Introduction

Depending on the criteria adopted, treatment-refractory schizophrenia occurs in anywhere from 30% to 60% of patients with schizophrenia.¹ Clozapine has superior efficacy in the subset of patients with schizophrenia considered refractory to antipsychotic therapy, actively suicidal, or violent.^2-5 Current guidelines recommend a trial of clozapine after a patient fails two trials of antipsychotic dosed adequately and for an appropriate duration (4 to 6 weeks).^5-7 Clozapine's combination of dopamine D2 receptor and serotonin 5HT2A receptor blockade is the prototype that the class of atypical antipsychotics was built upon; however, its binding profile is more extensive. Clozapine also has high affinity for α_1,2A-C-adrenergic, H₁-histaminic, 5-HT_2B,C- serotonergic, and M_1-4 muscarinic receptors. The unique pharmacology of clozapine may be responsible for both greater efficacy and a greater side-effect burden. This complex side-effect profile and the resulting complex monitoring requirements are often cited as the primary reasons for its underutilization.⁸ Beyond the well-known risk for agranulocytosis, there are numerous other adverse effects, ranging from bothersome to potentially fatal, that require close attention.^9,10 Bothersome side effects, such as sialorrhea, sedation, nocturnal enuresis, and orthostatic hypotension, can often limit the rate of titration and the tolerability of clozapine. More severe side effects include a dose-dependent increase in seizure risk, metabolic syndrome, myocarditis, cardiomyopathy, delirium, bowel obstruction, and paralytic ileus.

Clozapine is metabolized predominantly by N-demethylation, N-oxidation, and aromatic hydroxylation. The main cytochrome P-450 hepatic isoenzymes involved in its metabolism are CYP1A2 and CYP3A4. In vitro data also show that CYP2D6, CYP2C19, and CYP2C9 play a role.¹¹ Norclozapine (N-desmethylclozapine) and clozapine-n-oxide are the two primary active metabolites created through hepatic metabolism, with the former being more active than the latter.¹² The involvement of the P450 enzymatic system, as one would expect, means that clozapine has an extensive list of pharmacokinetic interactions. Some of these interactions are not always apparent and may not be routinely flagged by drug-interaction programs. These include induction of metabolism by the polycyclic aromatic hydrocarbons in cigarette or cannabis smoke, inhibition of metabolism by caffeine consumption, and other variables such as age, race, and gender.^1,13,14

It has been documented that smoking as few as eight cigarettes per day can completely induce the CYP1A2 enzyme and cause a reduction of approximately 50% in clozapine serum level. Potentially toxic serum clozapine levels can result from reversal of this enzyme induction after smoking cessation. This can occur either voluntarily, in the context of a formal smoking cessation program, or involuntarily, on admission to a smoke-free facility.¹⁵ It is also important to keep in mind the potential need for dose adjustment if the patient plans to return to smoking upon discharge from a smoke-free facility. Caffeine can appreciably inhibit CYP1A2 with the equivalent of three cups of coffee per day; a high percentage of psychiatric patients consume this much caffeine on a regular basis.^1,16 Considering that smoking and consuming caffeinated beverages are often positively correlated, the competing interactions can potentially result in unstable clozapine serum levels.¹⁷

Patient demographics, genetic variables, and health status may affect the pharmacokinetics of clozapine. The metabolism of clozapine can decrease as a result of progressive reductions in liver function in aging patients.¹⁸ Men tend to have a higher rate of activity of hepatic isoenzymes than women, and both may have variations in metabolism based upon single-nucleotide polymorphisms in the CYP450 genes.¹¹ These single-nucleotide polymorphisms tend to be distributed based on ethnic background.^19,20 Other aberrations between dose and level may arise because of a common side effect of clozapine: constipation. Delays in gastrointestinal transit time can cause the absorption of clozapine to lag. If the dose were increased in response to this delay, the resultant level would be higher than expected upon initiation of an appropriate bowel regimen.¹ Infection and inflammation have both been shown to raise serum clozapine levels.²¹ Human error may also contribute to variation in the reported levels. If the evening dose is taken less than 12 hours before the blood draw, the level may be falsely elevated. Alternatively, if the dose is missed completely, the level will appear low.¹² All of these individual variations need to be taken into account when interpreting serum clozapine levels.

Therapeutic drug monitoring (TDM) has been a component of psychotropic medication administration for many years, especially when prescribing lithium and tricyclic antidepressants.²² The primary goal of TDM is to maximize treatment efficacy while minimizing treatment emergent adverse events. TDM can also serve as a marker of treatment adherence, a method to evince underlying variations in metabolism, and a system to prevent overdose.²¹ Current guidelines recommend the TDM of clozapine in routine practice; however, unlike other agents for which TDM is commonplace, the inter- and intraindividual variations in clozapine metabolism, coupled with a lack of agreement on a therapeutic range, make its adoption a challenge.^5,6,22 There is also a poor correlation between dose and level for a given dosage. Past studies have shown that patients stabilized on a fixed daily dose of clozapine can exhibit between an eight- to 45-fold interindividual variation in serum clozapine levels. ^23,24 In our research presented at the 2014 College of Psychiatric and Neurologic Pharmacists annual meeting, we performed a retrospective chart review of serum clozapine and norclozapine levels from March 2012 to April 2014, in a population of treatment-refractory patients with schizophrenia (n = 21). Within this population we found only a 36% shared variance between the clozapine dose and the serum level.²⁵ The result is consistent with previous findings of approximately 30% variation reported in earlier studies.²⁵ This highlights the lack of translatability between dose and serum level that prevents the provider from accurately concluding that a given dose will result in a given serum level for an individual patient.

