A Systematic Review of Clozapine Concentration–Dose Ratio from Therapeutic Drug Monitoring Studies in Children and Adolescents Treated with Clozapine for Mental Disorders

Abstract

Background:

Therapeutic drug monitoring of clozapine in children and adolescents has received insufficient attention. Calculation of concentration-to-dose (C/D) ratios from trough steady-state concentrations estimate drug clearance.

Methods:

A systematic electronic literature search was conducted in 3 article databases from inception until January 10, 2023, and articles reporting clozapine concentrations in children and adolescents were retrieved. The pharmacokinetic quality of the studies was assessed, and clozapine C/D ratios were calculated using the sample mean clozapine dose and concentration.

Results:

Of the 37 articles of potential interest, only 7 reported clozapine trough and steady-state concentrations. After excluding case reports and a study confounded by fluvoxamine, 4 studies on psychosis from Europe and the United States were included. The clozapine C/D ratios were similar to published adult values and ranged from 0.82 to 1.24 with a weighted mean of 1.08 ng/mL per mg/d. The weighted means were 334 mg/d for the dose and 380 ng/mL for the concentration. The stratified analysis of the weighted mean clozapine C/D ratios from 2 studies showed lower values in 52 male (1.05 ng/mL per mg/d) than in 46 female (1.46 ng/mL per mg/d) children and adolescents, with values similar to those reported for European adult nonsmokers. Two female adolescents had high clozapine C/D ratios (2.54 ng/mL per mg/d), an Asian American patient with borderline obesity and a patient with intellectual disability with low dosage (mean = 102 mg/d) and concentration (mean = 55 ng/mL).

Conclusions:

Reports on clozapine therapeutic drug monitoring in children and adolescents are limited in number and quality. Future studies should focus on basic pharmacokinetic issues, such as stratification by sex, smoking, and relevant comedications with inductive or inhibitory properties.

INTRODUCTION

Clozapine is an atypical antipsychotic effective in the treatment of refractory schizophrenia, schizoaffective disorder, and bipolar disorder in adults.^1–3 In children and adolescents, its use is supported by the recent literature^4,5 that has found significant improvement in negative⁶ and positive symptoms.^7,8 Clozapine is superior to other atypical antipsychotics in psychosis.⁵ In low-to-moderate doses, clozapine seems to have clinical benefits in adolescents with other diagnoses such as bipolar disorder or intermittent explosive disorder with careful evaluation of the individual risk-to-benefit ratio.⁹ In a double-blinded, randomized, clinical trial, clozapine was more effective for externalization factors (particularly delinquency) and for global functioning than risperidone.¹⁰ However, in clinical practice, in children and adolescents, frequent clozapine use is restricted^11,12 due to the experience that children are more vulnerable and sensitive to adverse drug reactions (ADRs).^13,14 Studies on clozapine ADRs in children and adolescents include (1) a multicenter study in 8 United States (US) or Australian children's hospitals,¹⁵ (2) a clozapine protocol for children,¹⁶ (3) 2 systematic reviews,^17,18 and (4) a narrative review.¹⁹

The exact mechanism of action of clozapine is poorly understood.²⁰ Clozapine is associated with a transient D₂ receptor occupancy¹⁷ that appears sufficient to obtain an antipsychotic effect and is associated with fewer ADRs related to postsynaptic dopamine receptor blockade, such as extrapyramidal side effects²¹ and prolactin elevations.²² In addition, clozapine has the unique property of being the most selective antagonist for D₄ of any antipsychotic agent,²³ and its main active metabolite, norclozapine, has promuscarinic activity.²⁰ Clozapine also has an antagonistic effect on 5-HT_2A receptors that may help with negative and cognitive symptoms through the prefrontal release of dopamine.²⁴

Clozapine is metabolized through 3 pathways, the most important of which involves its demethylation to norclozapine, followed by oxidation to clozapine-N-oxide and glucuronidation.²⁵ Demethylation is mainly catalyzed by the cytochrome P450 1A2 (CYP1A2) enzyme, whose activity is influenced by sex and smoking and varies across ancestries.²⁶ Nonsmoking women need the lowest (estrogens inhibit CYP1A2), whereas smoking men need the highest minimum therapeutic dose of clozapine because tobacco smoking is a CYP1A2 inducer. Patients of Asian (of ancestries ranging from Pakistan to Japan) or American ancestry (indigenous or native Americans) require the lowest clozapine doses, whereas those of European ancestry require intermediate doses and those of African ancestry require the highest doses.^27,28

In adults with schizophrenia, the average minimum therapeutic dose required to reach a plasma or serum concentration of 350 ng/mL was used to compare individuals and groups based on clozapine clearance. The minimum therapeutic dose ranges are as follows: (1) 166 mg/d (nonsmoker women) to 270 mg/d (smoker men) in patients of Asian (or American) ancestry, (2) 236–368 mg/d in patients of European ancestry, (3) and 300–600 mg/d in patients of African ancestry (based on US recommended doses).²⁶ The elimination of five-sixth of norclozapine is renally,²⁹ which includes free norclozapine and its glucuronidation metabolites.³⁰ The glucuronidation metabolites can also be excreted by the biliary system to the gut.

