion, metabolic alkalosis, high urine chloride, and hypertension has been narrowed to primary or secondary hyperaldosteronism, surreptitious mineralocorticoid ingestion, Cushing syndrome, licorice ingestion, Liddle syndrome, or one of the 3 hydroxylase deficiencies (11-, 17-, and 21-) (Figure 1).

Although clues in the history, physical examination, and imaging may suggest a specific cause of his abnormal laboratory values, the next step in the diagnostic workup is to measure the plasma renin and aldosterone levels (Table 3).

HYPERALDOSTERONISM

4. Hyperaldosteronism is associated with which of the following patterns of renin and aldosterone values?

• High renin, high aldosterone, normal ratio of plasma aldosterone concentration (PAC) to plasma renin activity (PRA)
• Low renin, low aldosterone, normal PAC–PRA ratio
• Low renin, high aldosterone, high PAC–PRA ratio
• High renin, low aldosterone, low PAC–PRA ratio

The pattern of low renin, high aldosterone, and high PAC–PRA ratio is associated with hyperaldosteronism.

Primary hyperaldosteronism

Primary hyperaldosteronism is one of the most common causes of resistant hypertension and is underappreciated, being diagnosed in up to 20% of patients referred to hypertension specialty clinics.⁷ Potassium levels may be normal, likely contributing to its lack of recognition in this target population.

Primary hyperaldosteronism should be suspected in patients who have a plasma aldosterone PAC–PRA ratio greater than 20 with elevated plasma aldosterone concentrations
(> 15 ng/dL).

Persistently elevated aldosterone levels in the setting of elevated plasma volume is proof that aldosterone secretion is independent of the renin-angiotensin-aldosterone axis, and therefore is autonomous (secondary to adrenal tumor or hyperplasia). Further testing in the form of oral salt loading, saline infusion, or fludrocortisone (a sodium-retaining steroid) administration is thus required to confirm inappropriate, autonomous aldosterone secretion.⁹

After establishing the diagnosis of primary hyperaldosteronism, one should determine the subtype (ie, due to an adrenal carcinoma, unilateral hypersecreting adenoma, or unilateral or bilateral hyperplasia). Further testing includes adrenal computed tomography (CT) to rule out adrenal carcinomas, which are suspected with adenomas larger than 4 cm. Though part of the diagnostic workup, CT as a means of confirmational testing alone does not preclude the possibility of bilateral adrenal hyperplasia in some patients, even in the presence of an adrenal adenoma. For this reason, adrenal venous sampling is required to definitively determine whether the condition is due to a hypersecreting adrenal adenoma or unilateral or bilateral hyperplasia.^9,10

Treatment of primary hyperaldosteronism depends on the subtype of the disease and involves salt restriction in addition to an aldosterone antagonist (spironolactone or eplerenone in the case of bilateral disease) or surgery (unilateral disease).^9,11,12

Secondary hyperaldosteronism

Secondary hyperaldosteronism should be suspected when plasma renin and aldosterone levels are both elevated with a PAC–PRA ratio less than 10.

This pattern is most commonly seen with diuretic use but can also be a consequence of renal artery stenosis or, rarely, a renin-secreting tumor.¹³ Renal artery stenosis is a common finding in patients with hypertension undergoing cardiac catheterization, which is not surprising as more than 90% of such stenoses are atherosclerotic.⁷ Renin-secreting tumors are exceedingly rare, with fewer than 100 cases reported in the literature, and are more common in younger individuals.¹³

Our patient has low-normal aldosterone and plasma renin

On further testing, this patient’s plasma aldosterone level is 2.55 ng/dL (normal < 15 ng/dL), his plasma renin activity is 0.53 ng/mL/hour (normal 0.2–2.8 ng/mL/hour), and his PAC–PRA ratio is therefore 4.81.

The categories discussed thus far have included primary and secondary hyperaldosteronism, which typically do not present with low to normal levels of both renin and aldosterone. Surreptitious mineralocorticoid use could present in this manner, but is unlikely in this patient, whose medications do not include fludrocortisone.

The low-normal values thus lead to consideration of a third category: apparent mineralocorticoid excess. Diseases in this category such as Cushing disease or adrenocorticotropic hormone (ACTH) excess are characterized by increases in corticosteroids so that the potassium depletion, metabolic alkalosis, and hypertension are not a consequence of renin and aldosterone but rather the excess corticosteroids.¹⁴

Causes of apparent mineralocorticoid excess

There are several possible causes of mineralocorticoid excess associated with hypertension and hypokalemic metabolic alkalosis not due to renin and aldosterone.

