Mauro Pittiruti, MD

Short Peripheral Intravenous Catheter

Approximately 200 million peripheral intravenous catheters (PIVs) are placed annually in the United States, making them the most common intravenous catheter.[3] PIVs are short devices, 3 to 6 cm in length, that enter and terminate in peripheral veins (Figure 1A). Placement is recommended in forearm veins rather than those of the hand, wrist, or upper arm, as forearm sites are less prone to occlusion, accidental removal, and phlebitis.[4] Additionally, placement in hand veins impedes activities of daily living (eg, hand washing) and is not preferred by patients.[5] PIV size ranges from 24 gauge (smallest) to 14 gauge (largest); larger catheters are often reserved for fluid resuscitation or blood transfusion as they accommodate greater flow and limit hemolysis. To decrease risk of phlebitis and thrombosis, the shortest catheter and smallest diameter should be used. However, unless adequately secured, smaller diameter catheters are also associated with greater rates of accidental removal.[4, 5]

Figure 1

By definition, PIVs are short‐term devices. The CDC currently recommends removal and replacement of these devices no more frequently than every 72 to 96 hours in adults. However, a recent randomized controlled trial found that replacing PIVs when clinically indicated (eg, device failure, phlebitis) rather than on a routine schedule added 30 hours to their lifespan without an increase in complications.[6] A systematic review by the Cochrane Collaboration echoes these findings.[3] These data have thus been incorporated into recommendations from the Infusion Nurses Society (INS) and the National Health Service in the United Kingdom.[5, 7] In hospitalized patients, this approach is relevant, as it preserves venous access sites, maximizes device dwell, and limits additional PIV insertions. In turn, these differences may reduce the need for invasive VADs such as PICCs. Furthermore, the projected 5‐year savings from implementation of clinically indicated PIV removal policies is US$300 million and 1 million health‐worker hours in the United States alone.[4]

PIVs offer many advantages. First, they are minimally invasive and require little training to insert. Second, they can be used for diverse indications in patients requiring short‐term (1 week) venous access. Third, PIVs do not require imaging to ensure correct placement; palpation of superficial veins is sufficient. Fourth, PIVs exhibit a risk of bloodstream infection that is about 40‐fold lower than more invasive, longer‐dwelling VADs[8] (0.06 bacteremia per 1000 catheter‐days).

Despite these advantages, PIVs also have important drawbacks. First, a quarter of all PIVs fail through occlusion or accidental dislodgement.[4] Infiltration, extravasation, and hematoma formation are important adverse events that may occur in such cases. Second, thrombophlebitis (pain and redness at the insertion site) is frequent, and may require device removal, especially in patients with catheters 20 guage.[9] Third, despite their relative safety, PIVs can cause localized or hematogenous infection. Septic thrombophlebitis (superficial thrombosis and bloodstream infection) and catheter‐related bloodstream infection, though rare, have been reported with PIVs and may lead to serious complications.[8, 10] In fact, some suggest that the overall burden of bloodstream infection risk posed by PIVs may be similar to that of CICCs given the substantially greater number of devices used and greater number of device days.[8]

PIVs and other peripheral VADs are not suitable for infusion of vesicants or irritants, which require larger, central veins for delivery. Vesicants (drugs that cause blistering on infusion) include chemotherapeutic agents (eg, dactinomycin, paclitaxel) and commonly used nonchemotherapeutical agents (eg, diazepam, piperacillin, vancomycin, esmolol, or total parenteral nutrition [TPN]).[11] Irritants (phlebitogenic drugs) cause short‐term inflammation and pain, and thus should not be peripherally infused for prolonged durations. Common irritants in the hospital setting include acyclovir, dobutamine, penicillin, and potassium chloride.

Of note, about one‐quarter of PIV insertions fail owing to difficult intravenous access.[12] Ultrasound‐guided peripheral intravenous (USGPIV) catheter placement is emerging as a technique to provide peripheral access for such patients to avoid placement of central venous access devices. Novel, longer devices (>8 cm) with built‐in guide wires have been developed to increase placement success of USGPIVs. These new designs provide easier access into deeper arm veins (brachial or basilic) not otherwise accessible by short PIVs. Although studies comparing the efficacy of USGPIV devices to other VADs are limited, a recent systematic review showed that time to successful cannulation was shorter, and fewer attempts were required to place USGPIVs compared to PIVs.[13] A recent study in France found that USGPIVs met the infusion needs of patients with difficult veins with minimal increase in complications.[14] Despite these encouraging data, future studies are needed to better evaluate this technology.

Midline Catheter

A midline is a VAD that is between 7.5 to 25 cm in length and is typically inserted into veins above the antecubital fossa. The catheter tip resides in a peripheral upper arm vein, often the basilic or cephalic vein, terminating just short of the subclavian vein (Figure 1B). Midline‐like devices were first developed in the 1950s and were initially used as an alternative to PIVs because they were thought to allow longer dwell times.[15] However, because they were originally constructed with a fairly rigid material, infiltration, mechanical phlebitis, and inflammation were common and tempered enthusiasm for their use.[15, 16] Newer midline devices obviate many of these problems and are inserted by ultrasound guidance and modified Seldinger technique.[17] Despite these advances, data regarding comparative efficacy are limited.