Therapeutic drug monitoring (TDM) is used to improve patient response to important life-sustaining drugs, such as clozapine, and to decrease the risk of ADRs.^31,32 Steady state is usually defined as more than 5 half-lives from the last clozapine dosage change. The US clozapine package insert has a mean elimination half-life of 12 hours, but there is no description of how this half-life was calculated.²⁶ In the clinical setting, with repeated dosing, and in patients with average metabolism, the half-life of clozapine is approximately 24 hours or 1 day. Average metabolism refers to patient who have normal metabolism and exclude those who have too low metabolism called clozapine poor metabolizers (PMs) and too high metabolism called ultrarapid metabolizers (UMs). Therefore, at least 5 days (5 half-lives) from the last dosage change are required in most patients to achieve a steady state, but 10 days may be required in patients who are clozapine PMs.³³ Based on the short half-life proposed by the clozapine package insert, most published articles define trough time as at least 12 hours from the last dose. In 46 adults with schizophrenia, Jakobsen et al³⁴ found that the impact of deviations in clozapine TDM sampling time within the period of 10–14 hours postdose appeared to be of minor importance when looking at median percentage differences. Assuming a half-life of 24 hours, the trough state in a patient taking a single dose at night would occur before taking the night dose, when venipuncture is not easy. In clinical settings, trough concentrations are typically obtained early in the morning before the morning clozapine dose in patients dosed twice daily because there is little variation within a day owing to the much longer half-life of clozapine in most patients. In patients receiving only 1 clozapine dose, that is, at night, it is reasonable to give the evening daily dose earlier in the afternoon before the blood draw the next day, for example, at 5 pm, so that blood is drawn approximately 24 hours after the last clozapine dose when venipuncture facilities are still available. Future studies must compare 12-hour blood concentrations with true 24-hour trough blood concentrations in the patients receiving 1 evening dose of clozapine. Owing to the linear kinetics of clozapine, such an adjustment formula would be reliable and clinically useful. In individual patients, however, it is reasonable to target the same trough conditions by blood collection at the same hours of early morning and the last dose administration at the same hours. The stability of these 2 parameters is a reasonable way to define trough conditions in complex clinical settings.

If there are no changes on clozapine metabolism between the measures, clozapine follows linear kinetics in plasma concentrations from 150 ng/mL to saturation,³⁵ which means that there is a linear relationship between the clozapine doses and concentrations. Changes in factors that inhibit or induce clozapine metabolism, including comedications, smoking, or inflammation, can interfere with the linear kinetics. Thus, the relationship between clozapine dose and concentration can be represented by the C/D ratio,³⁶ which is stable in an individual as long as 2 conditions are met: (1) the absence of changes in clozapine metabolism–inducing or clozapine metabolism–inhibiting factors and (2) clozapine concentrations higher than 150 ng/mL.³³ Clozapine C/D ratios from trough steady-state concentrations in the range of linear kinetics are a good measure of clozapine clearance. As clozapine follows linear kinetics: (1) the relationship between clozapine concentration and dose is linear, and (2) the clozapine C/D ratio is constant in the same individual as long as the variables that influence clozapine clearance remain unchanged. The relationship between the C/D ratio and clozapine clearance has received limited attention. In one of the first pharmacokinetic studies conducted in 18 patients with schizophrenia using a mathematical model, Guitton et al³⁵ proposed that clozapine follows linear kinetics between plasma concentrations of 145 and 1411 ng/mL. This means that at very low concentrations or when the concentrations are close to saturation, clozapine may not follow linear kinetics. Thus, the clozapine international guideline²⁶ has proposed that when trough steady-state plasma clozapine concentrations are lower than 150 ng/mL, they should not be used to estimate the minimum therapeutic dose. There is no linear relationship between clozapine concentration and dose below 150-ng/mL concentration, and the clozapine C/D ratio is not stable.

