Central and Peripheral Medication Administration Practices in Pediatric Patients

Care of the pediatric critically ill patient includes obtaining appropriate intravenous (IV) access for secure, safe, and timely medication administration. Many options exist for IV access but risks and benefits of each access type should be considered. Various vascular access devices are used in the critical care setting. Peripheral access is often gained in distal veins yet—particularly in pediatric patients—peripheral devices have a high first-attempt failure rate and short patency time, making them not ideal for long-term drug administration.1 Central venous access devices (CVADs) and peripherally inserted central catheters provide direct administration to the central venous system and can allow for more stable long-term infusion needs.2 Midline catheters are inserted peripherally but terminate distal to a CVAD at or below the axillary vein.Several medication properties should be considered when deciding on central versus peripheral administration, including osmolarity, pH, and concentration. The cumulative risk of phlebitis and the caustic nature of some medications due to these properties will help guide the practitioner, pharmacist, and bedside nurse to determine the ideal prescribing, preparation, and administration techniques to avoid extravasation. Medications that are caustic, high volume or concentration, or a high osmolarity may require central venous catheter administration to avoid burning and to limit the likelihood of extravasation. Central venous access devices are not without risk, therefore the risks and benefits should be balanced when deciding on the placement of central venous access. The purpose of this article is to review complications of central venous catheters, to explain medication properties that should be evaluated when deciding on central or peripheral administration, and to evaluate the literature on peripheral administration of historically centrally administered medications. Specific medications are discussed below in a body system–based approach.Although central venous catheters ensure more secure access, their placement and maintenance are not without risk. A systematic review of 74 pediatric cohort studies aimed to define the complications and rates of complications with CVADs.3 The researchers noted that 25% (95% CI, 20.9%-29.2%) of pediatric CVADs failed before completion of therapy. The highest probability of failure was noted in hemodialysis catheters, whereas implanted devices had the lowest failure rates. Central venous access device–associated bloodstream infections accounted for 10.3% (95% CI, 8.9%-11.6%) of the device failures, with other failures being related to occlusion/breakage (7.4%), dislodgement/migration (4.7%), local infection/phlebitis (3.1%), thrombosis (1.7%), and breakage/rupture (1.6%).3 Local and systemic infections can be a result of skin flora at the insertion site, poor catheter care practices contaminating the device's internal hub, or the infusion of a contaminated product. These infections prolong hospital stay, increase the cost of care, and increase the relative risk of death.4 Occlusion can be secondary to fibrin sheath development or medication precipitation, and thrombosis at the catheter tip is possible. Insertion-related complications of CVADs have been reported in the range of 7% to 18%.5 These adverse effects include pneumothorax, arterial puncture, hemorrhage, and cardiac dysrhythmias. Because of these risks, central venous catheter placement risks and benefits should be weighed.Osmolarity is the number of milliosmoles per liter of solution (mOsm/L). It is calculated based on the osmolality (mOsm of solute/kg of solvent) and the solution's density (g/mL). Although the commonly accepted osmolarity of a peripherally administered medication is 900 mOsm/L, drugs within this range are still hyperosmolar as compared with the osmolarity of human serum, 275 to 295 mOsm/L.6,7 Some have suggested that a threshold of 450 to 600 mOsm/L may be more appropriate for peripheral medication administration.8 Infusion of medications with high osmolarity may result in osmotic changes that could lead to red blood cell damage, phlebitis, and necrosis at the injection site.9 Rarely is the osmolarity of a specific medication included in the general medication package insert, leading to potential for inadvertent hyperosmolar drug administration. Slowing the rate of a hyperosmolar infusion may also improve tolerability in a peripheral vein, suggesting that faced with a hyperosmolar substance, strategies to improve tolerability could include further dilution and slower rates of infusion.10The pH of a substance is the degree of acidity or basicity of a solution. Normal blood pH is fairly close to neutral at 7.35 to 7.45. Medications and infusates that are highly acidic or basic can result in venous damage