Chronic licorice ingestion in high volumes is one such cause and is thought to result in inhibition of 11B-hydroxysteroid dehydrogenase or possibly cortisol oxidase by licorice’s active component, glycyrrhetinic acid. This inhibition results in an inability to convert cortisol to cortisone. The cortisol excess binds to mineralocorticoid receptors, and acting like aldosterone, results in hypertension and hypokalemic metabolic alkalosis as well as feedback inhibition of renin and aldosterone levels.¹⁵

Partial hydroxylase deficiencies, though rare, should also be considered as a cause of hypokalemic metabolic alkalosis, hypertension, and, potentially, hirsutism and clitoromegaly in women. They can be diagnosed with elevated levels of 17-ketosteroids and dehydroepiandrosterone sulfate, both of which, in excess, may act on aldosterone receptors in a manner similar to cortisol.¹⁶

Liddle syndrome, a rare autosomal dominant condition, may also present with suppressed levels of both renin and aldosterone. In contrast to the disorders of nonaldosterone mineralocorticoid excess, however, the sodium channel defect in Liddle syndrome is characterized by a primary increase in sodium reabsorption in the collecting tubule and potassium wasting. The resultant volume expansion leads to suppressed renin and aldosterone levels and hypertension with low potassium and elevated bicarbonate concentrations.¹⁷

Liddle syndrome is commonly diagnosed in childhood but may go unrecognized due to occasional absence of hypokalemia at presentation. Potassium-sparing diuretics such as amiloride or triamterene are the mainstays of treatment.¹⁸

Rates of cardiovascular and all-cause mortality are increased in patients with long-term hypercortisolism, even after plasma concentrations of cortisol are normalized.²¹

Figure 2 shows the cascade of the hypothalamic-pituitary-adrenal axis.

 

 

TESTING FOR HYPERCORTISOLISM IN OUR PATIENT

Given the patient’s clinical presentation and laboratory and imaging findings with normal plasma renin and aldosterone levels, a workup for suspected hypercortisolism is initiated.

Initial diagnostic testing for hypercortisolism depends on the degree of clinical suspicion. In those with low probability of the disease, testing should consist of 1 of the following, as a single negative test may be sufficient to rule out the disease:

• 24-hour urinary cortisol levels
• Overnight dexamethasone suppression testing
• Late-night salivary cortisol measurements.

In those with a high index of suspicion, 2 of the aforementioned tests should be performed, as 1 normal result may not be sufficient to exclude the diagnosis.^22,23

A 24-hour urinary cortisol collection and overnight dexamethasone suppression test are obtained. His 24-hour urinary free cortisol level is elevated at 6,600 µg (normal 4–100), and suppression testing with 8 mg of dexamethasone (a form of “high-dose” testing)demonstrates only an 8% decline in serum cortisol levels. Cortisol should generally drop more than 90%.

Morning serum cortisol concentration is less than 5 µg/dL (140 nmol/L) in most patients with Cushing disease (ie, a pituitary tumor), and is usually undetectable in normal subjects. Only about 50% of neuroendocrine ACTH-secreting tumors will suppress with this test.

The patient’s clinical presentation, in conjunction with his diagnostic testing, are thus consistent with Cushing syndrome.

CUSHING SYNDROME

Cushing syndrome is most often exogenous or iatrogenic, ie, a result of supraphysiologic doses of glucocorticoids used to treat a variety of inflammatory, autoimmune, and neoplastic conditions.

Endogenous Cushing syndrome, on the other hand, is rare, with an estimated prevalence of 0.7 to 2.4 cases per million per year. ACTH-dependent causes account for 80% of endogenous Cushing syndrome cases, with ACTH-secreting pituitary adenomas (Cushing disease) accounting for 75% to 80% and ectopic ACTH secretion accounting for 15% to 20%. Less than 1% of cases are due to tumors that produce corticotropin-releasing hormone (CRH).