The aforementioned intrinsic variations in an already highly complex patient population can play out in a myriad ways in clinical practice. This requires in-depth knowledge to use clozapine serum levels to their full potential. For example, a clinically stable patient may present with no apparent signs or symptoms of toxicity but receive what is deemed a “critically high” serum level based on the reference range of the facility. In such a case, the knee-jerk reaction to decrease the total clozapine dose may place the patient at risk of relapse and clinical decompensation without necessarily decreasing their exposure to side effects. This is just one scenario, albeit one that arises frequently in the monitoring of serum clozapine levels. These pharmacologic variables make the TDM of clozapine and its interpretation highly cumbersome, but does the resulting information possess enough clinical value to justify the additional effort? This review will examine where we currently stand and where we are heading in regards to the clinical utility of serum clozapine and norclozapine levels and how to effectively incorporate them into routine clinical practice.

What Is a Therapeutic Clozapine Level?

The serum clozapine level threshold associated with higher rates of therapeutic response has been examined in numerous studies.^24,26-28 At present, the agreed-upon level above which patients are more likely to respond is 350 ng/mL, though some articles cite a range between 250 and 420 ng/mL.^12,21-23 Although some patients may respond at lower serum levels, current guidelines recommend achieving a level of at least 350 ng/mL before a patient is deemed a clozapine nonresponder.⁵ In a study of 29 patients with schizophrenia who were titrated to a clozapine dose of 400 mg/day, only 22% of those below a serum cut-off point of 350 ng/mL were designated responders (a 20% reduction in symptom rating) while those above the cutoff had a response rate of 64%.²⁸ In a follow-up study, treatment response was seen in five of the seven nonresponders once their serum clozapine levels were raised above 350 ng/mL.²⁹ These findings have been replicated by other groups with similar results.^24,26,27,30 It has also been suggested that some patients may be slow responders and require more time on the medication rather than a rapid titration to a potentially higher than necessary dosage.³¹

A prospective relapse prevention study of 23 outpatients with schizophrenia stabilized on clozapine monotherapy demonstrated that a 40% reduction in clozapine level resulted in a statistically significant increase in likelihood of relapse.³² Relapse was defined as an increase of three or more points on the psychosis items of the Brief Psychiatric Rating Scale or psychiatric hospitalization.³² These data highlight the utility of monitoring clozapine levels as a method to ensure proper optimization and prevent relapse. After analyzing the data from our clozapine clinic, we found that patients who were psychiatrically hospitalized had significantly (t = 2.942, df =19, P < .05) higher serum clozapine levels.²⁵ This could have resulted from clozapine dose increases in an unsuccessful attempt to treat patients who were relapsing to hospitalization.²⁵ Recording serum clozapine levels when patients are psychiatrically and medically stable may serve as a valuable reference should they decompensate in the future.

Although there is consensus on the lower limit of therapeutic efficacy for clozapine levels, so far an agreed-upon upper limit of therapeutic efficacy remains unclear.²¹ As stated previously, most studies suggest anywhere from 250 to 420 ng/mL as a minimum threshold for response; however, at levels above 450 ng/mL there seems to be a fall off in clozapine response that is not well delineated or studied.²¹ Some retrospective studies identify much higher levels (600 and 838 ng/mL) as nontherapeutic.^22,33 One such study reported diminishing rates of response at clozapine levels above 700 ng/mL.³⁴ The authors suggested that an inverse curvilinear concentration-response curve would best describe this plateau in clozapine response at higher levels. The tricyclic antidepressant nortriptyline has a similar concentration-response curve, which is explained by a greater anticholinergic side-effect burden outweighing additional benefit at higher serum levels.³⁴ The anticholinergicity of clozapine similarly increases at higher dosages. Therefore, titrating clozapine nonresponders up to these higher serum levels may result in diminishing gains in response and increased anticholinergic side effects, which would outweigh any additional symptomatic benefit.^34,35 Given that many studies examine linear and not curvilinear concentration-response curves, patterns of response for higher serum levels may be obscured.^11,21

Much of the available information examining higher levels of serum clozapine are sporadic case reports written about adverse reactions or overdose; therefore, it is difficult to determine their effect on clinical response.²¹ Focusing on the typical patient response may not answer the question of when to hold the titration in a nonresponsive patient and consider the poorly studied alternatives of clozapine augmentation or switching. In any case, increased use of rating scales, such as the Brief Psychiatric Rating Scale, in tandem with the monitoring of serum levels in clozapine clinics could lead to a rich data source that reflects everyday response patterns.