Midlines offer longer dwell times than standard PIVs owing to termination in the larger diameter basilic and brachial veins of the arm. Additionally, owing to their length, midlines are less prone to dislodgement. As they are inserted with greater antisepsis than PIVs and better secured to the skin, they are more durable than PIVs.[5, 9, 18] Current INS standards recommend use of midlines for 1 to 4 weeks.[5] Because they terminate in a peripheral vein, medications and infusions compatible with midlines are identical to those that are infused through a PIV. Thus, TPN, vesicants or irritants, or drugs that feature a pH <5 or pH >9, or >500 mOsm should not be infused through a midline.[15] New evidence suggests that diluted solutions of vancomycin (usually pH <5) may be safe to infuse for short durations (<6 days) through a midline, and that concentration rather than pH may be more important in this regard.[19] Although it is possible that the use of midlines may extend to agents typically not deemed peripheral access compatible, limited evidence exists to support such a strategy at this time.

Midlines offer several advantages. First, because blood flow is greater in the more proximal veins of the arm, midlines can accommodate infusions at rates of 100 to 150 mL/min compared to 20 to 40 mL/min in smaller peripheral veins. Higher flow rates offer greater hemodilution (dilution of the infusion with blood), decreasing the likelihood of phlebitis and infiltration.[20] Second, midlines do not require x‐ray verification of tip placement; thus, their use is often favored in resource‐depleted settings such as skilled nursing facilities. Third, midlines offer longer dwell times than peripheral intravenous catheters and can thus serve as bridge devices for short‐term intravenous antibiotics or peripheral‐compatible infusions in an outpatient setting. Available evidence suggests that midlines are associated with low rates of bloodstream infection (0.30.8 per 1000 catheter‐days).[17] The most frequent complications include phlebitis (4.2%) and occlusion (3.3%).[20] Given these favorable statistics, midlines may offer a good alternative to PIVs in select patients who require peripheral infusions of intermediate duration.

Intraosseous Vascular Access

Intraosseous (IO) devices access the vascular system by piercing cortical bone. These devices provide access to the intramedullary cavity and venous plexi of long bones such as the tibia, femur, or humerus. Several insertion devices are now commercially available and have enhanced the ease and safety of IO placement. Using these newer devices, IO access may be obtained in 1 to 2 minutes with minimal training. By comparison, a central venous catheter often requires 10 to 15 minutes to insert with substantial training efforts for providers.[21, 22, 23]

IO devices thus offer several advantages. First, given the rapidity with which they can be inserted, they are often preferred in emergency settings (eg, trauma). Second, these devices are versatile and can accommodate both central and peripheral infusates.[24] Third, a recent meta‐analysis found that IOs have a low complication rate of 0.8%, with extravasation of infusate through the cortical entry site being the most common adverse event.[21] Of note, this study also reported zero local or distal infectious complications, a finding that may relate to the shorter dwell of these devices.[21] Some animal studies suggest that fat embolism from bone may occur at high rates with IO VADs.[25] However, death or significant morbidity from fat emboli in humans following IO access has not been described. Whether such emboli occur or are clinically significant in the context of IO devices remains unclear at this time.[21]

Peripherally Inserted Central Catheter

PICCs are inserted into peripheral veins of the upper arm (eg, brachial, basilica, or cephalic vein) and advanced such that the tip resides at the cavoatrial junction (Figure 1C). PICCs offer prolonged dwell times and are thus indicated when patients require venous access for weeks or months.[26] Additionally, they can accommodate a variety of infusates and are safer to insert than CICCs, given placement in peripheral veins of the arm rather than central veins of the chest/neck. Thus, insertion complications such as pneumothorax, hemothorax, or significant bleeding are rare with PICCs. In fact, a recent study reported that PICC insertion by hospitalists was associated with low rates of insertion or infectious complications.[27]

However, like CICCs, PICCs are associated with central lineassociated bloodstream infection (CLABSI), a serious complication known to prolong length of hospital stay, increase costs, and carry a 12% to 25% associated mortality.[28, 29] In the United States alone, over 250,000 CLASBI cases occur per year drawing considerable attention from the CDC and Joint Commission, who now mandate reporting and nonpayment for hospital‐acquired CLABSI.[30, 31, 32] A recent systematic review and meta‐analysis found that PICCs are associated with a substantial risk of CLABSI in hospitalized patients.[33] Importantly, no difference in CLABSI rates between PICCs and CICCs in hospitalized patients was evident in this meta‐analysis. Therefore, current guidelines specifically recommend against use of PICCs over CICCs as a strategy to reduce CLABSI.[34] Additionally, PICCs are associated with 2.5‐fold greater risk of deep vein thrombosis (DVT) compared to CICCs; thus, they should be used with caution in patients with cancer or those with underlying hypercoagulable states.