ACTH-independent Cushing syndrome is diagnosed in 20% of endogenous cases and is most commonly caused by a unilateral adrenal tumor. Rare causes of ACTH-independent disease include adrenal carcinoma, McCune-Albright syndrome, and adrenal hyperplasia.²⁴

The patient’s ACTH is high

To determine whether this is an ACTH-dependent or independent process, the next step is to order an ACTH level. His ACTH level is high at 107 pg/mL (normal < 46 pg/mL), confirming the diagnosis of ACTH-dependent Cushing syndrome.

To find out if this ACTH-dependent process is due to a pituitary adenoma, magnetic resonance imaging (MRI) of the pituitary is obtained but is normal.

Large masses (> 6 mm) strongly suggest Cushing disease, but these tumors are often small and may be missed even with more advanced imaging techniques. Corticotropin-secreting adenomas arising from normal cells in the pituitary retain some sensitivity to glucocorticoid negative feedback and CRH stimulation, and thus high-dose dexamethasone suppression testing in conjunction with CRH stimulation testing can be used to differentiate Cushing disease from ectopic ACTH secretion.^24,25 Both of these tests have poor diagnostic accuracy, however, and thus inferior petrosal sampling remains the gold standard for the diagnosis of Cushing disease.

ACTH-SECRETING TUMORS

5. Cushing syndrome due to ectopic ACTH secretion is most commonly attributed to which of the following tumors?

• Small-cell lung carcinoma
• Pancreatic carcinoma
• Medullary thyroid carcinoma
• Gastrinoma

Severe cases of Cushing syndrome are often attributable to ectopic ACTH secretion due to an underlying malignancy, most commonly small-cell lung carcinoma or neuroendocrine tumors of pulmonary origin. Other causes include pancreatic and thymic neuroendocrine tumors, gastrinomas, and medullary thyroid carcinoma.^25,26

Because most ACTH-producing tumors are intrathoracic, initial imaging in cases of suspected ectopic ACTH secretion should focus on the chest, with CT the usual first choice. Octreotide scintigraphy can also be useful in localizing disease, as many neuroendocrine tumors express somatostatin receptors. Specialized positron-emission tomography scans may also be helpful in tumor identification.²⁴

 

 

TREATMENT OF CUSHING SYNDROME DUE TO ECTOPIC ACTH SECRETION

6. Which of the following is most appropriate medical therapy for suppression of cortisol secretion in Cushing syndrome due to ectopic ACTH secretion?

• Spironolactone
• Dexamethasone
• Somatostatin
• Estrogen
• Ketoconazole

Hyperglycemia, hypokalemia, hypertension, psychiatric disturbances, venous thromboembolism, and systemic infections appear to be common in ectopic ACTH syndrome and often correlate with the degree of hypercortisolemia. Severe Cushing syndrome due to ectopic ACTH secretion is an emergency requiring prompt control of cortisol secretion.

First-line treatments include steroidogenesis inhibitors (ketoconazole, metyrapone, etomidate, mitotane) and glucocorticoid receptor antagonists (mifepristone). High-dose spironolactone and eplerenone can also be used to treat the hypertension and hypokalemia associated with mineralocorticoid receptor stimulation. Definitive treatment involves surgical resection, chemotherapy, or radiotherapy when applicable.^24,25

After confirmation of the diagnosis, the patient is prescribed ketoconazole and spironolactone, with substantial improvement. He subsequently is started on combination chemotherapy and radiation therapy for his small-cell lung carcinoma.

DISCUSSION

The differential diagnosis for hypokalemia is broad and relies on information obtained during the history and physical examination, followed by interpretation of selected laboratory results. Myriad pathologies in diverse organ systems, eg, diarrhea, renal tubular acidosis, and adrenal disease, may be responsible for a low serum potassium. Further categorizing potassium depletion on the basis of an associated acid-base disturbance, such as metabolic alkalosis, allows one to use an algorithmic approach that can identify specific etiologies responsible for both the potassium and the acid-base disturbances.

Using the spot urine chloride in the setting of hypokalemic metabolic alkalosis with or without hypertension can narrow the differential diagnosis and allow additional clinical findings to guide clinical problem-solving and decision-making, even for conditions not commonly encountered in routine medical practice.

Obtaining renin and aldosterone measurements in patients with potassium depletion, metabolic alkalosis, high urine chloride excretion, and hypertension permits further categorization into 3 clinical groups: elevated aldosterone and renin (secondary hyperaldosteronism), elevated aldosterone and low renin (primary hyperaldosteronism), or apparent mineralocorticoid excess wherein neither renin nor aldosterone are responsible for the syndrome.

