| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Department of Paediatrics, The Hospital for Sick Children, 555 University Ave., Toronto, Ontario M5G 1X8, Canada, and the University of Toronto, Toronto, Ontario, Canada.
 |
Abstract |
|---|
 Top Abstract IntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
Key Words: indexing terms: newborns • pharmacokinetics • drug administration
|
Introduction |
|---|
|
TopAbstractIntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
Here I will review differences in disposition and response to drugs in infants and children. Developmental changes will be dealt with, including specific problems associated with drug administration to this age group.
|
Pharmacokinetic Aspects |
|---|
|
TopAbstractIntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
Gastric acidity.
At birth, gastric pH is in the
neutral range (between 6 and 8), owing to the presence of alkaline
amniotic fluids. Within a day, the pH generally falls to between 1 and
3. However, gastric acidity is poorly maintained in neonates. Only by
age 3 years does the production of acidity reach adult capacity
(H+ per hour, 0.15 mmol/10 kg body wt. in neonates vs 2
mmol/10 kg in adults after a stimulation test)
(1)(2). These age-dependent changes correspond
closely to the development of the gastric mucosa. The relatively
alkaline milk consumed by the infant further decreases gastric acidity.
These differences in gastric pH have been shown to affect the absorption of several drugs. For example, oral penicillin, ampicillin, and nafcillins (all acid labile) achieve higher concentrations in neonates than in children or adults (3)(4)(5)(6). Higher concentrations of these agents in neonates may also result from lower elimination rates (see below). The relative contributions of these two mechanisms have yet to be defined.
On the other hand, acidic drugs, such as nalidixic acid, are better absorbed in their nonionized form. In alkaline pH, a larger fraction of the drug is ionized and is therefore less well absorbed by the small infant (7).
Gastric emptying.
In infants younger than 6 months,
gastric emptying is much slower than in older children and adults. In
normal adults, gastric emptying is biphasic, a rapid (10–20 min) first
phase being followed by an exponentially slower phase. In the preterm
infant, gastric emptying is slow and linear. The prolonged gastric
emptying in the small infant (6–8 h) results in increased absorption
of the penicillins (ampicillin, nafcillin) and decreased absorption of
acidic drugs (e.g., nalidixic acid).
Theoretically, one would expect that drugs relatively poorly absorbed in adults might have an improved absorption rate in young infants, owing to prolonged contact with the gastrointestinal mucosa secondary to slow gastric emptying. However, limited data suggest that certain drugs, including amoxicillin, rifampin, chloramphenicol, and cephalosporins, demonstrate delayed and incomplete absorption in neonates and small infants (8)(9)(10). After birth, there is a gradual improvement of gastrointestinal absorption of drugs, and by age 3 months, absorption may be comparable with or even more complete than that seen in adults (11).
Intramuscular absorption.
The bioavailability of drugs
after intramuscular injection depends on the perfusion in the area of
injection, the rate of drug penetration through the capillary
endothelium, and the apparent volume into which the drug has been
distributed. Several physiological factors distinguish neonates from
older children and adults, including vasomotor instability, less
muscular mass and subcutaneous fat, and a higher proportion of water.
In addition, the pathological states of hypovolemia, hypothermia, and
hypoxemia may decrease perfusion and potentially decrease intramuscular
bioavailability. Only sparse information exists comparing intramuscular
administration of drugs in various age groups. Apparently, the time
needed to achieve peak concentration (Tmax) for
drugs administered by the intramuscular route is comparable for
infants, children, and adults for aminoglycosides, ampicillin, and
carbenicillin (11)(12)(13)(14), whereas preterm infants exhibit
delayed absorption of chloramphenicol, rifampicin, cephalexin,
cephaloridine, and benzylpenicillin
(10)(11)(15)(16).
Percutaneous absorption.
Two major factors determine the
rate and extent of percutaneous absorption and may cause excessive
absorption of an agent applied to the skin in the neonate and small
infants. The thickness of the epidermal stratum corneum is inversely
related to absorption, whereas the state of skin hydration directly
influences absorption. Several antiseptic agents have been implicated
in cases of severe toxicity in neonates after percutaneous absorption.
For many years, hexachlorophene (PHisohex) was used routinely for the
"total body bath" to guard against skin bacteria, in the belief
that absorption occurred only through wounds and burns. However, cases
of severe central nervous system toxicity with generalized seizures
reported in neonates with intact skin led to findings that 3% of
infants so bathed had blood concentrations of hexachlorophene
comparable with those causing seizures in animals (17).
Subsequently, the American Academy of Pediatrics recommended this
compound not be used on extensive areas (18). Similarly,
increased skin permeability was shown with boric acid when used as a
skin antiseptic in infants. Given the findings of severe intoxications
and the existence of safer and less-penetrating antiseptic agents, it
is now thought that boric acid should not be used at all.
