Therapeutic Drug Monitoring in Neonates: What Makes them Unique?

INTRODUCTION: Therapeutic drug monitoring (TDM) refers to the interpretation of quantified drug concentrations in strategically timed samples of bodily fluids, with the aim to maximize therapeutic benefit, while minimizing toxicity. In essence, TDM criteria for neonates are similar to those for adults, but specific issues should be considered. This review focusses on the relevance of these specific issues: larger variability in pharmacokinetics (PK), and non-PK related factors, sampling opportunities, analytical techniques, therapeutic range. Specific issues: Larger variability in PK, and non-PK related factors in neonates compared to adults result in a less clear relation between the administered dose and the concentration measured. Sophisticated dosing regimens derived from population PK-models can partly overcome this variability, thereby reducing the need for TDM. Dosing can be further individualized using Bayesian forecasting as a tool for TDM. Besides PK related factors, concentrations of endogenous substances (e.g. immunoglobulin A, plasma protein) in neonates differ from those in adults, which may complicate interpretation of measured drug concentrations. Blood sampling opportunities in neonates are limited by the small blood volume and the need to minimize painful procedures. Dried blood spot sampling may be less invasive. This method has been facilitated by more sensitive analytical techniques, such as chromatography followed by mass spectrometry. For the same reason, saliva is gaining attention as an alternative non-invasive bodily fluid. Lastly, reference values for therapeutic ranges of drugs in neonates are mostly adapted from adult studies, although pharmacodynamics may be quite different in neonates. This review concludes with recommendations for future research on these specific issues.

Sildenafil and retinopathy of prematurity risk in very low birth weight infants.

OBJECTIVE: To examine the effect of sildenafil therapy on development of severe retinopathy of prematurity (ROP) requiring surgical intervention in premature infants. STUDY DESIGN: We identified premature infants who were discharged from Pediatrix Medical Group neonatal intensive care units from 2003 to 2012 and who received an ophthalmologic exam. We matched each infant exposed to sildenafil before first eye exam to three nonexposed infants using propensity scoring to control for differences in baseline infant characteristics. We evaluated the association between sildenafil exposure and development of severe ROP using conditional logistic regression. RESULT: Of the 57 815 infants meeting inclusion criteria, 88 were exposed to sildenafil. We matched 81/88 (92%) sildenafil-exposed with 243 nonexposed infants. There was no difference in the proportion of infants who developed severe ROP in the sildenafil-exposed vs nonexposed groups (17/81 (21%) vs 38/243 (16%), P=0.27). On adjusted analysis, there was no difference in severe ROP in the sildenafil-exposed vs nonexposed infants (odds ratio=1.46, 95% confidence interval=0.76 to 2.82, P=0.26). CONCLUSION: We did not observe an association between risk of severe ROP and sildenafil exposure before first eye exam in this cohort of premature infants.

Possible effects of repeated exposure to ibuprofen and acetaminophen on the intestinal immune response in young infants.

There has been an exponential increase in the frequency of immune deviations in young children. Consequently, research investigating environmental causes for this increase has become a Public Health priority. We have summarized the experimental observations and epidemiological data that could link repeated acetaminophen and ibuprofen exposure in early infancy to this increase. Recent observations on the maturational immunity of the intestinal sub-mucosal lamina propria underscore indeed the importance of prostaglandins (PGE2s). PGE2 appearing at this sub-mucosal level is a product of arachidonic acid metabolism mediated by type-2 cyclooxygenase (COX-2) situated on the membrane of many immune cells. Moreover, it seems that acetaminophen - like ibuprofen - also carries a non-selective inhibitory action on peripheral COXs, besides its central action. This inhibitory action of acetaminophen on COX2 only relates to physiological, low arachidonic acid concentrations. This explains the difference in anti-inflammatory effects. The impact of repeated inhibition of mucosal PGE2 synthesis due to COX-inhibitor exposure on maturational immunity has been demonstrated in animal experiments. Repeatedly exposed young animals do not develop tolerance to food antigens and exhibit autoimmune deviations. Several recent epidemiological studies have also reported on the magnitude of acetaminophen and ibuprofen exposure in children and the increase in immune deviations, it is important to better understand the potential negative impact of repeated inhibitions of prostaglandin synthesis by COX2s during infancy. Since acetaminophen and ibuprofen are commonly administered analgesics and antipyretics, a well-designed prospective strategy for pharmacovigilance and -epidemiology of COX-inhibitor exposure in infancy is urgently needed.