Table 1 describes the typical clozapine C/D ratios in adults after stratification by ancestry and sex-smoking subgroups.^26,37–39 The minimum therapeutic dose of clozapine was calculated by dividing 350 by the clozapine C/D ratio.

TABLE 1.

C/D Ratios in Adults According to Ancestry (Modified From de Leon et al²⁶)

	Asian Ancestry37/Native Americans38		European ancestry39
	N	Clozapine C/D ratio	N	Clozapine C/D ratio
		(ng/mL per mg/d)		(ng/mL per mg/d)
♀ nonsmokers	252	2.10	218	1.48
♂ nonsmokers	227	1.68	340	1.18
♀ smokers	16	1.22	269	0.99
♂ smokers	137	1.30	546	0.95

According to the consensus range of 350–600 ng/mL,²² the therapeutic index is 1.7 (600/350 = 1.7), indicating that clozapine may have the narrowest therapeutic index among second-generation antipsychotics.²⁶ Clozapine PMs⁴⁰ or those taking inhibitors⁴¹ have higher clozapine C/D ratios. Patients taking inducers⁴¹ or those who are clozapine UMs have lower clozapine C/D ratios than average patients.^40,42

Clozapine PM status can be explained by genetic or nongenetic causes.⁴⁰ Clozapine genetic PMs have been described in adults with rare CYP1A2 mutations with no or little CYP1A2 activity.⁴⁰ Importantly, the CYP1A2 variants have different distributions in different DNA ancestral groups.⁴⁰ The CYP1A2*6 allele, present in approximately 1% of people of European ancestry, does not have enzyme activity⁴³ and may account for clozapine PM status.⁴⁰ Nongenetic causes of clozapine PM status include obesity,^37,44 coprescription of inhibitors of clozapine metabolism (including fluvoxamine and oral contraceptives⁴⁵), and infections that release cytokines that inhibit CYP1A2.⁴⁶ The inhibition by infections manifests clinically by elevations of the C-reactive protein concentrations.^47,48 These nongenetic causes are usually called phenoconversion because they are potentially reversible and are more frequently responsible than genetic causes for clozapine PM status.^37,49

The clozapine UM status, which leads to low clozapine C/D ratios, has been described in patients consuming inducers of clozapine metabolism. These include potent inducers, such as rifampicin, moderate inducers, such as phenytoin or carbamazepine, and occasional patients who are very sensitive to weak inducers, such as smoking and/or valproate.⁴² Attaining clozapine UM status during weak induction is rare and seen in no more than 1%–2% of clozapine-treated patients. A much more frequent cause of very low clozapine C/D ratios is nonadherence, which is very frequent and may occur in at least one-third of outpatients treated with clozapine.³³ Partial nonadherence usually manifests as a low clozapine C/D ratio for the corresponding clozapine dose. The clozapine-to-norclozapine ratio is calculated by dividing the plasma clozapine concentration by the norclozapine concentration; in a systematic review of 2317 adult patients across 19 studies, the weighted mean of the C/D ratio was 1.73 but the range varied widely from 1.19 to 3.37.²⁵ An inverted clozapine-to-norclozapine ratio, meaning a ratio of <1 in the case where plasma concentration of norclozapine is higher than that of clozapine, is another sign of partial nonadherence as long as the concentration is collected at trough and steady state, and the doses provide plasma concentrations higher than 150 ng/mL.^33,35 Nondetectable concentrations of clozapine and even more of norclozapine are a sign of full nonadherence. Different laboratories have different limits for nondetectable concentrations, but clozapine concentrations may take up to 7–9 days to go from a therapeutic to a nondetectable concentration, whereas in the same patient, norclozapine concentrations would take 12–14 days to become undetectable. Norclozapine has a longer half-life and stays in the plasma longer than clozapine.³³

In children and adolescents with schizophrenia, clozapine is typically started at very low doses and gradually increased based on symptom improvement. Clozapine is sometimes used to treat autoaggressive or heteroaggressive behaviors in children and adolescents with intellectual disabilities (IDs). Clinical experience indicates that in both adults and children with ID, clozapine doses tend to be lower than those in patients with schizophrenia; however, this has never been systematically studied.^26,50,51

Studies on clozapine TDM in children and adolescents have not received sufficient attention. Therefore, this study aimed to fill this information gap. To achieve this objective, we conducted a systematic review of the literature on clozapine concentrations to calculate clozapine C/D ratios in children and adolescents treated with clozapine.