The Eutectic Mixture of Local Anesthetics (EMLA) is the first effective noninvasive method for skin anesthesia. One of its components, prilocaine, is metabolized in the liver to o-toluidine, which can cause methemoglobinemia. The immature activity of the enzyme diaphorase in their erythrocytes makes infants more prone to methemoglobinemia in such circumstances; nonetheless, as my colleagues and I recently showed, even in preterm infants, applying EMLA before heel pricking does not produce methemoglobinemia (19).
distribution
Several determinants of drug distribution also are markedly
different in neonates and infants compared with adults. Some of the
factors influencing the distribution of antimicrobial drugs, such as
protein binding and compartmentalization of body water, change
continuously during the first years of life—as does, therefore, drug
distribution in this group.
Protein binding.
In general, protein binding of drugs is
lower in newborns than in older children and adults—the result of
lower albumin values, the lower affinity of fetal albumin for drugs,
and the presence of endogenous compounds, such as bilirubin, that
compete for protein binding. Examples of drugs for which lower protein
binding has been documented in neonates are phenytoin, salicylates,
ampicillin, nafcillin, sulfisoxazole, and sulfamethoxyphrazine
(3)(20)(21)(22). Consequently, greater free
fractions of these drugs are circulating and thus are able to penetrate
various tissue compartments, yielding higher distribution volumes
(Vd).
Body water.
In the newborn infant, water makes up
70–75% of the body weight, compared with 50–55% in adults, and the
extracellular water component is greater (40% of body weight vs 20%
in adults). Infants have less fat tissue, 15% vs 20% of body weight
in adults, and 25% less muscle tissue. These characteristics affect
the Vd of drugs that are mainly distributed
in body water and, to a lesser extent, of lipophilic drugs. Gentamicin
and amikacin have been shown to have a greater distribution in
neonates, which tends to decrease gradually during childhood
(11). These changes are explained mainly by changes in the
percentage of body water, given that the aminoglycosides bind minimally
to plasma proteins.
The larger Vd in infants and small children
means that, at equal doses (per body weight), the peak concentrations
produced in their blood will be lower than in adults. However, the mean
serum concentration at steady state is independent of the
Vd and therefore is unaffected, according
to the equation:
 |
elimination
Hepatic metabolism.
Although almost all liver metabolic
processes can be demonstrated in infants, even in the fetus, their
rates are by far slower in the newborn infant, especially if the infant
is preterm, than in the older child and adult. The maturation of
different phase 1 reactions (e.g., hydroxylation, deacetylation, and
oxidation) and phase 2 reactions (e.g., conjugation) may vary
extensively. For almost all the drugs studied that are metabolized by
the liver, clearance rates are slower at birth than later. The
deacetylation of rifampicin (9), glucuronidation of
chloramphenicol (8), oxidation of nalidixic acid
(7), and hydrolysis of clindamycin (23)
appear to be much slower in newborn infants. Similarly, clearance rates
for tetracyclines, phenobarbital, and phenytoin all appear to increase
with age (24). The direct result of this immaturity of
metabolic degradation of drugs is a prolonged drug half-life
(T1/2), which is inversely correlated with
clearance rate:
 |
Renal elimination.
Many drugs, including penicillins,
cephalosporins, digoxin, and the aminoglycosides, are eliminated
unchanged by the kidney. All of them are filtered in the glomerulus,
whereas some of them are also reabsorbed and secreted by the tubular
cell (e.g., penicillins, cephalosporins, and phenobarbital). At birth,
both the glomerular filtration rate (GFR) and tubular secretion
processes are reduced; filtration, however, is relatively more
developed. The preterm infant has fewer glomeruli than the full-term
neonate, who in fact has the same number as an adult
(25)(26)(27)(28). The maturation process of kidney
structure and function is associated with prolongation and maturation
of the tubules, increase in renal blood flow, and improvement of
filtration efficiency. In addition, blood flow is shifted from the
deeper to the more superficial nephrons. Improvement of the GFR depends
on both gestational and postnatal age. The rate of tubular secretion is
similarly impaired, owing to poor perfusion and undeveloped energy
supply. Generally, it takes 6–12 months for the various renal
functions to reach adult values.
Investigators often assume that drugs that are metabolically inactivated by the liver are not influenced by reduced renal function. However, in most cases the metabolite—whether inactive or active—is eliminated by filtration and (or) tubular secretion in the kidney. With chloramphenicol, for example, the conjugated metabolite is mainly secreted by the renal tubules. Reduced tubular secretion in preterm infants could result in increased serum concentrations of the inactivated (conjugated) drug, such that subsequent intestinal ß-glucuronidase hydrolysis of the inactive metabolite would lead to increased enterohepatic recycling of the parent drug.