Emerging biomarkers and metabolomics for assessing toxic nephropathy and acute kidney injury (AKI) in neonatology.

Identification of novel drug-induced toxic nephropathy and acute kidney injury (AKI) biomarkers has been designated as a top priority by the American Society of Nephrology. Increasing knowledge in the science of biology and medicine is leading to the discovery of still more new biomarkers and of their roles in molecular pathways triggered by physiological and pathological conditions. Concomitantly, the development of the so-called "omics" allows the progressive clinical utilization of a multitude of information, from those related to the human genome (genomics) and proteome (proteomics), including the emerging epigenomics, to those related to metabolites (metabolomics). In preterm newborns, one of the most important factors causing the pathogenesis and the progression of AKI is the interaction between the individual genetic code, the environment, the gestational age, and the disease. By analyzing a small urine sample, metabolomics allows to identify instantly any change in phenotype, including changes due to genetic modifications. The role of liquid chromatography-mass spectrometry (LC-MS), proton nuclear magnetic resonance (1H NMR), and other emerging technologies is strategic, contributing basically to the sudden development of new biochemical and molecular tests. Urine neutrophil gelatinase-associated lipocalin (uNGAL) and kidney injury molecule-1 (KIM-1) are closely correlated with the severity of kidney injury, representing noninvasive sensitive surrogate biomarkers for diagnosing, monitoring, and quantifying kidney damage. To become routine tests, uNGAL and KIM-1 should be carefully tested in multicenter clinical trials and should be measured in biological fluids by robust, standardized analytical methods.

Optimising the management of fever and pain in children.

Fever and pain in children, especially associated with infections, such as otitis media, are very common. In paediatric populations, ibuprofen and paracetamol (acetaminophen) are both commonly used over-the-counter medicines for the management of fever or mild-to-moderate pain associated with sore throat, otitis media, toothache, earache and headache. Widespread use of ibuprofen and paracetamol has shown that they are both effective and generally well tolerated in the reduction in paediatric fever and pain. However, ibuprofen has the advantage of less frequent dosing (every 6-8 h vs. every 4 h for paracetamol) and its longer duration of action makes it a suitable alternative to paracetamol. In comparative trials, ibuprofen has been shown to be at least as effective as paracetamol as an analgesic and more effective as an antipyretic. The safety profile of ibuprofen is comparable to that of paracetamol if both drugs are used appropriately with the correct dosing regimens. However, in the overdose situation, the toxicity of paracetamol is not only reached much earlier, but is also more severe and more difficult to manage as compared with an overdose of ibuprofen. There is clearly a need for advanced studies to investigate the safety of these medications in paediatric populations of different ages and especially during prolonged use. Finally, the recently reported association between frequency and severity of asthma and paracetamol use needs urgent additional investigations.

Determinants of variability in clearance of exogenous compounds in neonates.

Although the general principles of disposition and elimination of exogenous compounds apply in neonates, their specific characteristics warrant a tailored approach. Children display maturation in drug disposition, and these maturational changes are most prominent in the first year of life. Elimination clearance is mainly either through metabolic or renal elimination clearance. Almost all phase I and phase II metabolic processes display ontogeny in a iso-enzyme specific pattern. Variation in phenotypic metabolic clearance is based on constitutional, environmental and genetics factors. In early life, it mainly reflects ontogeny, but other covariates may also become relevant. The impact of various covariates like postmenstrual age, postnatal age, disease state characteristics and polymorphisms are illustrated based or 'probe' drugs (paracetamol, tramadol, propofol) administered as part of their medical treatment in critically ill neonates. Renal elimination clearance in early life is low and almost completely depends on glomerular filtration. Despite this overall low clearance, interindividual variability is already extensive and can be explained by covariates like postmenstrual age, postnatal age, co-administration of a non-selective cyclo-oxygenase inhibitor or growth restriction. These findings are illustrated by observations on amikacin, vancomycin and cefazolin disposition in perinatal life. These maturational changes all have impact on the pharmaco/toxicokinetics and -dynamics. We hereby would like to extent the adagio of Paracelsus that 'all is toxic, it only depends on the dose' by making the point that the 'patient' is also relevant.

Developmental pharmacology: neonates are not just small adults...