The patient in our case had apparent mineralocorticoid excess as a consequence of an ACTH-producing small-cell carcinoma.

I have been assured by a very knowing American of my acquaintance in London, that a young healthy child well nursed is at a year old, a most delicious, nourishing, and wholesome food, whether stewed, roasted, baked, or boiled, and I make no doubt that it will equally serve in a fricassee, or ragout.

—Jonathan Swift, A Modest Proposal¹

Large-scale cancer screening programs have the unintended consequences of false-positive results and overdiagnosis, leading to anxiety and overtreatment. The magnitude of these harms continues to be clarified after decades of screening.

Recognizing the trade-off between benefits and harms, the US Preventive Services Task Force (USPSTF) has changed several of its recommendations in recent years. Breast cancer screening recommendations have gone from yearly mammograms starting at age 40 to biennial mammograms starting at age 50 for women at average risk.² Prostate cancer screening is no longer recommended for men age 70 and older, and even for men between 55 and 69, screening is now an individual decision.³

Newer screening programs are targeting high-risk groups rather than the general population, with the aim of increasing the likelihood of benefits and limiting the harms. For example, lung cancer screening is recommended only for current smokers or smokers who have quit within the past 15 years, are between 55 and 80, and have at least a 30 pack-year smoking history.⁴

The movement toward less-frequent screening and screening in a narrower population has evoked strong reactions from advocates of cancer screening. One professor of radiology writes, “It borders on unethical to suggest that the benefit of having your life saved by screening and living another 40 years can be balanced against the ‘harm’ of being recalled for additional mammographic views for what proves to not be a cancer.”⁵ Another notes, “It does not make any sense to throw away the lives saved by screening to avoid over-treating a small number of cancers.”⁶ Both of these authors defend the position that the goal of screening is to minimize cause-specific mortality, irrespective of overdiagnosis, overtreatment, or false-positive results. In other words, harm should have little to no weight in screening recommendations.

Although the debate on cancer screening is moving toward a more balanced discussion of benefits and harms, many patients are still subjected to screening that is more aggressive than the USPSTF recommends, which may be due to an underlying belief that no harm is greater than the benefit of saving a life.

IS MORE-AGGRESSIVE SCREENING THE ANSWER?

One may wonder if more-aggressive screening could prevent deaths that occur despite standard screening. For example, more-frequent screening or use of additional screening methods such as ultrasonography or magnetic resonance imaging has been suggested for patients at high risk of breast cancer.

A MODEST PROPOSAL

If one holds the view that benefits alone should be considered when writing recommendations about screening, the logical conclusion extends beyond screening. We would therefore like to propose a different approach to reducing cancer deaths in the general population:

Why not just remove everybody’s breasts, prostate gland, and colon before cancer arises?

 

 

TO CUT IS TO PREVENT

Currently, we offer prophylactic surgery to patients at high risk of cancer. For example, women with BRCA1/BRCA2 mutations are offered prophylactic mastectomy as one of several options for reducing risk of breast cancer. In 2013, the first case of prophylactic prostatectomy was performed in a man who had a BRCA1/BRCA2 mutation. Total colectomy is considered in men and women who have hereditary nonpolyposis colon cancer, instead of segmental resection, to prevent future cancer.

If prophylactic surgery were extended to the general population, it would greatly reduce the number of cancer deaths. Assuming that removing an organ almost always precludes development of cancer, we may predict that prophylactic mastectomy, prostatectomy, or colectomy would save the lives of most of the patients who are still dying of cancer of these organs. The effectiveness rates would approach, but not reach 100%; such is the case with prophylactic mastectomy.

Consider prostate-specific antigen (PSA) screening. Even using the favorable estimate of the impact of PSA screening, arising from the European Randomised Study of Screening for Prostate Cancer trial, 27 men have to be diagnosed, most undergoing local therapy (the trial was conducted before active surveillance became routine), to avert 1 death from prostate cancer over 13 years.⁹

Contrast this “number needed to diagnose” with the number needed to treat for a strategy of routine prostate removal at age 45 or 50. Given that the lifetime risk of death from prostate cancer approaches 3%, and few cases arise before this age, a prophylactic surgical strategy would avert 1 death per 33 operations. If proponents of screening are willing to accept a number needed to diagnose of 27 over a 13-year interval, they may be willing to consider a number needed to treat of 33 over a lifetime.