while also affecting the patient's overall pH. Since there is less control over the pH of a specific medication, further care should be taken when deciding on infusion location.In adult patients, an expert panel consisting of a multidisciplinary group of physicians, nurses, and pharmacists attempted to define which medication properties, specifically osmolarity, pH, and labeled vesicants, would be best suited for central administration.8 They reviewed 106 drugs with 183 different concentrations. Most of the medications (91.5%) had an osmolarity of less than 600 mOsm/L, with 15 medications having an osmolarity greater than 600 mOsm/L. Sixty drugs had a pH between 5 and 7.5, with 27 drugs (20.15%) having extreme pHs less than 4 or greater than 9. On the basis of their evaluation, Manrique-Rodríguez et al8 categorized these medications into 3 different risk categories (Table 1). They further suggest that CVADs are preferred for medications they defined as high risk. Peripheral catheters remain ideal for short-term therapies with medications that are low risk for either osmolarity or pH and when the patient's venous access is in good condition. They recommend midline catheters for medications with an osmolarity and pH of moderate risk intended for intermediate length. Although aimed at adults, these general recommendations can be extrapolated to use in the pediatric setting, taking into consideration that pediatric patients are more susceptible to extravasations due to the small size of their veins and capillary leak.11 Other groups have suggested similar risk stratifications for pediatric patients.12 Naik and colleagues2 suggest central administration when the infusate has a pH of less than 5 or greater than 9, is greater than 600 mOsm/L, is hypotonic or hypertonic, or is an irritant or vesicant. Peripheral administration would be appropriate if the medication has a pH between 5 and 9, is less than 600 mOsm/L or isotonic, and is not an irritant or vesicant.2Hyperosmolar therapies, including hypertonic saline and mannitol, are primarily used for the management of life-threatening intracranial hypertension. Hypertonic saline can be considered for the treatment of acute, serious symptomatic hyponatremia, and mannitol is additionally indicated for reducing intraocular pressure (see Table 2).Mannitol is available as a 20% or 25% product, with an osmolarity of 1100 mOsm/L and 1375 mOsm/L, respectively. Hypertonic saline is defined as any solution with a sodium concentration of greater than 0.9%, such as 3% (1027 mOsm/L) or 23.4% (8008 mOsm/L).12,13 Concern for extravasation-induced tissue injury has led to recommendations not to administer solutions exceeding 900 mOsm/L via a peripheral catheter.14 However, for emergent indications, there are concerns with delaying medication administration in order to obtain central access.A retrospective chart review of 101 pediatric patients who received 3% sodium chloride boluses during pediatric critical care transport did not find any infusion reactions locally or systemically.15 Patients in this study had a mean age and weight of 5.9 years and 27.6 kg, respectively. Of the 101 included patients, 95 patients received 3% sodium chloride through a peripheral catheter. The mean dose was 5.4 mL/kg (range, 1.2–24.0 mL/kg) and mean infusion time was 47 minutes (range, 9–180 min). On the basis of the results of this study, the authors concluded that 3% sodium chloride may be administered safely during pediatric critical care transport, and the administration method can include peripheral catheters.15In 2016, an article was published that aimed to describe the safety of continuous infusion of 3% sodium chloride in neurocritical care patients, specifically related to the occurrence of infusion site reactions requiring intervention.16 There were 203 adult patients included in the study. The mean duration of 3% sodium chloride infusion was 51 hours and 7 minutes, with a mean duration of peripheral infusion of 44 hours and 43 minutes. Approximately 74% of patients received 3% sodium chloride through a peripheral catheter only, with infusion duration ranging from 21 hours and 15 minutes to 74 hours and 48 minutes. The majority of patients (64.3%) had peripheral catheters placed in the antecubital site. Needle gauge for peripheral administration sites varied from 16 to 22 gauge, with the majority having 20-gauge (46.5%) or 18-gauge (47.6%) access. Extravasation or phlebitis occurred in 7.4% of patients, but only 2.3% of patients required an infusion site change secondary to an infusion-related reaction. On the basis of this data, the authors recommended re-evaluation of the requirement of central access for continuous 3% sodium chloride infusion.16 Although this study involved adult patients, data collected and evaluated may be extrapolated to pediatric patients. When extrapolating to pediatric patients, the bedside