Of particular import to hospitalists is the fact that PICC placement is contraindicated in patients with stage IIIB or greater chronic kidney disease (CKD). In such patients, sequelae of PICC use, such as phlebitis or central vein stenosis, can be devastating in patients with CKD.[35] In a recent study, prior PICC placement was the strongest predictor of subsequent arteriovenous graft failure.[36] For this reason, Choosing Wisely recommendations call for avoidance of PICCs in such patients.[37]

Centrally Inserted Central Catheter

CICCs are CVADs placed by puncture and cannulation of the internal jugular, subclavian, brachiocephalic, or femoral veins (Figure 1D) and compose the vast majority of VADs placed in ICU settings.[38, 39] Central termination of CICCs allows for a variety of infusions, including irritants, vesicants, and vasopressors, as well as blood withdrawal and hemodynamic monitoring. CICCs are typically used for 7 to 14 days, but may remain for longer durations if they remain complication free and clinically necessary.[40] A key advantage of CICCs is that they can be placed in emergent settings to facilitate quick access for rapid infusion or hemodynamic monitoring. In particular, CICCs are inserted in the femoral vein and may be useful in emergency settings. However, owing to risk of infection and inability to monitor central pressures, these femoral devices should be replaced with a proper CICC or PICC when possible. Importantly, although CICCs are almost exclusively used in intensive or emergency care, PICCs may also be considered in such settings.[41, 42] CICCs usually have multiple lumens and often serve several simultaneous functions such as both infusions and hemodynamic monitoring.

Despite their benefits, CICCs have several disadvantages. First, insertion requires an experienced clinician and has historically been a task limited to physicians. However, this is changing rapidly (especially in Europe and Australia) where specially trained nurses are assuming responsibility for CICC placement.[43] Second, these devices are historically more likely to be associated with CLABSI, with estimates of infection rates varying between 2 and 5 infections per 1000 catheter‐days.[44] Third, CICCs pose a significant DVT risk, with rates around 22 DVTs per 1000 catheter‐days.[45] However, compared to PICCs, the DVT risk appears lower, and CICC use may be preferable in patients at high risk of DVT, such as critically ill or cancer populations.[46] An important note to prevent CICC insertion complications relates to use of ultrasound, a practice that has been associated with decreased accidental arterial puncture and hematoma formation. The role of ultrasound guidance with PICCs as well as implications for thrombotic and infectious events remains less characterized at this point.[47]

Tunneled Central Venous Access Devices

Tunneled devices (either CICCs or PICCs) are characterized by the fact that the insertion site on the skin and site of ultimate venipuncture are physically separated (Figure 1E). Tunneling limits bacterial entry from the extraluminal aspect of the CVAD to the bloodstream. For example, internal jugular veins are often ideal sites of puncture but inappropriate sites for catheter placement, as providing care to this area is challenging and may increase risk of infection.[34] Tunneling to the infraclavicular area provides a better option, as it provides an exit site that can be adequately cared for. Importantly, any CVAD (PICCs or CICCs) can be tunneled. Additionally, tunneled CICCs may be used in patients with chronic or impending renal failure where PICCs are contraindicated because entry into dialysis‐relevant vessels is to be avoided.[48] Such devices also allow regular blood sampling in patients who require frequent testing but have limited peripheral access, such as those with hematological malignancies. Additionally, tunneled catheters are more comfortable for patients and viewed as being more socially acceptable than nontunneled devices. However, the more invasive and permanent nature of these devices often requires deliberation prior to insertion.

Of note, tunneled devices and ports may be used as long‐term (>3 months to years) VADs. As our focus in this review is short‐term devices, we will not expand the discussion of these devices as they are almost always used for prolonged durations.[7]

Short Peripheral Intravenous Catheter

Approximately 200 million peripheral intravenous catheters (PIVs) are placed annually in the United States, making them the most common intravenous catheter.[3] PIVs are short devices, 3 to 6 cm in length, that enter and terminate in peripheral veins (Figure 1A). Placement is recommended in forearm veins rather than those of the hand, wrist, or upper arm, as forearm sites are less prone to occlusion, accidental removal, and phlebitis.[4] Additionally, placement in hand veins impedes activities of daily living (eg, hand washing) and is not preferred by patients.[5] PIV size ranges from 24 gauge (smallest) to 14 gauge (largest); larger catheters are often reserved for fluid resuscitation or blood transfusion as they accommodate greater flow and limit hemolysis. To decrease risk of phlebitis and thrombosis, the shortest catheter and smallest diameter should be used. However, unless adequately secured, smaller diameter catheters are also associated with greater rates of accidental removal.[4, 5]

Figure 1

By definition, PIVs are short‐term devices. The CDC currently recommends removal and replacement of these devices no more frequently than every 72 to 96 hours in adults. However, a recent randomized controlled trial found that replacing PIVs when clinically indicated (eg, device failure, phlebitis) rather than on a routine schedule added 30 hours to their lifespan without an increase in complications.[6] A systematic review by the Cochrane Collaboration echoes these findings.[3] These data have thus been incorporated into recommendations from the Infusion Nurses Society (INS) and the National Health Service in the United Kingdom.[5, 7] In hospitalized patients, this approach is relevant, as it preserves venous access sites, maximizes device dwell, and limits additional PIV insertions. In turn, these differences may reduce the need for invasive VADs such as PICCs. Furthermore, the projected 5‐year savings from implementation of clinically indicated PIV removal policies is US$300 million and 1 million health‐worker hours in the United States alone.[4]

PIVs offer many advantages. First, they are minimally invasive and require little training to insert. Second, they can be used for diverse indications in patients requiring short‐term (1 week) venous access. Third, PIVs do not require imaging to ensure correct placement; palpation of superficial veins is sufficient. Fourth, PIVs exhibit a risk of bloodstream infection that is about 40‐fold lower than more invasive, longer‐dwelling VADs[8] (0.06 bacteremia per 1000 catheter‐days).