Neonatal drug dosing needs to be based on the physiological characteristics of the newborn and the pharmacokinetic parameters of the drug. Size-related changes can in part be modelled based on allometry and relates to the observation that metabolic rate relates to weight by a kg 0.75 trend. Until adult metabolic activity has been reached, ontogeny, i.e. isoenzyme-specific maturation and maturation of renal clearance also contributes to drug metabolism, making isoenzyme-specific documentation of maturation necessary. Changes in body composition and ontogeny are most prominent in neonates. The body fat content (/kg) is markedly lower and the body water content (/kg) is markedly higher in neonates. These findings have an impact on the distribution volume of both lipophilic and hydrophilic drugs. Drugs are cleared either by metabolism or elimination. While the first is mainly hepatic, the second route is mainly renal. Both hepatic metabolism and renal clearance display maturation in early life although other covariables (e.g. polymorphisms, co-administration of drugs, first pass metabolism, disease characteristics) further contribute to the interindividual variability in drug disposition. Documentation of these maturational processes based on in vivo 'case' studies is of value since these drug-specific observations can subsequently be extrapolated to other drugs which are either already being prescribed or even considered for use in neonates by the introduction of these observations in 'generic physiologically-based pharmacokinetic' models.

Covariates of tramadol disposition in the first months of life.

BACKGROUND: Data on contributors to between-individual variability in overall tramadol clearance and O-demethyl tramadol (M1) formation in preterm neonates and young infants are limited. METHODS: A population pharmacokinetic analysis of tramadol and M1 was undertaken using non-linear mixed effects model. Covariate analysis included weight, postmenstrual age (PMA), postnatal age (PNA), creatinaemia, (cardiac) surgery, cardiac defect, and cytochrome (CYP)2D6 polymorphisms, classified by CYP2D6 activity score. RESULTS: In 57 patients (25-54 weeks PMA), 593 observations were collected. Tramadol clearance was described using a two-compartment, zero-order input, first-order elimination linear model. An additional compartment was used to characterize M1. Tramadol clearance at term age was 17.1 litre h(-1) (70 kg)(-1) (CV, 37.2%). Size (37.8%) and PMA (27.3%) contribute to this variability. M1 formation clearance (CL2M1, i.e. the contribution of M1 synthesis to M clearance) was 4.11 litre h(-1) (70 kg)(-1) (CV, 110.9%) at term age. Size and PMA were the major contributors to the variability (52.7%); the CYP2D6 activity score contributes 6.4% to this variability. CONCLUSIONS: Overall tramadol clearance estimates confirm earlier reports while CL2M1 variability is explained by size, PMA, and CYP2D6 polymorphisms. The CL2M1 is very low in preterm neonates, irrespective of the CYP2D6 polymorphism with subsequent rapid maturation. The slope of this increase depends on the CYP2D6 activity score. The current pharmacokinetic observations suggest a limited micro-opioid receptor-mediated analgesic effect of M1 in preterm neonates and a potential CYP2D6 polymorphism-dependent effect beyond term age.

Rectal acetaminophen does not reduce morphine consumption after major surgery in young infants.

BACKGROUND: The safety and value of acetaminophen (paracetamol) in addition to continuous morphine infusion has never been studied in newborns and young infants. We investigated the addition of acetaminophen to evaluate whether it decreased morphine consumption in this age group after major thoracic (non-cardiac) or abdominal surgery. METHODS: A randomized controlled trial was performed in 71 patients given either acetaminophen 90-100 mg kg(-1) day(-1)or placebo rectally, in addition to a morphine loading dose of 100 microg kg(-1) and 5-10 microg kg(-1) h(-1) continuous infusion. Analgesic efficacy was assessed using Visual Analogue Scale (VAS) and COMFORT scores. Extra morphine was administered if VAS was > or = 4. RESULTS: We analysed data of 54 patients, of whom 29 received acetaminophen and 25 received placebo. Median (25-75th percentile) age was 0 (0-2) months. Additional morphine bolus requirements and increases in continuous morphine infusion were similar in both groups (P = 0.366 and P = 0.06, respectively). There was no significant difference in total morphine consumption, respectively, 7.91 (6.59-14.02) and 7.19 (5.45-12.06) mug kg(-1) h(-1) for the acetaminophen and placebo group (P = 0.60). COMFORT [median (25-75th percentile) acetaminophen 10 (9-12) and placebo 11 (9-13)] and VAS [median (25-75th percentile) acetaminophen 0.0 (0.0-0.2) and placebo 0.0 (0.0-0.3)] scores did not differ between acetaminophen and placebo group (P = 0.06 and P = 0.73, respectively). CONCLUSIONS: Acetaminophen, as an adjuvant to continuous morphine infusion, does not have an additional analgesic effect and should not be considered as standard of care in young infants, 0-2 months of age, after major thoracic (non-cardiac) or abdominal surgery.