clinician should consider the catheter gauge as well as the site of insertion.The safety of peripheral administration of 3% sodium chloride and mannitol was evaluated in a retrospective cohort study in adult and pediatric patients.14 Of the 192 included patients, 85 received 3% sodium chloride and 107 received mannitol (either 20% or 25%). There were 49 pediatric patients included in the study. Seventy-eight of the patients in the 3% sodium chloride group received a bolus, and 7 received continuous infusions. The mean bolus dose of 3% sodium chloride was 6 mL/kg. Of the 107 patients in the mannitol group, 106 received bolus doses, and 1 received a continuous infusion. The mean bolus dose of mannitol was 0.97 g/kg; most patients (84.1%) received mannitol 25%. The primary outcome was extravasation incidence, with no extravasations reported in either group. Patient demographics in this study were variable, however details of peripheral access (site of infusion, gauge of needle) would have been beneficial for further clinical applicability.14Based on the above studies, it is likely unnecessary to delay administration of bolus hyperosmolar medications until central access is obtained. Continuous infusions of hypertonic saline may be safe for peripheral administration and may be considered in patients when central access is not available or if that is the only indication for central access. The infusion site should be monitored regularly for infusion site–related reactions (ie, extravasation, phlebitis).Pentobarbital, a barbiturate, is most commonly used for seizure cessation in pediatric patients. It may also be used for sedation (preoperative, procedural, or mechanically ventilated) or to reduce elevated intracranial pressure. Because of its alkalotic properties (pH 9.5), care should be taken to prevent extravasation. Administration using a large vein or via an infusing IV catheter at a port farthest from the insertion site may decrease the risk of extravasation.17 Patients requiring continuous infusions of pentobarbital frequently have other indications for central venous catheter access. Intravenous catheter compatibility should be confirmed, as pentobarbital may be incompatible with other solutions and the bedside clinician may need to adjust other continuous infusions.Management of an acutely ill patient with sepsis will require rapid administration of fluids and potentially vasoactive medications such as epinephrine, norepinephrine, or dopamine to maintain cardiac output and perfusion. These medications work both centrally and peripherally at various adrenergic receptors throughout the body. It has been shown that morbidity and mortality increase significantly when shock persists. Carcillo and colleagues18 found that with each hour that pediatric advanced life support guidelines were not implemented, mortality increased by 40%. Similarly, other researchers have found significant increases in mortality associated with delays in inotropic resuscitation.19, 20 Imrain et al21 explored whether earlier administration of epinephrine in the fluid resuscitation phase would lead to better outcomes in pediatric patients. Patients in the study received usual fluid resuscitation, however 1 group was initiated with epinephrine after 40 mL/kg while the second group received a total of 60 mL/kg prior to epinephrine initiation. These researchers found that children in the earlier epinephrine initiation group had lower mortality (10% vs 33%, P=.04) and were less likely to require mechanical ventilation (10% vs 41%, P=.006). Some have suggested that the delay in inotropic infusion may be related to the placement of a CVAD.22Central access for these medications has historically been recommended due to the medications' peripheral vasoconstriction effects and potential for harm with extravasation. Neither the 2020 Pediatric Surviving Sepsis Guidelines nor the American Society of Health-System Pharmacists (ASHP) Standardize 4 Safety initiative issued a recommendation for or against peripheral vasoactive infusion.23,24 The majority (82%) of the expert panel authors of the Surviving Sepsis guidelines reported sometimes using peripheral access for initial vasoactive infusion in acute settings. Most also preferred the use of epinephrine or dopamine peripherally over norepinephrine.23 The American College of Critical Care Medicine has recommended that epinephrine may be administered via peripheral IV, as no data demonstrate worse extravasation-related tissue damage than that associated with dopamine.25 The primary literature describing these practices can help to direct in which patients it may be appropriate to initiate peripheral vasoactive infusions.In one of the largest retrospective analyses on peripheral vasoactive administration, out of 756 children who received vasoactive therapy, 231 (30.6%) patients had vasopressors initiated peripherally.26 The majority