Despite these advantages, PIVs also have important drawbacks. First, a quarter of all PIVs fail through occlusion or accidental dislodgement.[4] Infiltration, extravasation, and hematoma formation are important adverse events that may occur in such cases. Second, thrombophlebitis (pain and redness at the insertion site) is frequent, and may require device removal, especially in patients with catheters 20 guage.[9] Third, despite their relative safety, PIVs can cause localized or hematogenous infection. Septic thrombophlebitis (superficial thrombosis and bloodstream infection) and catheter‐related bloodstream infection, though rare, have been reported with PIVs and may lead to serious complications.[8, 10] In fact, some suggest that the overall burden of bloodstream infection risk posed by PIVs may be similar to that of CICCs given the substantially greater number of devices used and greater number of device days.[8]

PIVs and other peripheral VADs are not suitable for infusion of vesicants or irritants, which require larger, central veins for delivery. Vesicants (drugs that cause blistering on infusion) include chemotherapeutic agents (eg, dactinomycin, paclitaxel) and commonly used nonchemotherapeutical agents (eg, diazepam, piperacillin, vancomycin, esmolol, or total parenteral nutrition [TPN]).[11] Irritants (phlebitogenic drugs) cause short‐term inflammation and pain, and thus should not be peripherally infused for prolonged durations. Common irritants in the hospital setting include acyclovir, dobutamine, penicillin, and potassium chloride.

Of note, about one‐quarter of PIV insertions fail owing to difficult intravenous access.[12] Ultrasound‐guided peripheral intravenous (USGPIV) catheter placement is emerging as a technique to provide peripheral access for such patients to avoid placement of central venous access devices. Novel, longer devices (>8 cm) with built‐in guide wires have been developed to increase placement success of USGPIVs. These new designs provide easier access into deeper arm veins (brachial or basilic) not otherwise accessible by short PIVs. Although studies comparing the efficacy of USGPIV devices to other VADs are limited, a recent systematic review showed that time to successful cannulation was shorter, and fewer attempts were required to place USGPIVs compared to PIVs.[13] A recent study in France found that USGPIVs met the infusion needs of patients with difficult veins with minimal increase in complications.[14] Despite these encouraging data, future studies are needed to better evaluate this technology.

Midline Catheter

A midline is a VAD that is between 7.5 to 25 cm in length and is typically inserted into veins above the antecubital fossa. The catheter tip resides in a peripheral upper arm vein, often the basilic or cephalic vein, terminating just short of the subclavian vein (Figure 1B). Midline‐like devices were first developed in the 1950s and were initially used as an alternative to PIVs because they were thought to allow longer dwell times.[15] However, because they were originally constructed with a fairly rigid material, infiltration, mechanical phlebitis, and inflammation were common and tempered enthusiasm for their use.[15, 16] Newer midline devices obviate many of these problems and are inserted by ultrasound guidance and modified Seldinger technique.[17] Despite these advances, data regarding comparative efficacy are limited.

Midlines offer longer dwell times than standard PIVs owing to termination in the larger diameter basilic and brachial veins of the arm. Additionally, owing to their length, midlines are less prone to dislodgement. As they are inserted with greater antisepsis than PIVs and better secured to the skin, they are more durable than PIVs.[5, 9, 18] Current INS standards recommend use of midlines for 1 to 4 weeks.[5] Because they terminate in a peripheral vein, medications and infusions compatible with midlines are identical to those that are infused through a PIV. Thus, TPN, vesicants or irritants, or drugs that feature a pH <5 or pH >9, or >500 mOsm should not be infused through a midline.[15] New evidence suggests that diluted solutions of vancomycin (usually pH <5) may be safe to infuse for short durations (<6 days) through a midline, and that concentration rather than pH may be more important in this regard.[19] Although it is possible that the use of midlines may extend to agents typically not deemed peripheral access compatible, limited evidence exists to support such a strategy at this time.

Midlines offer several advantages. First, because blood flow is greater in the more proximal veins of the arm, midlines can accommodate infusions at rates of 100 to 150 mL/min compared to 20 to 40 mL/min in smaller peripheral veins. Higher flow rates offer greater hemodilution (dilution of the infusion with blood), decreasing the likelihood of phlebitis and infiltration.[20] Second, midlines do not require x‐ray verification of tip placement; thus, their use is often favored in resource‐depleted settings such as skilled nursing facilities. Third, midlines offer longer dwell times than peripheral intravenous catheters and can thus serve as bridge devices for short‐term intravenous antibiotics or peripheral‐compatible infusions in an outpatient setting. Available evidence suggests that midlines are associated with low rates of bloodstream infection (0.30.8 per 1000 catheter‐days).[17] The most frequent complications include phlebitis (4.2%) and occlusion (3.3%).[20] Given these favorable statistics, midlines may offer a good alternative to PIVs in select patients who require peripheral infusions of intermediate duration.