Life threatening central nervous system manifestations and hypothermia due to maneb intoxication in a child: a case report.

Maneb, manganese ethylene-bis-dithiocarbamate, is a fungicide pesticide used in the agriculture and bulb flower culture sector. Toxicological effects for humans have been reported in literature and are diverse. They vary from allergic reactions (dermatitis, conjunctivitis, and bronchitis), central nervous system effects (muscarinic, nicotinic, central and extrapyramidal) and renal toxicity (acute renal failure).A 7-year old girl was admitted to the pediatric intensive care unit because of status epilepticus. Physical examination showed respiratory insufficiency, convulsions, and severe hypothermia (32.5 degrees C). The patient was intubated and her convulsions were successfully treated with benzodiazepines. Except for a combined metabolic and respiratory acidosis and hyperglycemia, diagnostic investigations on admission (full blood count, electrolytes, liver and renal functions, cerebrospinal fluid investigation, toxicology screening of blood and urine for barbiturates and benzodiazepines, blood culture, herpes PCR, and a CT scan of the brain) were normal. Within 24 hours, there was a complete recovery of all neurological signs. Within 72 hours, the patient was discharged from the hospital. Liquid chromatography-mass spectrometric investigation of her blood showed amounts of maneb, which can explain all symptoms and signs. However, effects of this magnitude on the central nervous system have not previously been reported in humans.

Maturational changes in the in vivo activity of CYP3A4 in the first months of life.

OBJECTIVE: To document maturational changes of the in vivo activity of CYP3A4 in the first months of life. METHODS: The contribution of tramadol (M), O-demethyl tramadol (M1, CYP2D6-mediated) and N-demethyl tramadol (M2, CYP3A4-mediated) to the overall elimination of tramadol and the log M/M2 was assessed in 24-hour urine collections during continuous intravenous tramadol administration. Correlations with perinatal characteristics (postnatal age (PNA) and postmenstrual age (PMA)) were studied. RESULTS: Of the total amount of tramadol administered in a 24-hour interval to 25 neonates and young infants (PMA 25 - 53 weeks), 34.5% (SD 6.1) were retrieved in the urine as parent compound or metabolite in a 24-hour interval. This retrieved material consisted primarily of tramadol 79% (SD 18), M1 10% (SD 17) and M2 3% (SD 3.4). The contribution of M (r2 = -0.53), M1 (r2 = 0.46) and M2 (r2 = 0.16) to overall M elimination correlated with increasing PMA. The mean log M/M2 was 1.44 (SD 0.46) and there was an inverse correlation between the log M/M2 ratio and PMA (r2 = -0.43, 95% CI for r = -0.84 to -0.34, p = 0.0006) and PNA (r2 = -0.25, 95% CI for r = -0.78 to -0.16, p = 0.008). The maturational half-life of the log M/M2 ratio was 16 - 20 weeks. In a multiple regression model, PMA was the only significant variable accounting for the interindividual variability in log M/M2. CONCLUSIONS: PMA was found to be the most important maturational change determing the in vivo activity of CYP3A4. The activity of CYP3A4 is relatively delayed in the first months of life compared to the developmental changes in CYP2D6 activity described earlier, however, the overall weak correlations reflect that PMA explains only in part the interindividual variability observed.

Morphine in ventilated neonates: its effects on arterial blood pressure.

OBJECTIVE: To study the effects of continuous morphine infusion on arterial blood pressure in ventilated neonates. DESIGN: Blinded randomised placebo controlled trial. SETTING: Level III neonatal intensive care unit in two centres. PATIENTS: A total of 144 ventilated neonates. Inclusion criteria were postnatal age <3 days, ventilation <8 hours, and indwelling arterial line. Exclusion criteria were severe asphyxia, severe intraventricular haemorrhage, major congenital anomalies, neuromuscular blockers. INTERVENTION: Arterial blood pressure was measured before the start and during the first 48 hours of masked infusion of drug (morphine/placebo; 100 microg/kg + 10 microg/kg/h). OUTCOME MEASURES: Arterial blood pressure and blood pressure variability. RESULTS: There were no significant differences in overall mean arterial blood pressure between the morphine group (median (interquartile range) 36 mm Hg (6) and the placebo group (38 mm Hg (6)) (p = 0.11). Although significantly more morphine treated patients (70%) showed hypotension than the placebo group (47%) (p = 0.004), the use of volume expanders and vasopressor drugs was not significantly different (morphine group, 44%; placebo group, 48%; p = 0.87), indicating the limited clinical significance of this side effect. Blood pressure variability was not influenced by routine morphine analgesia (p = 0.81) or additional morphine (p = 0.80). Patients with and without intraventricular haemorrhage showed no differences in blood pressure (Mann-Whitney U test 1953; p = 0.14) or incidence of hypotension (chi(2) test 1.16; df 1; p = 0.28). CONCLUSIONS: Overall arterial blood pressure, use of inotropes, and blood pressure variability were not influenced by morphine infusion. Therefore the clinical impact of hypotension as a side effect of low dose morphine treatment in neonates is negligible.