Knowledge of these dynamic principles is crucial in planning a rational dose schedule for drugs. In general, the preterm infant will need lower doses or longer dose intervals (or both) than the full-term infant to maintain similar steady-state concentrations. In short, the clinician must consider both the degree of maturation of renal function and possible disease or drug-induced (e.g., by aminoglycosides) impairment of renal elimination of antimicrobial drugs.
|
Specific Problems Associated with Drug Administration in Neonates and Infants |
|---|
|
TopAbstractIntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
infusion rates, drug volumes, and injection sites
Roberts was the first to draw attention to the importance of the
infusion rate and portal of injection on the speed and completeness of
drug administration by the intravenous route (29). A
full-term neonate receiving 100–140 mL of water per kilogram of body
weight per day generally will require an infusion rate of 10–20 mL/h
in the absence of oral intake. A preterm infant weighing <1 kg often
will be receiving intravenous fluids at a rate of 3–5 mL/h. Gould and
Roberts demonstrated that, at an infusion rate of 3 mL/h, actual
infusion of drug begins 160 min after injection into the system close
to the reservoir and will be complete only after 12 h
(30). This delay in the administration of a drug may
compromise therapy. Drug administration will be faster when the portal
for injection is closer to the neonate. In addition, if serum
concentrations of the drug are being monitored, accurate timing will be
difficult. Attempts to increase the infusion rate are often not
possible: Infants in intensive care units receive an average of 10
intravenous doses per day (31) and therefore cannot
tolerate an excessive infusion rate. In the original observations
(30), the injection portals were not necessarily as close
as possible to the infant. This problem can be minimized by injecting
the drug at a closer site.
Delivery of drugs to neonates should utilize pumps calibrated to accurately deliver very small volumes. If desired, these small volumes can then be infused at a rate much higher than the maintenance infusion rate, thus guaranteeing rapid delivery in only a small volume of fluid. The volume in which the drug is prepared is another important factor (29): the greater the volume, the more time required to deliver the full dose. Consequently, flow rate in conjunction with dosage volume will dictate the time needed for full delivery of a given dose.
loss of drugs while changing intravenous sets
Gould and Roberts drew attention to the fact that, owing to
frequent intravenous set replacement, multiple-dose administration, and
slow infusion rates in neonates, as much as 38% of the total drug
doses intended could be lost in the discarded sets (30).
When a drug is discontinued, the physician should keep in mind the
amount of drug still in the line; flushing the set with drug-free
solution or changing the set will guarantee discontinuation of the drug
delivery.
filter devices
Filters are often connected to the intravenous tubing in neonatal
units to clean the infusion fluid of bubbles, particles, and
microorganisms. In 1975, however, Wagman et al. documented the binding
of antimicrobial agents to filter devices (32). In 1981,
Rajchgot et al. noticed that subtherapeutic serum concentrations of
gentamicin and cloxacillin were achieved in neonates given these drugs
intravenously through a site proximal to a filter chamber
(33). Subsequent analysis has shown that drugs delivered
by slow infusion rates tend to be sequestered in the filter according
to their specific gravity; the higher the specific gravity (e.g., that
for cloxacillin), the more likely is the drug to stay in the filter
chamber.
After these findings, the manufacturer devised a new filter chamber to permit full delivery of injected drugs even with slow infusion rates. To avoid accumulation of the drug in the tubing during slow infusion rates, one injects the dose into a specially designed syringe chamber that permits any extra fluid to drain into a small additional bag. Only after the infusion volume first pushes the piston back to its original place is the infusion fluid able to pass the unidirectional valve; this guarantees propagation of the full dose without dilution through the filter chamber. Full recovery of injected doses of gentamicin and cloxacillin has been documented in studies in vitro (34). In subsequent in vivo studies with 70 neonates serving as their own controls, we showed that concentrations produced by injecting the dose into the new device were comparable with those achieved by a direct intravenous bolus into the site nearest the neonate (35).
"dilution intoxication"
Neonates require very small doses of intravenous preparations. To
meet the required dose, one often must dilute the stock solution. For
example, if a neonate requires 2.5 mg of gentamicin from a 50 mg/mL
stock solution (adult preparation), then 0.05 mL of the stock solution
is required. Because dealing with such small injection volumes is
difficult, the physician often dilutes this dose, e.g., by adding
another 0.05 mL of isotonic saline. In such an instance, the dead space
of the tuberculin syringe is filled as the 0.05-mL saline "pushes"
gentamicin from the dead space into the syringe; the child is likely to
receive more than the intended 2.5 mg. In a recent study, Berman et al.