METHODS

This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statements of quality.⁵²

Article Search Strategy

A systematic electronic literature search without language restrictions was conducted using MEDLINE/PubMed, TripDatabase, and Web of Science databases from inception until January 10, 2023. The search terms used were “clozapine/blood” with the following filters: Humans, Child: birth-18 years. The inclusion criterion for the studies was young people (children and adolescents) with a psychiatric diagnosis during stable clozapine treatment. We also included only studies that reported trough and steady-state clozapine concentrations.

Data Extraction and Outcomes

The selected data were extracted and entered by 1 author (S.J.-F.) and verified by another author (J.d.L.). All inconsistencies regarding study inclusion and data abstraction were resolved by consensus.

Quality Assessment of Included Studies

To assess the quality of the studies included in this review, we combined 2 methods. Kanji et al⁵³ developed a checklist to assess specific types of pharmacokinetic studies, that is, those reporting population pharmacokinetic models. The reviewed studies did not include population pharmacokinetic models; therefore, the Methods and Results sections of the checklists could not be applied to our studies. Table 2 presents information from the checklist developed by Kanji et al⁵³ in 3 sections: title/abstract, background, and discussion/conclusion.

TABLE 2.

Checklist of Information for Population Pharmacokinetic Studies^*

Model Reference and Items	Alfaro et al61	Bez et al59	Jiménez-Fernández et al63	Kumra et al7	Piscitelli et al62	Schulz et al58	Wohkittel et al60	Item Compliance Rate (%)
Title/abstract
Drug and patient population								100
Name of the drug(s) studied								100
Route of administration								0
Population studied								100
Primary objective								85.7
Major findings								71.4
Background
Pharmacokinetic data								0
Study rationale								85.7
Special objectives/hypotheses								71.4
Discussion/conclusion
Study limitations								42.9
Study findings								100
Study compliance rate (%)	8/11 72.7	5/11 45.5	5/11 45.5	9/11 81.8	9/11 81.8	8/11 72.7	9/11 81.8

^*Following the recommendations of the study in Clinical Pharmacokinetics by Kanji et al⁵³

Table 3 shows the methodology that we previously used to assess and calculate C/D ratios by combining multiple studies.^25,54–56 Table 4 shows the methodology that we previously used to assess the quality of laboratory assays measuring plasma and serum concentrations.^{25,36,54–56}

TABLE 3.

CLO and Total C/D Ratios in Studies With Mean Cs at Steady-State and Trough Conditions

Author	Country	Ancestry	Size		Concentrations (ng/mL)			CLO/NCLO	D (mg/d)	All CD Ratios Were Calculated by us (ng/mL per mg/d)		Confounding Factors
			N	TDM Samples	CLO	NCLO	Total CLO + NCLO			CLO	Total CLO + NCLO	M/F	Smoker	Inducers/Inhibitor*	Obesity or Inflammation
Alfaro et al61	US	ND	40 S/PD	30	333.5 ± 147.9 19 M: 318.2 ± 139.8 11 F: 360 ± 164.5	123.1 ± 66.6 19M: 122.5 ± 64.2 11F: 124.0 ± 73.9	458.6 19 M: 440.7 11 F: 484	2.73 M: 2.59 F: 2.90	269.9 ± 17	1.243 19 M: 1.18 11 F: 1.33	1.699 19 M:1.63 11F: 1.79	11/30	ND	ND	ND
Bez et al59	US	Asian American	1S	1	763	380	1143	2.01	300	2.54	3.81	0/1	ND	No	Borderline of obesity Inflammation: ND
Jiménez-Fernández et al63	Spain	European ancestry	1 ID	23	Mean 55 C is too low				95.6	0.59*		0/1	No	No	No obesity Inflammation: ND
Kumra et al7	US	European ancestry 2 Other 16	18 S	14	517.9 ± 287.8				487.5 ± 150.6	1.062		8/10	ND	No	ND
Piscitelli et al62	US	ND	11 S	11	378.34				350	1.081	1.918	9/2	ND	No	ND
Schulz et al58	Germany	ND	16 S	96	251 ± 140	263 ± 146	517	0.95	307 ± 160	0.817	1.674	13/7	ND	No	ND
Wohkittel et al60	Germany	ND	68 PD (89.7%) ID (6%) ED (4.3%)	68	387.0 ± 238.4 33 M: 349.4 ± 194.6 35 F:422.4 ± 271.5				318.8 33 M: 359.8 35 F: 280.0	1.214 33 M: 0.97 35 F: 1.51		33/35	64%	Antidepressant (citalopram, fluvoxamine) 38.2%	ND (mean BMI: 24.6 ± 5.1)
Weighted mean of 4 studies†	2 US and 2 Germany	NC	71 52 M 46 F	NC	380 338 M 408 F		NC		334 336 M 267 F	1.08 1.05 M 1.46 F	NC	52/46	NC	NC	ND