O-demethylation of tramadol in the first months of life.

OBJECTIVE: Assess in vivo O-demethylation activity in the first months of life. METHODS: Time-concentration profiles of tramadol (M) and O-demethyl tramadol (M1) in plasma and urine were simultaneously collected in the first 24 h of continuous intravenous tramadol administration in neonates and young infants. M and M1 were determined by high performance liquid chromatography. Correlations between perinatal characteristics [postnatal age (PNA), postmenstrual age (PMA)] and the contribution of metabolites (M, M1) to overall tramadol elimination and to the plasma and urine log M/M1 were calculated. RESULTS: Plasma samples were available in 20/29 and complete 24-h urine collections were available in 25/29 neonates (25-53 weeks PMA). Mean plasma log M/M1 value (>4 h, n=86) was 0.8 (SD 0.4). A significant correlation between plasma log M/M1 and PMA (r=-0.73, P<0.0001) and PNA (r=-0.58, P<0.005) was observed. In a multiple regression model, only PMA remained an independent variable. Mean urine log M/M1 was 0.94 (SD 0.7). Significant correlations of the urine log M/M1 ratio with PMA (r=-0.73, P<0.0001) and PNA (r=-0.56, P=0.0035) were observed. In a multiple regression model with the urine log M/M1 ratio as dependent variable, only PMA remained an independent variable. The maturational half-life of the log M/M1 ratio in early neonatal life in the age range evaluated is about 12-16 weeks without plateau. CONCLUSIONS: O-demethylation activity was already observed in early neonatal life. A significant correlation with PMA was documented, but PMA can only partially explain the observed variability in O-demethylation activity. Polymorphism therefore likely already contributes to the interindividual variability observed in neonates.

Tramadol disposition in the very young: an attempt to assess in vivo cytochrome P-450 2D6 activity.

BACKGROUND: Tramadol is potentially a very useful pain relief medication in neonates and infants. It is primarily metabolized into O-demethyl tramadol (M1) by CYP2D6. Data concerning tramadol disposition and CYP2D6 activity in young infants are not available. METHODS: A population pharmacokinetic analysis of tramadol and M1 time-concentration profiles was undertaken using non-linear mixed-effects models (NONMEM), based on newly collected data on tramadol and M1 time-concentration profiles in neonates and young infants (n=20) and published studies on intravenous tramadol in children and adults. M1 formation served as a surrogate for CYP2D6 activity. RESULTS: Tramadol clearance was described using a two-compartment linear model with zero-order input and first-order elimination. Clearance increased from 25 weeks post-conception age (PCA) (5.52 litre h(-1) [70 kg](-1)) to reach 84% of the mature value by 44 weeks PCA (standardized to a 70 kg adult using allometric '1/4 power' models). The central volume of distribution decreased from 25 weeks PCA (256 litre [70 kg](-1)) to reach 120% of its mature value by 87 weeks PCA. Formation clearance to M1 contributed 43% of tramadol clearance, but had no relationship with PCA. There was a weak non-linear relationship between PCA and M1 metabolite clearance. CONCLUSIONS: Maturational clearance of tramadol is almost complete by 44 weeks PCA. A target concentration of 300 microg litre(-1) is achieved after a bolus of tramadol hydrochloride 1 mg kg(-1) and can be maintained by infusion of tramadol hydrochloride 0.09 mg kg(-1) h(-1) at 25 weeks PCA, 0.14 mg kg(-1) h(-1) at 30 weeks PCA, 0.17 mg kg(-1) h(-1) at 35 weeks PCA, 0.18 mg kg(-1) h(-1) at 40 weeks, 0.19 mg kg(-1) h(-1) at 50 weeks PCA to 1 yr, 0.18 mg kg(-1) h(-1) at 3 yr and 0.12 mg kg(-1) h(-1) in adulthood. CYP2D6 activity was observed as early as 25 weeks PCA, but the impact of CYP2D6 polymorphism on the variability in pharmacokinetics, metabolism and pharmacodynamics of tramadol remains to be established.