of patients received dopamine (82%), with the remaining receiving norepinephrine (19%) and epinephrine (13%). Fifty-three patients received more than 1 vasopressor concomitantly via the same peripheral catheter. Most peripheral catheters were located in the arm (58%), hand (25%), or lower extremity (8%), and the most common catheter size was 20 or 22 gauge (75%). Median duration of infusion varied by vasoactive medication (norepinephrine, 205 min; dopamine, 180 min; and epinephrine, 160 min). As compared with patients who had vasoactive medications administered through a CVAD, patients with peripheral access were older, larger, and had lower illness severity. Also, peripheral administration was initiated more often at night, and those patients were less likely to be supported with mechanical ventilation and had shorter length of stays.26 It is worth noting that the concentrations of epinephrine (100 μg/mL) and norepinephrine (100 μg/mL) were higher than those recommended by ASHP's Standardize 4 Safety initiative (epinephrine: 20 μg/mL, 40 μg/mL; norepinephrine: 16 μg/mL, 32 μg/mL, 64 μg/mL), highlighting that at more dilute concentrations, safety may be maintained.24Fifty-four percent of patients had a CVAD placed after a median duration of 140 minutes of peripheral vasoactive infusion. Four of these patients had an extravasation (1.7%), which exclusively occurred in the hand (epinephrine, n=2; dopamine, n=2).26 Two of the extravasations occurred after prolonged (>12-h) vasoactive infusion, suggesting that longer durations may carry a higher risk. The median doses at the time of extravasation were similar to the maximum doses seen in the entire cohort. Three of the extravasation events required additional therapy (phentolamine, n = 2; terbutaline, n=1; nitroglycerin, n=1). None of the extravasation events was associated with long-term disability in the affected limb, although 1 patient later died after withdrawal of life-sustaining support.26Oftentimes when pediatric patients are being transported to a higher level of care, they may only have peripheral access. D'Souza and colleagues27 collected data on 45 children receiving vasoactive agents via a peripheral venous catheter on transport.27 These agents included epinephrine (n=18), norepinephrine (n=23), and both epinephrine and norepinephrine (n=4). Upon pediatric intensive care unit (PICU) admission, 28 patients (62%) did not require CVAD insertion, and 17 of those patients had their vasoactive agents discontinued within 6 hours of PICU admission. No extravasation events were noted in this cohort.27 However, in another evaluation of 75 transport patients receiving vasoactive agents peripherally, 11 patients (15%) developed IV infiltrates within the first 24 hours of facility arrival.28 Most of these patients (90%) were treated with dopamine monotherapy, and the authors associated infiltrations with longer durations of peripheral IV administration and higher doses. Given the mixed data, it is likely prudent for the clinician to evaluate any IV catheter that is placed at an alternative site of care when admitting the patient, and if the IV catheter is not functioning properly, to replace or switch any vasoactive agents to an alternative access point.The delay in getting a CVAD and starting peripheral administration of medications may give the clinical team time to reassess the patient and their long-term need for central access. Data from previous studies suggests that those patients who were able to avoid a CVAD were larger, had lower illness severity, and required less invasive support. This suggests that with a rapidly improving clinical trajectory, the clinician may be able to avoid the placement of a CVAD and potentially additional interventions such as arterial catheters.26Amiodarone, an antiarrhythmic agent, is used to treat tachyarrhythmias. Injection site reactions such as erythema, edema, necrosis, extravasation, infiltration, inflammation, thrombophlebitis, and phlebitis have been reported with infusion. An in-line filter is recommended during administration to reduce the risk of phlebitis. For pulseless ventricular tachycardia or ventricular fibrillation, undiluted IV amiodarone may be administered through the largest bore IV catheter, as an IV push. For perfusing arrhythmias, amiodarone requires dilution. When available, central access is preferred for amiodarone administration, and it is recommended to use a central venous catheter for concentrations greater than 2 mg/mL.29Parenteral nutrition (PN) may be required in critically ill patients or those who are unable to receive nutritional requirements through the enteral route. Macronutrient components of PN include dextrose, amino acids, and fat.Parenteral nutrition may be administered in either a 2-in-1 or 3-in-1 formulation. A 2-in-1 admixture contains dextrose, amino acids, electrolytes, vitamins, minerals, and fluids