^*Only potent inducers, such as carbamazepine, phenytoin, and phenobarbital, were considered. Only potent inhibitors, such as fluvoxamine or ciprofloxacin, were considered. The possible effects of valproate, which may be both a mild inducer and mild inhibitor, were ignored. Bez et al,⁵⁹ stated that the patient was taking 4 mg/d of drospirenone, a progesterone derivative. The limited available data suggest that although estrogens can inhibit clozapine metabolism, progesterone derivatives do not inhibit clozapine metabolism.⁶⁷

^†71 patients with psychotic disorders and not confounded by fluvoxamine and obesity.

BMI, body mass index; C, concentration; CLO, clozapine; D, dose; ED, eating disorder; F, female; ID, intellectual disability; M, male; NCLO, norclozapine; NC, not calculable; ND, not described; S, schizophrenia; PD, psychotic disorder; US, United States.

TABLE 4.

Review of Methodology and Analytical Quality of Laboratory Studies Measuring CLO and NCLO

Author	Method	Description	Detection/Quantification		CV Intraday		CV Interday
			CLO	NCLO	CLO	NCLO	CLO	NCLO
Alfaro et al61	HPLC plasma	Yes	No/20	No/20	<15%	<15%	<15%	<15%
Bez et al59	ND serum	No	No	No	No	No	No	No
Jiménez-Fernández et al63	ND plasma	No	No	Not studied	No	Not studied	No	Not studied
Kumra et al7	HPLC serum	No*	No	Not studied	No	Not studied	No	Not studied
Piscitelli et al62	HPLC plasma	No	No	No	<10%	<10%	<10%	<10%
Schulz et al58	HPLC serum	Schulz et al64	No	No	<10%	<10%	<6%	<6%
Wohkittel et al60	HPLC serum	Yes	10/No	Not studied	<2%	Not studied	<2%	Not studied

^*They refer to a particular laboratory but do not provide a description of the method.

CLO, clozapine; CV, coefficient of variation; NCLO, norclozapine; ND, not described.

Combined Analysis to Calculate Mean C/D Ratio

This review expands on the theoretical work of Shapiro,⁵⁷ who proposed the term “combined analysis” for systematic reviews of naturalistic studies, providing a mean outcome that may reflect heterogeneous samples. In this systematic review of TDM studies, we performed a combined analysis to obtain the weighted mean clozapine C/D ratio. To calculate the weighted mean clozapine C/D ratio, the mean of each study was weighted by the number of subjects, and larger studies were assigned more weight.

RESULTS

Study Selection

Figure 1 shows our method for identifying 37 potential articles and the subsequent exclusion of 30 articles, yielding 7 articles describing 7 studies.^7,58–63 Seven other studies were excluded because we could not verify whether the concentrations were taken during trough and steady-state conditions.

FIGURE 1.

Flow chart.

Quality of the Studies

Overall, the quality of the included studies was low. Table 2 shows that according to the used checklist,⁵³ most articles had problems with the title/abstract and background, and only 3 of 7 studies discussed any limitations. Table 3 indicates that no article provided clozapine C/D ratios; therefore, we calculated them using the mean dose and mean clozapine concentration. Among the confounding factors, 5^{7,58,59,61,62} of the 7 studies did not describe smoking status, which is a weak inducer of clozapine metabolism and decreases clozapine C/D ratios, which is why clozapine C/D ratios need to be stratified by smoking. Fluvoxamine, an extremely potent inhibitor of clozapine metabolism, was discussed in 1⁶⁰ of the 7 studies, which probably contributed to significant increases in the clozapine C/D ratio in patients who took it. Therefore, patients taking fluvoxamine should be excluded from the calculation of the mean C/D ratio in average patients, as they may be phenoconverted to clozapine PMs.

Table 4 indicates that 2^59,63 of the 7 studies did not present the analytical method used to measure clozapine and the other 5^{7,58,60–62,64} studies used high-performance liquid chromatography (HPLC). Only 2 studies^60,61 provided limits of quantification or detection, and the stability was measured using the coefficient of variation. Three studies measured plasma concentrations^61–63 and 4 studies^7,58–60 measured serum concentrations.