The pharmacological treatment of neonatal pain.

Optimal analgesia remains a major challenge for all involved in the care of (critically) ill newborns. The rapid changes in liver metabolism involving maturation of liver enzymes and renal clearance of drugs render (extreme) very low birth weight infants different from newborns of later postconceptional age with regards to the use of opioids such as morphine and fentanyl. Acute and/or procedural pain has been investigated fairly recently in randomized controlled trials and there are now guidelines. The long-term effects of opioid use in this particular age group of vulnerable babies await further evaluation.

Randomised controlled trial evaluating effects of morphine on plasma adrenaline/noradrenaline concentrations in newborns.

OBJECTIVES: To determine the effects of continuous morphine infusion in ventilated newborns on plasma concentrations of adrenaline (epinephrine) and noradrenaline (norepinephrine) and their relation to clinical outcome. DESIGN: Blinded, randomised, placebo controlled trial. SETTING: Level III neonatal intensive care units in two centres. PATIENTS: A total of 126 ventilated neonates (inclusion criteria: postnatal age <3 days, duration of ventilation <8 hours, indwelling arterial catheter for clinical purposes; exclusion criteria: severe asphyxia, severe intraventricular haemorrhage, major congenital anomalies, neuromuscular blockers). INTERVENTIONS: Plasma adrenaline and noradrenaline concentrations were determined in patients during blinded morphine (n = 60) and placebo (n = 66) infusion (100 microg/kg plus 10 microg/kg/h). RESULTS: Plasma concentrations at baseline (nmol/l with interquartile range in parentheses) were comparable in infants treated with morphine (adrenaline, 0.22 (0.31); noradrenaline, 2.52 (2.99)) or placebo (adrenaline, 0.29 (0.46); noradrenaline, 2.44 (3.14)). During infusion, median adrenaline concentrations were 0.12 (0.28) and 0.18 (0.35) and median noradrenaline concentrations were 2.8 (3.7) and 3.8 (4.0) for the morphine and placebo treated infants respectively. Multivariate analyses showed that noradrenaline (p = 0.029), but not adrenaline (p = 0.18), concentrations were significantly lower in the morphine group than the placebo group. Furthermore, noradrenaline concentrations were related to the length of stay in the neonatal intensive care unit. CONCLUSIONS: Continuous morphine infusion significantly decreased plasma noradrenaline concentrations in ventilated newborns compared with placebo treatment. The results of this study support the idea that routine morphine administration decreases stress responses in ventilated neonates.

Unlicensed and off-label prescription of respiratory drugs to children.

Many respiratory drugs are not available in formulations suitable for infants and toddlers. Efficacy and safety research is mostly restricted to older children. However, respiratory drugs are frequently used in children for common diseases like asthma, upper and lower respiratory tract infections, rhinitis and sinusitis. The unlicensed and off-label use of respiratory drugs in children were studied. A population-based cohort study was conducted by using the computerised medical records in the Integrated Primary Care Information project. The study population comprised a random sample from all children aged 0-16 yrs who were registered with a general practitioner in 1998. All prescriptions for respiratory drugs during the study period were classified according to their licensing and off-label status. The study population comprised 13,426 patients (51.7% male, median age 8.7 yrs), of whom 2,502 (19%) received 5,253 prescriptions for respiratory drugs in 1998. A total of 3,306 (62.9%) prescriptions concerned licensed drugs. Of the remaining 1,947 prescriptions (37.1%), 882 (16.8%) were unlicensed for use in children, and 1,065 (20.3%) were prescribed off-label. The 1-yr cumulative risk of receiving an unlicensed or off-label prescription was 45% among children with at least one prescription for a respiratory drug. This population-based study showed that a large proportion of respiratory drugs prescribed by the general practitioner are unlicensed for use in children, or licensed but prescribed in an off-label manner. Results have to be interpreted with caution because they may unjustly suggest inaccurate prescribing, whereas it may be difficult to treat children with respiratory symptoms and diseases, because for many respiratory drugs paediatric data on safety and efficacy are insufficient. These findings underline the importance of research on suitable formulations, dosages and efficacy of respiratory drugs in children.