in one solution and another containing intravenous lipid emulsion (ILE). A 3-in-1 admixture contains all nutrients mixed in one solution. Many institutions use 2-in-1 admixtures for inpatient settings, whereas 3-in-1 admixtures are more commonly used in the long-term care or home setting. Although a 3-in-1 admixture is appealing—with the ability to administer all components from one solution—there are risks associated with these admixtures including but not limited to decreased stability, especially in pediatric patients; the potential for increased microbial growth and infectious complications; compatibility issues; and increased risk of catheter occlusion. For acute inpatient use, 2-in-1 admixtures are preferred for these reasons. However, for home use, 3-in-1 admixtures have benefits such as reduced risk of contamination due to decreased manipulation, ease of administration, and less ILE waste. Therefore, when evaluating risks and benefits for home administration, 3-in-1 admixtures may be preferred for pediatric patients. However, neonatal patients should always receive 2-in-1 admixtures regardless of setting, as minimizing risk of infection and increasing stability outweigh any administration benefits of 3-in-1 admixtures.30 Additional administration considerations for PN and ILE include the use of filters. A 1.2 micrometer filter may be used for all PN formulations, and a 0.22 micrometer filter may be used for 2-in-1 formulations.31Most PN is a hypertonic solution and has an acidic pH secondary to high quantities of dextrose and amino acids required to meet pediatric patients' nutritional needs. Therefore, central administration of PN is generally preferred. Infusion of PN through a peripheral catheter may be done in conjunction with careful consideration of the solution's osmolarity and close administration site monitoring, as the incidence of thrombophlebitis is related to the osmotic content of infused solutions as well as infusion rate.32 A commonly cited recommendation is to limit osmolarity of peripheral PN to 900 mOsm/L.Intravenous lipid emulsions are essential components of PN and secondary to their isotonicity (20% ILE osmolarity ranges from 270 to 350 mOsm/L) can be administered through a peripheral or central catheter and may be administered concurrently with PN.31,33,34 Anecdotal human experience led to the prospective evaluation of the effect of continuous administration of fat emulsion on the incidence of, probability of, and time to infiltration of peripheral intravenous sites in infants receiving peripheral PN.35 Infants received solutions of dextrose (group 1), dextrose/amino acids (group 2), or dextrose/amino acids plus fat emulsion (group 3); each solution had an osmolarity of approximately 900 mOsm/L. For those receiving fat emulsion, it was administered through a Y-site into the primary dextrose/ amino acid solution.The incidence of infiltration at 48 hours was 54% for group 1, 63% for group 2, and 42% for group 3 (P > .05). Infiltration frequency for the entire time for group 1 was 71%; for group 2, 66%; and for group 3, 67% (P>.05). Sites in group 1 lasted a mean (SD) of 54.9 (7.8) hours; group 2, 43.6 (4.2) hours; and group 3 26.3 (3.3) hours (P = .003).35 Although the incidence of infiltration did not differ among the 3 groups, there was a statistically significant difference in the time to infiltration between the 3 groups. A further comparison between groups demonstrated no difference in age, weight, race, gender, cannula gauge, cannula site, rate of infusion, amount of calories infused, or the presence or absence of fat emulsion between groups (P>.005). The sites that infiltrated at 48 hours were more likely to have had medication infusions than those that did not infiltrate at 48 hours (P<.003), however this difference was not maintained the entire time. On the basis of the results of this study, the authors concluded that the continuous administration of fat emulsion provides some protection to peripheral veins of infants receiving concomitant dextrose/amino acid solutions. Based on the increased infiltration rate at 48 hours for patients receiving multiple medications through the same access point, it may be reasonable to avoid coadministration of medications with PN and fat emulsion.35Potassium salts require dilution prior to parenteral administration. In general, the dose, concentration, and rate of administration may be dependent on patient condition and specific institutional policy. Potassium chloride is considered a vesicant at concentrations greater than 0.1 mEq/mL, and it is generally recommended to infuse concentrations greater than 0.1 mEq/mL through a central venous catheter. For noncritically ill patients, infusion rates range from 0.2 to 0.5 mEq/kg/h up to 10 to 20 mEq/h, with a generally accepted maximum rate of peripheral infusion of 10 mEq/h. Higher rates for critically ill patients