Clinical Characteristics of the Samples

Two studies were double-blinded, randomized, clinical trials,^7,58 3 were observational naturalistic samples,^60–62, and 2 were case reports.^59,63 The eligible studies for this review included a total of 145 children and adolescents (1 included young adults up to 22 years of age),⁵⁸ 6 of the 7 studies included patients diagnosed with schizophrenia or psychotic disorders,^7,59–62 and 1 described a patient with ID.⁶³

Clozapine C/D Ratios

The clozapine C/D ratio was extremely high at 2.54 ng/mL per mg/d in an Asian American female adolescent with borderline obesity.⁵⁹ The patient with ID⁶³ received a low dose of clozapine (mean dose = 102 mg/d) and had a mean clozapine concentration of 55 ng/mL, which is low for reaching linear pharmacokinetics and providing an accurate estimation of the minimum therapeutic dose for clozapine.⁴²

To calculate weighted mean clozapine C/D ratios in patients with psychotic disorders, we excluded the case diagnosed with ID⁶³ and 2 of the samples that were severely confounded by fluvoxamine cotreatment,⁶⁰ a potent inhibitor that likely phenoconverts the patient into a clozapine PM and border-line obesity.⁵⁹ The remaining 4 samples^7,58,61,62 included 71 patients with psychotic disorders not confounded by fluvoxamine. In these 4 studies, which were from the United States or Europe, the clozapine C/D ratio ranged between 0.82 and 1.24 ng/mL per mg/d. The weighted mean clozapine C/D ratio was 1.08 ng/mL per mg/d and was confounded by sex and smoking status. Only 2 studies^58,60 stratified clozapine concentrations by sex and none provided stratification by smoking status. After stratification by sex, we calculated weighted mean clozapine C/D ratios of 1.05 ng/mL per mg/d in 52 male and 1.46 ng/mL per mg/d in 46 female children and adolescents.

DISCUSSION

Unfortunately, 7 studies were excluded because we could not verify whether the concentrations were at trough and steady-state conditions. This is the basic pharmacokinetic requirement for interpreting the plasma concentrations of clozapine or any other drug. Only 7 published articles provided pharmacokinetic data for clozapine at trough and steady-state conditions. All 7 studies were from the United States or Europe, and 1 case report involved Asian American female adolescent with borderline obesity.⁵⁹

Comparison with Clozapine C/D Ratios in Adults

Table 1 shows that among the 252 adult nonsmoker women of Asian ancestry, the mean clozapine C/D ratio was 2.10 ng/mL per mg/d. One pediatric patient of Asian ancestry had a higher clozapine C/D ratio but was also an outlier—she was a US teenager of Asian ancestry⁵⁹ with unknown smoking status and a clozapine C/D ratio of 2.54 ng/mL per mg/d. Assuming she was a nonsmoker, her clozapine C/D ratio would have been higher than the mean in Asian nonsmoking women. This higher value was probably because the patient was overweight.

After excluding the 2 case reports and the study contaminated by fluvoxamine, the remaining 4 studies in patients with psychosis were from Europe or the United States and their clozapine C/D ratios ranged from 0.82 to 1.24 ng/mL per mg/d. The study did not provide clozapine concentrations stratified by ancestry; therefore, we could not calculate the C/D ratios stratified by ancestry. The mean clozapine C/D ratios from studies in adult patients of European ancestry ranged from as low as 0.95 ng/mL per mg/d in smoker men to as high as 1.48 ng/mL per mg/d in nonsmoker women (Table 1). The lower value of 0.82 ng/mL per mg/d, derived from our analysis of 5 studies, appeared too low.⁵⁸ The study by Schulz et al included 16 outpatients but provided no clozapine concentration stratified by sex and smoking status. The low value of 0.82 ng/mL per mg/d can be explained by the following 3 possible reasons, listed from the most to least probable reasons: (1) many of these outpatients were partly nonadherent, (2) most patients were smoker men, and (3) many of the patients were consuming other potent inducers, such as phenytoin or carbamazepine. Nonadherence appears very likely because the data from Schulz et al⁵⁸ presented a mean clozapine-to-norclozapine ratio of 0.95 (Table 2), which is lower than 1.³³

Our stratified analysis of the clozapine C/D ratios from 2 studies showed that 52 male children and adolescents had a weighted C/D ratio of 1.05 ng/mL per mg/d, which was lower than the 1.46 ng/mL per mg/d value obtained in the 46 female children and adolescents. These values are similar to the published mean values for nonsmoking adult patients of European ancestry (men, 1.18 ng/mL per mg/d and women, 1.48 ng/mL per mg/d; Table 1). Unfortunately, only 2 studies^60,61 specified whether the patients were nonsmokers.