Intraosseous Vascular Access

Intraosseous (IO) devices access the vascular system by piercing cortical bone. These devices provide access to the intramedullary cavity and venous plexi of long bones such as the tibia, femur, or humerus. Several insertion devices are now commercially available and have enhanced the ease and safety of IO placement. Using these newer devices, IO access may be obtained in 1 to 2 minutes with minimal training. By comparison, a central venous catheter often requires 10 to 15 minutes to insert with substantial training efforts for providers.[21, 22, 23]

IO devices thus offer several advantages. First, given the rapidity with which they can be inserted, they are often preferred in emergency settings (eg, trauma). Second, these devices are versatile and can accommodate both central and peripheral infusates.[24] Third, a recent meta‐analysis found that IOs have a low complication rate of 0.8%, with extravasation of infusate through the cortical entry site being the most common adverse event.[21] Of note, this study also reported zero local or distal infectious complications, a finding that may relate to the shorter dwell of these devices.[21] Some animal studies suggest that fat embolism from bone may occur at high rates with IO VADs.[25] However, death or significant morbidity from fat emboli in humans following IO access has not been described. Whether such emboli occur or are clinically significant in the context of IO devices remains unclear at this time.[21]

Peripherally Inserted Central Catheter

PICCs are inserted into peripheral veins of the upper arm (eg, brachial, basilica, or cephalic vein) and advanced such that the tip resides at the cavoatrial junction (Figure 1C). PICCs offer prolonged dwell times and are thus indicated when patients require venous access for weeks or months.[26] Additionally, they can accommodate a variety of infusates and are safer to insert than CICCs, given placement in peripheral veins of the arm rather than central veins of the chest/neck. Thus, insertion complications such as pneumothorax, hemothorax, or significant bleeding are rare with PICCs. In fact, a recent study reported that PICC insertion by hospitalists was associated with low rates of insertion or infectious complications.[27]

However, like CICCs, PICCs are associated with central lineassociated bloodstream infection (CLABSI), a serious complication known to prolong length of hospital stay, increase costs, and carry a 12% to 25% associated mortality.[28, 29] In the United States alone, over 250,000 CLASBI cases occur per year drawing considerable attention from the CDC and Joint Commission, who now mandate reporting and nonpayment for hospital‐acquired CLABSI.[30, 31, 32] A recent systematic review and meta‐analysis found that PICCs are associated with a substantial risk of CLABSI in hospitalized patients.[33] Importantly, no difference in CLABSI rates between PICCs and CICCs in hospitalized patients was evident in this meta‐analysis. Therefore, current guidelines specifically recommend against use of PICCs over CICCs as a strategy to reduce CLABSI.[34] Additionally, PICCs are associated with 2.5‐fold greater risk of deep vein thrombosis (DVT) compared to CICCs; thus, they should be used with caution in patients with cancer or those with underlying hypercoagulable states.

Of particular import to hospitalists is the fact that PICC placement is contraindicated in patients with stage IIIB or greater chronic kidney disease (CKD). In such patients, sequelae of PICC use, such as phlebitis or central vein stenosis, can be devastating in patients with CKD.[35] In a recent study, prior PICC placement was the strongest predictor of subsequent arteriovenous graft failure.[36] For this reason, Choosing Wisely recommendations call for avoidance of PICCs in such patients.[37]

Centrally Inserted Central Catheter

CICCs are CVADs placed by puncture and cannulation of the internal jugular, subclavian, brachiocephalic, or femoral veins (Figure 1D) and compose the vast majority of VADs placed in ICU settings.[38, 39] Central termination of CICCs allows for a variety of infusions, including irritants, vesicants, and vasopressors, as well as blood withdrawal and hemodynamic monitoring. CICCs are typically used for 7 to 14 days, but may remain for longer durations if they remain complication free and clinically necessary.[40] A key advantage of CICCs is that they can be placed in emergent settings to facilitate quick access for rapid infusion or hemodynamic monitoring. In particular, CICCs are inserted in the femoral vein and may be useful in emergency settings. However, owing to risk of infection and inability to monitor central pressures, these femoral devices should be replaced with a proper CICC or PICC when possible. Importantly, although CICCs are almost exclusively used in intensive or emergency care, PICCs may also be considered in such settings.[41, 42] CICCs usually have multiple lumens and often serve several simultaneous functions such as both infusions and hemodynamic monitoring.