There may be harms such as perioperative and postoperative complications. Mastectomy could lead to emotional stress from altered body image. Prostatectomy can have long-term complications such as urinary incontinence and sexual dysfunction. Nevertheless, prophylactic organ removal would save far more lives than current screening practices. It also could decrease mental burden, as patients could rest assured that they will never develop cancer, whereas screening often involves ambiguous test results, follow-up tests, and interventions, increasing patient anxiety.

FINDING THE BALANCE BETWEEN BENEFITS AND HARMS

In truth, we do not really advocate universal mastectomy, prostatectomy, and colectomy to prevent cancer, no more than Swift¹ really wanted to eat the children of Ireland to alleviate poverty and famine in that country.  Rather, we use it as an extreme proposal to highlight the scope and depth of harms that inevitably arise from screening.

If proponents of aggressive screening believe that the goal is to reduce cause-specific mortality as much as possible, giving little weight or consideration to overdiagnosis and overtreatment, then they ought to embrace universal prophylactic surgery as well. Recognition of this logical consequence reminds us that we must make screening recommendations that balance benefits and harms.

Considering an extreme perspective can help in recognizing our bias toward saving lives from cancer and discounting the harms. Aggravating this bias, it is impossible to know whether an individual patient has avoided fatal cancer or undergone unnecessary treatment. Moreover, changing practice is more difficult if it involves rolling back interventions that were once the standard.

Balancing benefits and harms is especially difficult when trying to compare the benefit of preventing a single cancer death against a harm that is less serious but more common. Medicine has always involved difficult trade-offs, as seen in cost-benefit analysis of new treatments or balancing quality of life with quantity of life in a single patient. In addition, each individual may place different values on benefits of screening and avoiding possible harms.

There is an undeniable trade-off with screening, and we must make a conscious decision on where to draw the line when harms outweigh the benefits. We must proceed with caution when subjecting large numbers of men and women to the possibility of psychological burden and decreased quality of life.

Given the growing appreciation of the harms of screening, it is likely that future guidance will continue to move toward less- frequent screening or focusing resources on high-risk populations, where the absolute magnitude of benefit is greater. Cancer screening is also likely to become an individual decision based on personal values and informed decisions.

I have been assured by a very knowing American of my acquaintance in London, that a young healthy child well nursed is at a year old, a most delicious, nourishing, and wholesome food, whether stewed, roasted, baked, or boiled, and I make no doubt that it will equally serve in a fricassee, or ragout.

—Jonathan Swift, A Modest Proposal¹

Large-scale cancer screening programs have the unintended consequences of false-positive results and overdiagnosis, leading to anxiety and overtreatment. The magnitude of these harms continues to be clarified after decades of screening.

Recognizing the trade-off between benefits and harms, the US Preventive Services Task Force (USPSTF) has changed several of its recommendations in recent years. Breast cancer screening recommendations have gone from yearly mammograms starting at age 40 to biennial mammograms starting at age 50 for women at average risk.² Prostate cancer screening is no longer recommended for men age 70 and older, and even for men between 55 and 69, screening is now an individual decision.³

Newer screening programs are targeting high-risk groups rather than the general population, with the aim of increasing the likelihood of benefits and limiting the harms. For example, lung cancer screening is recommended only for current smokers or smokers who have quit within the past 15 years, are between 55 and 80, and have at least a 30 pack-year smoking history.⁴

The movement toward less-frequent screening and screening in a narrower population has evoked strong reactions from advocates of cancer screening. One professor of radiology writes, “It borders on unethical to suggest that the benefit of having your life saved by screening and living another 40 years can be balanced against the ‘harm’ of being recalled for additional mammographic views for what proves to not be a cancer.”⁵ Another notes, “It does not make any sense to throw away the lives saved by screening to avoid over-treating a small number of cancers.”⁶ Both of these authors defend the position that the goal of screening is to minimize cause-specific mortality, irrespective of overdiagnosis, overtreatment, or false-positive results. In other words, harm should have little to no weight in screening recommendations.

Although the debate on cancer screening is moving toward a more balanced discussion of benefits and harms, many patients are still subjected to screening that is more aggressive than the USPSTF recommends, which may be due to an underlying belief that no harm is greater than the benefit of saving a life.