The Relevance of Oral Contraceptives in Female Adolescents Taking Clozapine

Stratified analysis of clozapine C/D ratios according to oral contraceptive use could not be performed, which is recommended if they are coprescribed in female patients taking clozapine because according to data from women in Finland⁶⁵ and the United States,⁶⁶ 10%–20% of young women take daily oral contraceptives. The estrogens found in oral contraceptives appear to have stronger inhibitory effects than endogenous estrogens on CYP1A2 and other CYPs.⁶⁷ Based on longitudinal case reports⁴⁵ in average adult patients, adding oral contraceptives is equivalent to decreasing clearance by approximately half and therefore requires cutting the clozapine dose in half; TDM provides a better basis for such dose adjustments.⁴⁵ For this reason, the international clozapine guideline for women taking oral contraceptives recommends using the slower titrations for patients with PM status in their respective ancestry groups.²⁶ To our knowledge, there are no studies on the use of oral contraceptives in female adolescents taking clozapine yet but may be reasonably extrapolated from the studies in adults. Nevertheless, when dealing with a female adolescent taking oral contraceptives, the best way to personalize dosing is using TDM.²⁶

The Relevance of Caffeine Intake in Adolescents Taking Clozapine

No study mentioned caffeine intake, but it is important because (1) caffeine intake may be associated with schizophrenia, (2) caffeine intake in adolescents and young adults living in Western countries is increasing because of the increased use of energy beverages containing caffeine, (3) caffeine has specific pharmacodynamic properties, and (4) caffeine has specific pharmacokinetic properties.

Studies in adults with schizophrenia suggest that caffeine intake in patients with schizophrenia may be associated with an attempt to reduce negative symptoms, as well as tobacco and alcohol use, whereas the association between heavy caffeine intake and schizophrenia may be explained by the association between schizophrenia and smoking.^68–70 Caffeine is mainly metabolized by CYP1A2. Because tobacco smoking induces caffeine metabolism, smokers tend to need 2 to 3 times more caffeine than nonsmokers to reach the same plasma caffeine concentration.⁷¹

Frequent caffeine intake in Western countries in adults with schizophrenia is explained by coffee or tea consumption, whereas in teenagers, by the consumption of energy beverages. As in the general population of Western countries, adolescents and young adults in many countries have shown dramatically increased consumption of caffeine-containing energy beverages; this has been better studied in the United States.⁷² Different beverages and brands of energy beverages contain different amounts of caffeine. In the United States, there is a comprehensive database of the caffeine content of drinks developed by the “Caffeine Informer”⁷³ but no independent information on the reliability of these estimations. There is also a lack of standardized self-report methodologies for investigating adolescent caffeine use, which makes it difficult to measure real caffeine consumption and limits generalization from prior studies.⁷⁴

From a pharmacodynamic point of view, caffeine is an antagonist of adenosine receptors and may increase norepinephrine and dopamine release.⁷⁵ Vainer et al described an adult patient on clozapine with psychotic exacerbations associated with heavy caffeine intake.⁷⁶

From a pharmacokinetic point of view, caffeine is metabolized by CYP1A2, and at high doses, particularly in nonsmokers, caffeine can act as a competitive inhibitor and can lead to clozapine intoxication.^77,78 Owing to these inhibitory effects of caffeine on clozapine metabolism, a sudden addition of high caffeine intake may be equivalent to a sudden increase in clozapine dose⁷⁹ and can lead to clozapine-induced inflammation, which further inhibits CYP1A2, leading to an increased risk of intoxication.⁸⁰

In conclusion, child and adolescent psychiatrists should inquire about the use of caffeinated beverages in patients receiving clozapine. No studies have provided clear instructions to clinicians regarding the amount of caffeine that could be clinically relevant.²⁶ Until prospective studies are available, we recommend caution in increasing or decreasing the daily caffeine intake of >1 cup of coffee (or 2 cans of caffeinated soda) in nonsmokers and >3 cups of coffee (or 6 cans of caffeinated soda) in smokers. For example, when a smoker taking clozapine increases caffeine intake by 3 cups of coffee (eg, from 2 to 5 cups per day), clinicians should monitor for increased ADRs due to an increase in clozapine concentration.²⁶ Similarly, heavy caffeine intake should be avoided during titration; if avoiding caffeine entirely is not possible, consumption should be consistently maintained, and the titration schedule for clozapine PMs of the respective ancestry group should be used.²⁶