(36) demonstrated that, were digoxin to be diluted so, the
dead space would increase an intended dose of 5 µg to 12–18 µg
(average, 14 µg). Thus, dilution of a very small-volume dose to
increase accuracy results in potential overdosing. Similarly, a case of
morphine toxicity attributable to "dilution intoxication" has been
reported (37). Roberts has calculated the effect of
dilutional intoxication of several medications when prepared in this
routine way (38).
errors in administration of drug doses
The neglected problem of errors in administration of drug doses in
pediatric patients has gained some publicity in the last decade.
Perlstein et al. showed that ~8% of the calculations of drug volumes
from a stock solution were erroneous, in many cases 10-fold higher or
lower than the intended dose (39). These results were
confirmed in a subsequent study that demonstrated the existence of an
"accident-prone" subgroup of health personnel who tended to perform
significantly more computation errors than other staff members
(40). These subjects also had noticeably more errors in
the 10-fold range. Although this finding of 3–6% for 10-fold
calculation errors in a written test may seem excessive, a recent study
from Britain reported a comparable incidence of 10-fold errors in
administering intravenous acetylcysteine for acetaminophen toxicity
(41).
Computational errors are not the only explanation for dosing errors. We recently reported a case of identical twins who, because the written order was misread, received two 50-mg doses of gentamicin instead of 5 mg. All appropriate staff should be aware of the appropriate dosages for neonates. The resulting acute renal failure took months to resolve, and it is too soon to evaluate the twins for potential ototoxicity (42). Errors committed by parents of a sick infant are also likely if the intended dose is not clearly explained (42). Pharmacy errors are less likely; pharmacists appear to be more knowledgeable and better trained in these matters than are nurses and physicians (39).
Because medication errors may result in morbidity and even mortality, various solutions to this issue have been suggested [42],[t44]:
1) A written test for all personnel involved in the preparation of doses from stock solution. There is a good correlation between the results of these written tests (42) and the real-life occurrence rate [44]. Staff members who fail this test ought to be retrained before being allowed to prepare drug doses for infants.
2) Double-checking of calculations by two staff members. Statistically, if each person is computing the dose independently, the chance of an error should be decreased significantly; e.g., if each staff member has a 6% chance of making an error, the probability that both of them will make an error is 6% x 6% = 0.36%.
3) Use of standard tables of recommended dose and volume of stock solution of drugs given intravenously to neonates and small infants.
4) Preparation of patient–unit dose by clinical pharmacists, thus obviating errors by nursing staff and doctors. Although this system increases the financial burden, pharmacists are less likely to produce errors in their computations (41).
The common denominator of the specific problems delineated in this section is the very small doses of drugs given to neonates. Although the issues raised are not confined to antimicrobial drugs, antibiotics, in terms of number of doses, by far exceed all other medications used in the nursery (3). Only well-trained personnel familiar with the appropriate doses for this age group and with effective methods to calculate them and to deliver them intravenously will be able to guarantee optimal antimicrobial therapy to small infants.
|
Practical Guidelines for Therapeutic Drug Monitoring in Neonates |
|---|
|
TopAbstractIntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
sample volume
With blood volume of 80 mL/kg, a preterm infant of 500 g has
only ~40 mL of blood. Moreover, because of higher hematocrit in
infants, more blood has to be harvested to obtain similar volumes of
plasma. Repeated sampling has been associated with neonatal anemia;
therefore, only laboratories possessing micromethods with samples
limited to 75 µL or less should perform these tests in neonates. New
automated instruments usually require <75 µL of serum (including the
dead space of the sample containers). It is strongly recommended that
laboratories inform clinicians of the absolute minimal blood volumes
needed for therapeutic drug monitoring, and clinicians drawing blood
should obtain only the minimum volume their laboratory requires.
Moreover, the neonatal intensive care unit should maintain a record of
the amount of blood being drawn daily for each infant, so that
excessive blood loss can be avoided.
availability of tests
The most common tests necessitating therapeutic drug monitoring in
neonates are those for aminoglycosides and vancomycin (in suspected or
proven sepsis), theophylline/caffeine (in neonatal apneas), digoxin (in
congestive heart failure and supraventricular arrhythmias), and
phenobarbital (in neonatal seizures and ischemic encephalopathy). Of
these, only the digoxin assay should be available on a stat basis,
given the availability of an antidote in case of life-threatening
toxicity. Table 1
lists the drugs for which monitoring should be routinely
available for neonates.
|
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionPharmacokinetic AspectsSpecific Problems Associated...Practical Guidelines for...References |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  B. Magnani and R. Evans Codeine Intoxication in the Neonate Pediatrics, December 1, 1999; 104(6): 75e - 75. [Abstract] [Full Text]
|
|
|
|
 |
|
 C. A. Hammett-Stabler and T. Johns Laboratory guidelines for monitoring of antimicrobial drugs Clin. Chem., May 1, 1998; 44(5): 1129 - 1140. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