IS MORE-AGGRESSIVE SCREENING THE ANSWER?

One may wonder if more-aggressive screening could prevent deaths that occur despite standard screening. For example, more-frequent screening or use of additional screening methods such as ultrasonography or magnetic resonance imaging has been suggested for patients at high risk of breast cancer.

A MODEST PROPOSAL

If one holds the view that benefits alone should be considered when writing recommendations about screening, the logical conclusion extends beyond screening. We would therefore like to propose a different approach to reducing cancer deaths in the general population:

Why not just remove everybody’s breasts, prostate gland, and colon before cancer arises?

 

 

TO CUT IS TO PREVENT

Currently, we offer prophylactic surgery to patients at high risk of cancer. For example, women with BRCA1/BRCA2 mutations are offered prophylactic mastectomy as one of several options for reducing risk of breast cancer. In 2013, the first case of prophylactic prostatectomy was performed in a man who had a BRCA1/BRCA2 mutation. Total colectomy is considered in men and women who have hereditary nonpolyposis colon cancer, instead of segmental resection, to prevent future cancer.

If prophylactic surgery were extended to the general population, it would greatly reduce the number of cancer deaths. Assuming that removing an organ almost always precludes development of cancer, we may predict that prophylactic mastectomy, prostatectomy, or colectomy would save the lives of most of the patients who are still dying of cancer of these organs. The effectiveness rates would approach, but not reach 100%; such is the case with prophylactic mastectomy.

Consider prostate-specific antigen (PSA) screening. Even using the favorable estimate of the impact of PSA screening, arising from the European Randomised Study of Screening for Prostate Cancer trial, 27 men have to be diagnosed, most undergoing local therapy (the trial was conducted before active surveillance became routine), to avert 1 death from prostate cancer over 13 years.⁹

Contrast this “number needed to diagnose” with the number needed to treat for a strategy of routine prostate removal at age 45 or 50. Given that the lifetime risk of death from prostate cancer approaches 3%, and few cases arise before this age, a prophylactic surgical strategy would avert 1 death per 33 operations. If proponents of screening are willing to accept a number needed to diagnose of 27 over a 13-year interval, they may be willing to consider a number needed to treat of 33 over a lifetime.

There may be harms such as perioperative and postoperative complications. Mastectomy could lead to emotional stress from altered body image. Prostatectomy can have long-term complications such as urinary incontinence and sexual dysfunction. Nevertheless, prophylactic organ removal would save far more lives than current screening practices. It also could decrease mental burden, as patients could rest assured that they will never develop cancer, whereas screening often involves ambiguous test results, follow-up tests, and interventions, increasing patient anxiety.

FINDING THE BALANCE BETWEEN BENEFITS AND HARMS

In truth, we do not really advocate universal mastectomy, prostatectomy, and colectomy to prevent cancer, no more than Swift¹ really wanted to eat the children of Ireland to alleviate poverty and famine in that country.  Rather, we use it as an extreme proposal to highlight the scope and depth of harms that inevitably arise from screening.

If proponents of aggressive screening believe that the goal is to reduce cause-specific mortality as much as possible, giving little weight or consideration to overdiagnosis and overtreatment, then they ought to embrace universal prophylactic surgery as well. Recognition of this logical consequence reminds us that we must make screening recommendations that balance benefits and harms.

Considering an extreme perspective can help in recognizing our bias toward saving lives from cancer and discounting the harms. Aggravating this bias, it is impossible to know whether an individual patient has avoided fatal cancer or undergone unnecessary treatment. Moreover, changing practice is more difficult if it involves rolling back interventions that were once the standard.

Balancing benefits and harms is especially difficult when trying to compare the benefit of preventing a single cancer death against a harm that is less serious but more common. Medicine has always involved difficult trade-offs, as seen in cost-benefit analysis of new treatments or balancing quality of life with quantity of life in a single patient. In addition, each individual may place different values on benefits of screening and avoiding possible harms.

There is an undeniable trade-off with screening, and we must make a conscious decision on where to draw the line when harms outweigh the benefits. We must proceed with caution when subjecting large numbers of men and women to the possibility of psychological burden and decreased quality of life.

Given the growing appreciation of the harms of screening, it is likely that future guidance will continue to move toward less- frequent screening or focusing resources on high-risk populations, where the absolute magnitude of benefit is greater. Cancer screening is also likely to become an individual decision based on personal values and informed decisions.