Limitations

The results of this study should be interpreted in the context of the small sample size and additional limitations. First, we were able to stratify the data by sex in 2 studies but not by smoking. Second, other important confounders of clozapine metabolism, such as obesity, inflammation, or comedication with inhibitors/inducers, were mostly ignored by the included articles. Third, the studies provided no data on sexual maturation or clozapine concentrations corrected for body weight. In preadolescent children, it may be reasonable to consider the patient's weight when administering clozapine and interpreting clozapine TDM; thus, studies may need to adjust the C/D ratio by kilograms. Future studies should further examine this issue. Moreover, changes in clozapine metabolism may be triggered in adolescence. Exogenous estrogens from oral contraceptives are clinically relevant inhibitors of CYP1A2⁶⁷ and thereby of clozapine metabolism.⁴⁵ Although not well studied, lower clozapine metabolism in women after smoking stratification is likely explained by the inhibitory effects of endogenous estrogens on CYP1A2. Thus, clozapine TDM studies in children and adolescents should provide details on children versus adolescents, particularly in female children and adolescents. Fourth, most studies have paid little attention to the analytical quality of the determinations.

Strength

This is the first systematic review of clozapine TDM in children and adolescents. The strict pharmacokinetic approach used in the selection of articles in this review led to a small number of studies and the uncovering of very limited quality of the currently published evidence, from the pharmacokinetic point of view.

Future TDM Studies

The limited quality of the reviewed studies indicated areas of improvement for future studies on clozapine TDM in children and adolescents. These include (1) sample collection, (2) individual versus repeated samples, (3) assessment of the high risk of nonadherence, (4) consideration of clozapine PM status, and (5) clozapine UM status.

Child and adolescent psychiatrists should consider that clozapine samples should be collected under trough and steady-state conditions. Due to the variability of clozapine concentrations, it is better to calculate the mean of 3 values from the same individual rather than 1 single sample to estimate the clozapine C/D ratio of that individual.⁴² Child and adolescent psychiatrists should consider that, at least in adults, partial nonadherence is frequent in clozapine outpatients and at least one-third of them may not take clozapine consistently.³³ The influence of nonadherence has not been studied in children and/or adolescents taking clozapine; however, in many patients, the parents or caregivers may supervise adherence to clozapine. In our review of published studies, we found that nonadherence appeared very likely in the study by Schulz et al,⁵⁸ which provided a clozapine-to-norclozapine mean ratio of 0.95 (Table 2), which is lower than 1.

Clozapine PM status should be expected when patients are taking known inhibitors, such as fluvoxamine, oral contraceptives, or high intake of caffeine, or when they are obese. We recommend systematic measurement of C-reactive protein concentrations with clozapine TDM, which allows the ruling out of unknown inflammation, including any systemic infection. There is no commercial test for clozapine genetic PMs because the mutations are rare and vary according to ancestry.⁴⁰ Clozapine UM status can be explained by potent inducers; however, these drugs are rarely coprescribed with clozapine. However, it has been estimated that approximately 1%–2% of adult patients of European ancestry may be particularly sensitive to weak inducers such as smoking and/or valproate and can phenoconvert to clozapine UMs.⁴²

Future Studies Should Consider Pharmacodynamic Monitoring

TDM studies would benefit from including pharmacodynamic data, such as relevant biomarkers, objective symptom assessments, and the use of ADR ratings to complement TDM.⁸¹ The best example of a clozapine biomarker of pharmacodynamic activity is probably serum antimuscarinic activity.⁸² Both clozapine and norclozapine display antimuscarinic activity (except for muscarinic agonis at M1 and M4 receptors for norclozapine),²⁰ and their plasma concentrations may contribute to constipation. Moreover, patients may also be on anticholinergic medications to treat hypersalivation or other antipsychotics with clinically relevant antimuscarinic activity, such as olanzapine or quetiapine.⁸³ Thus, serum antimuscarinic activity may reflect all antimuscarinic activity from all drugs and may be a form of pharmacodynamic monitoring and better predict constipation than simply measuring plasma clozapine and norclozapine concentrations.

CONCLUSIONS

We identified 4 studies on children and adolescents from Europe and the United States with psychotic disorders who did not take fluvoxamine. Their clozapine C/D ratios ranged from 0.82 to 1.24 ng/mL per mg/d, which is similar to published values in adults. Even after stratification by sex, the values of the 52 male (1.05 ng/mL per mg/d) and 46 female children and adolescents (1.46 ng/mL per mg/d) were similar to published adult values. The limited quality of the available studies suggests that future clozapine TDM studies in children and adolescents must pay attention to the basic TDM principles highlighted in this article, including reporting trough and steady-state clozapine concentrations, smoking status, comedications that inhibit or induce clozapine metabolism, and inflammatory status.