Despite their benefits, CICCs have several disadvantages. First, insertion requires an experienced clinician and has historically been a task limited to physicians. However, this is changing rapidly (especially in Europe and Australia) where specially trained nurses are assuming responsibility for CICC placement.[43] Second, these devices are historically more likely to be associated with CLABSI, with estimates of infection rates varying between 2 and 5 infections per 1000 catheter‐days.[44] Third, CICCs pose a significant DVT risk, with rates around 22 DVTs per 1000 catheter‐days.[45] However, compared to PICCs, the DVT risk appears lower, and CICC use may be preferable in patients at high risk of DVT, such as critically ill or cancer populations.[46] An important note to prevent CICC insertion complications relates to use of ultrasound, a practice that has been associated with decreased accidental arterial puncture and hematoma formation. The role of ultrasound guidance with PICCs as well as implications for thrombotic and infectious events remains less characterized at this point.[47]

Tunneled Central Venous Access Devices

Tunneled devices (either CICCs or PICCs) are characterized by the fact that the insertion site on the skin and site of ultimate venipuncture are physically separated (Figure 1E). Tunneling limits bacterial entry from the extraluminal aspect of the CVAD to the bloodstream. For example, internal jugular veins are often ideal sites of puncture but inappropriate sites for catheter placement, as providing care to this area is challenging and may increase risk of infection.[34] Tunneling to the infraclavicular area provides a better option, as it provides an exit site that can be adequately cared for. Importantly, any CVAD (PICCs or CICCs) can be tunneled. Additionally, tunneled CICCs may be used in patients with chronic or impending renal failure where PICCs are contraindicated because entry into dialysis‐relevant vessels is to be avoided.[48] Such devices also allow regular blood sampling in patients who require frequent testing but have limited peripheral access, such as those with hematological malignancies. Additionally, tunneled catheters are more comfortable for patients and viewed as being more socially acceptable than nontunneled devices. However, the more invasive and permanent nature of these devices often requires deliberation prior to insertion.

Of note, tunneled devices and ports may be used as long‐term (>3 months to years) VADs. As our focus in this review is short‐term devices, we will not expand the discussion of these devices as they are almost always used for prolonged durations.[7]

Short Peripheral Intravenous Catheter

Approximately 200 million peripheral intravenous catheters (PIVs) are placed annually in the United States, making them the most common intravenous catheter.[3] PIVs are short devices, 3 to 6 cm in length, that enter and terminate in peripheral veins (Figure 1A). Placement is recommended in forearm veins rather than those of the hand, wrist, or upper arm, as forearm sites are less prone to occlusion, accidental removal, and phlebitis.[4] Additionally, placement in hand veins impedes activities of daily living (eg, hand washing) and is not preferred by patients.[5] PIV size ranges from 24 gauge (smallest) to 14 gauge (largest); larger catheters are often reserved for fluid resuscitation or blood transfusion as they accommodate greater flow and limit hemolysis. To decrease risk of phlebitis and thrombosis, the shortest catheter and smallest diameter should be used. However, unless adequately secured, smaller diameter catheters are also associated with greater rates of accidental removal.[4, 5]

Figure 1

By definition, PIVs are short‐term devices. The CDC currently recommends removal and replacement of these devices no more frequently than every 72 to 96 hours in adults. However, a recent randomized controlled trial found that replacing PIVs when clinically indicated (eg, device failure, phlebitis) rather than on a routine schedule added 30 hours to their lifespan without an increase in complications.[6] A systematic review by the Cochrane Collaboration echoes these findings.[3] These data have thus been incorporated into recommendations from the Infusion Nurses Society (INS) and the National Health Service in the United Kingdom.[5, 7] In hospitalized patients, this approach is relevant, as it preserves venous access sites, maximizes device dwell, and limits additional PIV insertions. In turn, these differences may reduce the need for invasive VADs such as PICCs. Furthermore, the projected 5‐year savings from implementation of clinically indicated PIV removal policies is US$300 million and 1 million health‐worker hours in the United States alone.[4]

PIVs offer many advantages. First, they are minimally invasive and require little training to insert. Second, they can be used for diverse indications in patients requiring short‐term (1 week) venous access. Third, PIVs do not require imaging to ensure correct placement; palpation of superficial veins is sufficient. Fourth, PIVs exhibit a risk of bloodstream infection that is about 40‐fold lower than more invasive, longer‐dwelling VADs[8] (0.06 bacteremia per 1000 catheter‐days).

Despite these advantages, PIVs also have important drawbacks. First, a quarter of all PIVs fail through occlusion or accidental dislodgement.[4] Infiltration, extravasation, and hematoma formation are important adverse events that may occur in such cases. Second, thrombophlebitis (pain and redness at the insertion site) is frequent, and may require device removal, especially in patients with catheters 20 guage.[9] Third, despite their relative safety, PIVs can cause localized or hematogenous infection. Septic thrombophlebitis (superficial thrombosis and bloodstream infection) and catheter‐related bloodstream infection, though rare, have been reported with PIVs and may lead to serious complications.[8, 10] In fact, some suggest that the overall burden of bloodstream infection risk posed by PIVs may be similar to that of CICCs given the substantially greater number of devices used and greater number of device days.[8]

PIVs and other peripheral VADs are not suitable for infusion of vesicants or irritants, which require larger, central veins for delivery. Vesicants (drugs that cause blistering on infusion) include chemotherapeutic agents (eg, dactinomycin, paclitaxel) and commonly used nonchemotherapeutical agents (eg, diazepam, piperacillin, vancomycin, esmolol, or total parenteral nutrition [TPN]).[11] Irritants (phlebitogenic drugs) cause short‐term inflammation and pain, and thus should not be peripherally infused for prolonged durations. Common irritants in the hospital setting include acyclovir, dobutamine, penicillin, and potassium chloride.