I have been assured by a very knowing American of my acquaintance in London, that a young healthy child well nursed is at a year old, a most delicious, nourishing, and wholesome food, whether stewed, roasted, baked, or boiled, and I make no doubt that it will equally serve in a fricassee, or ragout.

—Jonathan Swift, A Modest Proposal¹

Large-scale cancer screening programs have the unintended consequences of false-positive results and overdiagnosis, leading to anxiety and overtreatment. The magnitude of these harms continues to be clarified after decades of screening.

Recognizing the trade-off between benefits and harms, the US Preventive Services Task Force (USPSTF) has changed several of its recommendations in recent years. Breast cancer screening recommendations have gone from yearly mammograms starting at age 40 to biennial mammograms starting at age 50 for women at average risk.² Prostate cancer screening is no longer recommended for men age 70 and older, and even for men between 55 and 69, screening is now an individual decision.³

Newer screening programs are targeting high-risk groups rather than the general population, with the aim of increasing the likelihood of benefits and limiting the harms. For example, lung cancer screening is recommended only for current smokers or smokers who have quit within the past 15 years, are between 55 and 80, and have at least a 30 pack-year smoking history.⁴

The movement toward less-frequent screening and screening in a narrower population has evoked strong reactions from advocates of cancer screening. One professor of radiology writes, “It borders on unethical to suggest that the benefit of having your life saved by screening and living another 40 years can be balanced against the ‘harm’ of being recalled for additional mammographic views for what proves to not be a cancer.”⁵ Another notes, “It does not make any sense to throw away the lives saved by screening to avoid over-treating a small number of cancers.”⁶ Both of these authors defend the position that the goal of screening is to minimize cause-specific mortality, irrespective of overdiagnosis, overtreatment, or false-positive results. In other words, harm should have little to no weight in screening recommendations.

Although the debate on cancer screening is moving toward a more balanced discussion of benefits and harms, many patients are still subjected to screening that is more aggressive than the USPSTF recommends, which may be due to an underlying belief that no harm is greater than the benefit of saving a life.

IS MORE-AGGRESSIVE SCREENING THE ANSWER?

One may wonder if more-aggressive screening could prevent deaths that occur despite standard screening. For example, more-frequent screening or use of additional screening methods such as ultrasonography or magnetic resonance imaging has been suggested for patients at high risk of breast cancer.

A MODEST PROPOSAL

If one holds the view that benefits alone should be considered when writing recommendations about screening, the logical conclusion extends beyond screening. We would therefore like to propose a different approach to reducing cancer deaths in the general population:

Why not just remove everybody’s breasts, prostate gland, and colon before cancer arises?

 

 

TO CUT IS TO PREVENT

Currently, we offer prophylactic surgery to patients at high risk of cancer. For example, women with BRCA1/BRCA2 mutations are offered prophylactic mastectomy as one of several options for reducing risk of breast cancer. In 2013, the first case of prophylactic prostatectomy was performed in a man who had a BRCA1/BRCA2 mutation. Total colectomy is considered in men and women who have hereditary nonpolyposis colon cancer, instead of segmental resection, to prevent future cancer.

If prophylactic surgery were extended to the general population, it would greatly reduce the number of cancer deaths. Assuming that removing an organ almost always precludes development of cancer, we may predict that prophylactic mastectomy, prostatectomy, or colectomy would save the lives of most of the patients who are still dying of cancer of these organs. The effectiveness rates would approach, but not reach 100%; such is the case with prophylactic mastectomy.

Consider prostate-specific antigen (PSA) screening. Even using the favorable estimate of the impact of PSA screening, arising from the European Randomised Study of Screening for Prostate Cancer trial, 27 men have to be diagnosed, most undergoing local therapy (the trial was conducted before active surveillance became routine), to avert 1 death from prostate cancer over 13 years.⁹

Contrast this “number needed to diagnose” with the number needed to treat for a strategy of routine prostate removal at age 45 or 50. Given that the lifetime risk of death from prostate cancer approaches 3%, and few cases arise before this age, a prophylactic surgical strategy would avert 1 death per 33 operations. If proponents of screening are willing to accept a number needed to diagnose of 27 over a 13-year interval, they may be willing to consider a number needed to treat of 33 over a lifetime.