may be used; maximum reported rates range from 1 to 2 mEq/kg/h, with a maximum of 40 mEq/h. Continuous cardiac monitoring is recommended for patients with severe hypokalemia and for infusion rates greater than 0.5 mEq/kg/h.36Potassium phosphate dose and concentration are displayed as mmol of phosphate. Potassium phosphate contains 4.4 mEq of potassium for every 3 mmol of phosphate. Therefore, although the dose of potassium phosphate is based on the mmol of phosphate, care should be taken to ensure that the potassium dose, concentration, and rate are appropriate. Undiluted concentrated potassium phosphate (3 mmol/mL) is hypertonic, with an osmolarity of 7700 mOsm/L, making it necessary to dilute it in a larger volume of fluid for administration. Concentrated formulations should never be administered IV without being diluted.37For pediatric patients 12 years of age and older, the prescribing information recommends a maximum concentration for peripheral administration of phosphorous 0.064 mmol/mL (potassium, 0.1 mEq/mL), with a maximum infusion rate of approximately phosphorus 6.4 mmol/h (potassium, 10 mEq/h). The recommended maximum concentration for central administration is phosphorus 0.18 mmol/mL (potassium, 0.282 mEq/mL), with a maximum infusion rate of phosphorus 15 mmol/h (potassium, 23.5 mEq/h). A central venous catheter, in addition to continuous cardiac monitoring, is recommended for infusion rates of potassium higher than 0.5 mEq/kg/h.38,39 For pediatric patients less than 12 years of age, the manufacturer recommends further dilution to 0.027 mmol/mL (potassium, 0.04 mEq/mL) for peripheral administration and 0.055 mmol/mL (potassium, 0.08 mEq/mL) for central administration.40 No specific administration information is provided by the manufacturer for rate of administration, but administration practices for patients greater than 12 years of age could be applied. Tertiary resources suggest more concentrated infusions, 0.05 mmol/mL (potassium, 0.073 mEq/mL) for peripheral administration and 0.12 mmol/mL (potassium, 0.18 mEq/mL) for central administration.38Potassium phosphate can lead to venous thrombophlebitis primarily when administered peripherally. Diluting this salt form to reduce the osmolarity of the agent can help with peripheral administration.39,40 Local institutional policies and protocols related to potassium and phosphorus administration should be followed and can help the bedside clinician discern which catheter administration and concentration is most appropriate.Similar to potassium phosphate, the maximum suggested concentration of sodium phosphate is not based on the phosphate component. Sodium phosphate contains 4 mEq of sodium for every 3 mmol of phosphate; although sodium phosphate dosing is based on the phosphate component, clinicians should carefully consider the sodium load when using sodium phosphate. Sodium phosphate requires significant dilution prior to administration, as its osmolarity is 7000 mOsm/L. The maximum suggested concentration of sodium phosphate, based primarily on sodium concentration, is 0.05 mmol/mL for peripheral administration and 0.12 mmol/mL for central venous catheters. Sodium phosphate should be administered over 4 to 12 hours, depending on the severity of hypophosphatemia. Rapid infusions of phosphorus can result in a decrease in calcium concentrations and symptoms of hypocalcemic tetany, arrhythmias, ectopic calcification, and acute kidney injury.39,41Consideration should be made for other electrolytes with an elevated osmolarity. As discussed above, hypertonic saline products are more readily recognizable as hyperosmolar therapies given their use in practice as well as attention from a medication safety perspective.42 Medications like sodium bicarbonate 8.4%, which is commercially available in vials and carpujects as 1 mEq/mL, also have a high osmolarity (1560 mOsm/L). Neonatal bicarbonate products are more dilute, for example, sodium bicarbonate 4.2% (0.5 mEq/mL), and have lower osmolarities, making them more appropriate for administration to the neonatal/infant population and better suited for peripheral administration. Care should be taken if administering sodium bicarbonate 8.4% undiluted through peripheral access, and dilution should be considered.Although some electrolyte administration is dependent on concentration to determine appropriate central or peripheral administration, medications like calcium chloride, which even at low concentrations can cause significant tissue necrosis, should only be administered via a central venous catheter. Alternative salt forms like calcium gluconate are traditionally used for peripheral administration for hypocalcemia, yet these also carry a risk of tissue necrosis if extravasated. If central access is available, it is still preferable to administer calcium gluconate centrally to avoid this risk.Vancomycin is an antibiotic, most known for