Of note, about one‐quarter of PIV insertions fail owing to difficult intravenous access.[12] Ultrasound‐guided peripheral intravenous (USGPIV) catheter placement is emerging as a technique to provide peripheral access for such patients to avoid placement of central venous access devices. Novel, longer devices (>8 cm) with built‐in guide wires have been developed to increase placement success of USGPIVs. These new designs provide easier access into deeper arm veins (brachial or basilic) not otherwise accessible by short PIVs. Although studies comparing the efficacy of USGPIV devices to other VADs are limited, a recent systematic review showed that time to successful cannulation was shorter, and fewer attempts were required to place USGPIVs compared to PIVs.[13] A recent study in France found that USGPIVs met the infusion needs of patients with difficult veins with minimal increase in complications.[14] Despite these encouraging data, future studies are needed to better evaluate this technology.

Midline Catheter

A midline is a VAD that is between 7.5 to 25 cm in length and is typically inserted into veins above the antecubital fossa. The catheter tip resides in a peripheral upper arm vein, often the basilic or cephalic vein, terminating just short of the subclavian vein (Figure 1B). Midline‐like devices were first developed in the 1950s and were initially used as an alternative to PIVs because they were thought to allow longer dwell times.[15] However, because they were originally constructed with a fairly rigid material, infiltration, mechanical phlebitis, and inflammation were common and tempered enthusiasm for their use.[15, 16] Newer midline devices obviate many of these problems and are inserted by ultrasound guidance and modified Seldinger technique.[17] Despite these advances, data regarding comparative efficacy are limited.

Midlines offer longer dwell times than standard PIVs owing to termination in the larger diameter basilic and brachial veins of the arm. Additionally, owing to their length, midlines are less prone to dislodgement. As they are inserted with greater antisepsis than PIVs and better secured to the skin, they are more durable than PIVs.[5, 9, 18] Current INS standards recommend use of midlines for 1 to 4 weeks.[5] Because they terminate in a peripheral vein, medications and infusions compatible with midlines are identical to those that are infused through a PIV. Thus, TPN, vesicants or irritants, or drugs that feature a pH <5 or pH >9, or >500 mOsm should not be infused through a midline.[15] New evidence suggests that diluted solutions of vancomycin (usually pH <5) may be safe to infuse for short durations (<6 days) through a midline, and that concentration rather than pH may be more important in this regard.[19] Although it is possible that the use of midlines may extend to agents typically not deemed peripheral access compatible, limited evidence exists to support such a strategy at this time.

Midlines offer several advantages. First, because blood flow is greater in the more proximal veins of the arm, midlines can accommodate infusions at rates of 100 to 150 mL/min compared to 20 to 40 mL/min in smaller peripheral veins. Higher flow rates offer greater hemodilution (dilution of the infusion with blood), decreasing the likelihood of phlebitis and infiltration.[20] Second, midlines do not require x‐ray verification of tip placement; thus, their use is often favored in resource‐depleted settings such as skilled nursing facilities. Third, midlines offer longer dwell times than peripheral intravenous catheters and can thus serve as bridge devices for short‐term intravenous antibiotics or peripheral‐compatible infusions in an outpatient setting. Available evidence suggests that midlines are associated with low rates of bloodstream infection (0.30.8 per 1000 catheter‐days).[17] The most frequent complications include phlebitis (4.2%) and occlusion (3.3%).[20] Given these favorable statistics, midlines may offer a good alternative to PIVs in select patients who require peripheral infusions of intermediate duration.

Intraosseous Vascular Access

Intraosseous (IO) devices access the vascular system by piercing cortical bone. These devices provide access to the intramedullary cavity and venous plexi of long bones such as the tibia, femur, or humerus. Several insertion devices are now commercially available and have enhanced the ease and safety of IO placement. Using these newer devices, IO access may be obtained in 1 to 2 minutes with minimal training. By comparison, a central venous catheter often requires 10 to 15 minutes to insert with substantial training efforts for providers.[21, 22, 23]

IO devices thus offer several advantages. First, given the rapidity with which they can be inserted, they are often preferred in emergency settings (eg, trauma). Second, these devices are versatile and can accommodate both central and peripheral infusates.[24] Third, a recent meta‐analysis found that IOs have a low complication rate of 0.8%, with extravasation of infusate through the cortical entry site being the most common adverse event.[21] Of note, this study also reported zero local or distal infectious complications, a finding that may relate to the shorter dwell of these devices.[21] Some animal studies suggest that fat embolism from bone may occur at high rates with IO VADs.[25] However, death or significant morbidity from fat emboli in humans following IO access has not been described. Whether such emboli occur or are clinically significant in the context of IO devices remains unclear at this time.[21]