There may be harms such as perioperative and postoperative complications. Mastectomy could lead to emotional stress from altered body image. Prostatectomy can have long-term complications such as urinary incontinence and sexual dysfunction. Nevertheless, prophylactic organ removal would save far more lives than current screening practices. It also could decrease mental burden, as patients could rest assured that they will never develop cancer, whereas screening often involves ambiguous test results, follow-up tests, and interventions, increasing patient anxiety.

FINDING THE BALANCE BETWEEN BENEFITS AND HARMS

In truth, we do not really advocate universal mastectomy, prostatectomy, and colectomy to prevent cancer, no more than Swift¹ really wanted to eat the children of Ireland to alleviate poverty and famine in that country.  Rather, we use it as an extreme proposal to highlight the scope and depth of harms that inevitably arise from screening.

If proponents of aggressive screening believe that the goal is to reduce cause-specific mortality as much as possible, giving little weight or consideration to overdiagnosis and overtreatment, then they ought to embrace universal prophylactic surgery as well. Recognition of this logical consequence reminds us that we must make screening recommendations that balance benefits and harms.

Considering an extreme perspective can help in recognizing our bias toward saving lives from cancer and discounting the harms. Aggravating this bias, it is impossible to know whether an individual patient has avoided fatal cancer or undergone unnecessary treatment. Moreover, changing practice is more difficult if it involves rolling back interventions that were once the standard.

Balancing benefits and harms is especially difficult when trying to compare the benefit of preventing a single cancer death against a harm that is less serious but more common. Medicine has always involved difficult trade-offs, as seen in cost-benefit analysis of new treatments or balancing quality of life with quantity of life in a single patient. In addition, each individual may place different values on benefits of screening and avoiding possible harms.

There is an undeniable trade-off with screening, and we must make a conscious decision on where to draw the line when harms outweigh the benefits. We must proceed with caution when subjecting large numbers of men and women to the possibility of psychological burden and decreased quality of life.

Given the growing appreciation of the harms of screening, it is likely that future guidance will continue to move toward less- frequent screening or focusing resources on high-risk populations, where the absolute magnitude of benefit is greater. Cancer screening is also likely to become an individual decision based on personal values and informed decisions.

In Aleyadeh W, Hutt-Centeno E, Ahmed HM, Shah NP. Hypertension guidelines: treat patients, not numbers. Cleve Clin J Med 2019; 86(1):47–56. doi:10.3949/ccjm.86a.18027, on page 50, the following statement was incorrect: “In 2017, the American College of Physicians (ACP) and the American Academy of Family Physicians (AAFP) recommended a relaxed systolic blood pressure target, ie, below 150 mm Hg, for adults over age 60, but a tighter goal of less than 130 mm Hg for the same age group if they have transient ischemic attack, stroke, or high cardiovascular risk.9” In fact, the ACP and AAFP recommended a tighter goal of less than 140 mm Hg for this higher-risk group. This has been corrected online.

In Aleyadeh W, Hutt-Centeno E, Ahmed HM, Shah NP. Hypertension guidelines: treat patients, not numbers. Cleve Clin J Med 2019; 86(1):47–56. doi:10.3949/ccjm.86a.18027, on page 50, the following statement was incorrect: “In 2017, the American College of Physicians (ACP) and the American Academy of Family Physicians (AAFP) recommended a relaxed systolic blood pressure target, ie, below 150 mm Hg, for adults over age 60, but a tighter goal of less than 130 mm Hg for the same age group if they have transient ischemic attack, stroke, or high cardiovascular risk.9” In fact, the ACP and AAFP recommended a tighter goal of less than 140 mm Hg for this higher-risk group. This has been corrected online.

In Aleyadeh W, Hutt-Centeno E, Ahmed HM, Shah NP. Hypertension guidelines: treat patients, not numbers. Cleve Clin J Med 2019; 86(1):47–56. doi:10.3949/ccjm.86a.18027, on page 50, the following statement was incorrect: “In 2017, the American College of Physicians (ACP) and the American Academy of Family Physicians (AAFP) recommended a relaxed systolic blood pressure target, ie, below 150 mm Hg, for adults over age 60, but a tighter goal of less than 130 mm Hg for the same age group if they have transient ischemic attack, stroke, or high cardiovascular risk.9” In fact, the ACP and AAFP recommended a tighter goal of less than 140 mm Hg for this higher-risk group. This has been corrected online.

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.