its methicillin-resistant Staphylococcus aureus coverage. Vancomycin can be administered via peripheral or central access, typically over 60 minutes. Administration adverse effects can include vancomycin infusion reactions and thrombophlebitis.43 Vancomycin infusion reactions, characterized by flushing and pruritis, may occur with rapid administration and can be prevented by slowing the infusion over 90 to 120 minutes, further diluting the vancomycin, and administering antihistamines. Thrombophlebitis may also occur with vancomycin administration. Vancomycin is considered an irritant and has a pH between 2.5 and 4.5. Due to its irritant property, if extravasated, it can cause pain, tenderness, and necrosis. An in vitro analysis concluded that the local toxicity of vancomycin appears to be secondary to higher concentrations (>5 mg/mL) and less so from its low pH.44 The authors also noted that toxicity was also related to the length of infusion. As more time passed, cellular viability decreased at stable concentrations suggesting that the longer time that a patient receives vancomycin, the higher their risk of developing thrombophlebitis. Thrombophlebitis can be prevented by slowing infusion rates, diluting the solution, and rotating infusion sites.43, 45 Medication dilution is most quantifiable when performed preadministration (ie, dilution of vancomycin 500 mg in 100 mL of normal saline). Another method of dilution is postadministration, when the medication dilutes further in the blood. It has been suggested that due to higher blood flow through a central vessel, more concentrated medications administered via a CVAD may be better tolerated, but clinical outcomes of this are difficult to assess.46 Vancomycin is recommended to be administered peripherally or centrally at a concentration of ≤5 mg/mL. For fluid-restricted patients, concentrations of 10 mg/mL have been used, however with an increased risk of infusion reactions.43,45 Administration via a central venous catheter may be preferred if the patient experiences thrombophlebitis and central access is available.Micafungin is an antifungal agent of the echinocandin class used to prevent and treat fungal infections. There are postmarketing reports of local injection site phlebitis and thrombophlebitis that have been associated with peripheral administration, prompting recommendations for selection of administration site based on concentration. Micafungin can be prepared with a range of final concentrations, from 0.5 to 4 mg/mL; for concentrations greater than 1.5 mg/mL, a central venous catheter should be used.47, 48Chemotherapy can be a chemical vesicant or irritant with low or high pH and can be quite damaging to surrounding tissue in the event of an extravasation. Chemotherapy extravasation can be considered an oncologic emergency. Depending on the property of the chemotherapy, central administration is preferred to limit the risk of extravasation. Due to an oncology patient's need for frequent chemotherapy administrations and need for long-term access, most often central access via a port or tunneled catheter is obtained to limit peripheral administration of chemotherapy. For patients with surgically placed catheters that have become infected and need to be removed, peripheral access may be used temporarily before another central venous catheter can be placed. In these cases, risks and benefits should be weighed with respect to administration of certain chemotherapies. The European Society for Medical Oncology recommends a preventative approach to minimize extravasation risk.49 Specifically, they recommend selecting large veins in the forearm and avoiding cannula sites in the inner wrist, lower extremities, anticubital fossa, and dorsum of the hand. They also recommend that if infusions will last greater than 12–24 hours, central access should be obtained. When using a peripheral catheter, they recommend checking for blood flow and flushing with normal saline initially and between different drug infusions to check for signs of extravasation. The site should be regularly inspected for redness, swelling, or pain during infusion of any medication. Any extravasation should be quickly managed to limit tissue necrosis.As the care of pediatric patients evolves, the bedside clinician should remain vigilant about administration practices that are safe and based on evidence. Although primary literature suggests that some infusates historically considered for administration by central venous catheters only may be given in peripheral catheters, peripheral catheters should be inspected regularly to limit complications. Paradigms that analyze medications based on pH, osmolarity, and vesicant properties may be helpful to further define medications that should be administered centrally. Whenever possible, safe medication administration should be dictated by national guidelines and local policies.