Peripherally Inserted Central Catheter

PICCs are inserted into peripheral veins of the upper arm (eg, brachial, basilica, or cephalic vein) and advanced such that the tip resides at the cavoatrial junction (Figure 1C). PICCs offer prolonged dwell times and are thus indicated when patients require venous access for weeks or months.[26] Additionally, they can accommodate a variety of infusates and are safer to insert than CICCs, given placement in peripheral veins of the arm rather than central veins of the chest/neck. Thus, insertion complications such as pneumothorax, hemothorax, or significant bleeding are rare with PICCs. In fact, a recent study reported that PICC insertion by hospitalists was associated with low rates of insertion or infectious complications.[27]

However, like CICCs, PICCs are associated with central lineassociated bloodstream infection (CLABSI), a serious complication known to prolong length of hospital stay, increase costs, and carry a 12% to 25% associated mortality.[28, 29] In the United States alone, over 250,000 CLASBI cases occur per year drawing considerable attention from the CDC and Joint Commission, who now mandate reporting and nonpayment for hospital‐acquired CLABSI.[30, 31, 32] A recent systematic review and meta‐analysis found that PICCs are associated with a substantial risk of CLABSI in hospitalized patients.[33] Importantly, no difference in CLABSI rates between PICCs and CICCs in hospitalized patients was evident in this meta‐analysis. Therefore, current guidelines specifically recommend against use of PICCs over CICCs as a strategy to reduce CLABSI.[34] Additionally, PICCs are associated with 2.5‐fold greater risk of deep vein thrombosis (DVT) compared to CICCs; thus, they should be used with caution in patients with cancer or those with underlying hypercoagulable states.

Of particular import to hospitalists is the fact that PICC placement is contraindicated in patients with stage IIIB or greater chronic kidney disease (CKD). In such patients, sequelae of PICC use, such as phlebitis or central vein stenosis, can be devastating in patients with CKD.[35] In a recent study, prior PICC placement was the strongest predictor of subsequent arteriovenous graft failure.[36] For this reason, Choosing Wisely recommendations call for avoidance of PICCs in such patients.[37]

Centrally Inserted Central Catheter

CICCs are CVADs placed by puncture and cannulation of the internal jugular, subclavian, brachiocephalic, or femoral veins (Figure 1D) and compose the vast majority of VADs placed in ICU settings.[38, 39] Central termination of CICCs allows for a variety of infusions, including irritants, vesicants, and vasopressors, as well as blood withdrawal and hemodynamic monitoring. CICCs are typically used for 7 to 14 days, but may remain for longer durations if they remain complication free and clinically necessary.[40] A key advantage of CICCs is that they can be placed in emergent settings to facilitate quick access for rapid infusion or hemodynamic monitoring. In particular, CICCs are inserted in the femoral vein and may be useful in emergency settings. However, owing to risk of infection and inability to monitor central pressures, these femoral devices should be replaced with a proper CICC or PICC when possible. Importantly, although CICCs are almost exclusively used in intensive or emergency care, PICCs may also be considered in such settings.[41, 42] CICCs usually have multiple lumens and often serve several simultaneous functions such as both infusions and hemodynamic monitoring.

Despite their benefits, CICCs have several disadvantages. First, insertion requires an experienced clinician and has historically been a task limited to physicians. However, this is changing rapidly (especially in Europe and Australia) where specially trained nurses are assuming responsibility for CICC placement.[43] Second, these devices are historically more likely to be associated with CLABSI, with estimates of infection rates varying between 2 and 5 infections per 1000 catheter‐days.[44] Third, CICCs pose a significant DVT risk, with rates around 22 DVTs per 1000 catheter‐days.[45] However, compared to PICCs, the DVT risk appears lower, and CICC use may be preferable in patients at high risk of DVT, such as critically ill or cancer populations.[46] An important note to prevent CICC insertion complications relates to use of ultrasound, a practice that has been associated with decreased accidental arterial puncture and hematoma formation. The role of ultrasound guidance with PICCs as well as implications for thrombotic and infectious events remains less characterized at this point.[47]

Tunneled Central Venous Access Devices

Tunneled devices (either CICCs or PICCs) are characterized by the fact that the insertion site on the skin and site of ultimate venipuncture are physically separated (Figure 1E). Tunneling limits bacterial entry from the extraluminal aspect of the CVAD to the bloodstream. For example, internal jugular veins are often ideal sites of puncture but inappropriate sites for catheter placement, as providing care to this area is challenging and may increase risk of infection.[34] Tunneling to the infraclavicular area provides a better option, as it provides an exit site that can be adequately cared for. Importantly, any CVAD (PICCs or CICCs) can be tunneled. Additionally, tunneled CICCs may be used in patients with chronic or impending renal failure where PICCs are contraindicated because entry into dialysis‐relevant vessels is to be avoided.[48] Such devices also allow regular blood sampling in patients who require frequent testing but have limited peripheral access, such as those with hematological malignancies. Additionally, tunneled catheters are more comfortable for patients and viewed as being more socially acceptable than nontunneled devices. However, the more invasive and permanent nature of these devices often requires deliberation prior to insertion.

Of note, tunneled devices and ports may be used as long‐term (>3 months to years) VADs. As our focus in this review is short‐term devices, we will not expand the discussion of these devices as they are almost always used for prolonged durations.[7]