Upper Safety Thresholds for Serum Clozapine and Norclozapine

Elucidating a level above which serious side effects of clozapine are more likely to occur is an ongoing endeavor that has met with limited success. At the time of this writing, no consistent correlation has been found between serum clozapine level and agranulocytosis.¹¹ A dose-dependent increase in seizure risk was reported by Devinsky and associates in 1991.³⁶ The authors reported a seizure risk of 1% at doses less than 300 mg, 2.7% at doses from 300 to 600 mg, and 4.4% at doses above 600 mg. There has been little convincing evidence to support a relationship strictly between serum levels and seizure risk.^21,36 In order to study the risk of rare adverse events such as seizure or myocarditis, the sample size would have to be quite large.

Case reports detailing adverse events have shown high levels, drawn after the fact, in patients experiencing overdose, drug interactions, or serious illness.²¹ Most describe central nervous system side effects, such as delirium or seizure activity, over a wide range of levels from 1330 ng/mL to as high as 9100 ng/mL in overdose.^37,38 In clinical practice, many patients achieve what are considered high levels without similar untoward effects. One such patient, a 43-year-old woman with schizoaffective disorder reported by Greenwood-Smith and associates,¹² was maintained on clozapine 700 mg for 2 years with levels between 357 and 986 ng/mL. When a level of 1192 ng/mL was reported while she was still on the same dose and exhibiting no signs of toxicity, her total dosage was reduced by 50 mg, leading to a severe relapse and inpatient psychiatric hospitalization.¹² This highlights the importance of using clinical judgment in clozapine-level interpretation and viewing each case in the context of the patient's personal history and individual variations.

The Clozapine/Norclozapine Ratio

A limited number of studies have assessed the effect of the ratio of the clozapine parent compound to its metabolites, particularly norclozapine.^39-41 This ratio can provide practical information on the metabolism of clozapine. Because the half-life of the metabolite norclozapine is longer than that of its parent compound, a low ratio (<0.5) would suggest poor adherence in the previous 24 hours or rapid metabolism of the parent compound.¹ Alternatively, a high ratio (>3) could suggest inhibition of metabolism by a concomitant medication or saturation of metabolism.¹ Some evidence supports a better clinical response and fewer adverse events with higher clozapine:norclozapine ratios, even suggesting the addition of fluvoxamine to inhibit CYP 1A2 and further boost this ratio.^40,41 An interesting phenomenon is that the subset of patients who require a higher dosage of clozapine will invariably have higher clozapine:norclozapine ratios due to saturation of the CYP1A2.¹ A case report by Szegedi and associates⁴² presented a patient unresponsive to prior clozapine therapy and experiencing excessive daytime sedation limiting further dose increases. The patient was administered fluvoxamine 50 mg daily to inhibit clozapine metabolism, and thereby increase the clozapine:norclozapine ratio. The patient experienced a marked improvement in negative symptoms (anergia, social functioning, blunted affect) and absence of daytime sedation, even though the serum clozapine levels were higher with the combination therapy.⁴²

A 12-week open-label, randomized, prospective study by Lu and associates⁴¹ showed that increased serum norclozapine, rather than serum clozapine, was correlated with a significant increase in posttreatment body weight, body mass index, fasting blood glucose, and triglyceride levels. Patients in this study were maintained on either clozapine monotherapy or clozapine augmented with the antidepressant fluvoxamine.⁴¹ The adjunctive fluvoxamine was used as a means of shifting the clozapine:norclozapine ratio by inhibiting the metabolism of clozapine through CYP1A2. Norclozapine levels were, therefore, significantly higher in the clozapine monotherapy group than in the coadministration group.⁴¹ The authors suggest that a possible confound in this study was fluvoxamine augmentation having a direct effect on weight loss through modulation of corticotropin-releasing hormone (CRH) or CRH-like peptides, versus affecting weight by decreasing norclozapine levels.^4]

It is theorized that the manipulation of the clozapine:norclozapine ratio through targeted CYP enzyme inhibition may alter the subset of receptors affected.¹ Norclozapine has been shown to have higher antagonist activity at 5HT_1c, 5-HT_2C, and D_1,2 receptors than the parent compound with additional activity as an M₁-muscarinic partial agonist.^43-45 The more potent 5-HT_2C antagonism by norclozapine has been suggested as an underlying mechanism for causing increased clozapine-induced weight gain.^41,46 Other agents with high risk of metabolic adverse effects, such as olanzapine and risperidone, also share this 5-HT_2C antagonism. The partial M₁ agonism increases acetylcholine in the medial prefrontal cortex and hippocampus, which, in turn, can promote increased executive functioning and memory.^45,47 This property is unique to norclozapine, with clozapine having minor agonist activity at M₁ and other antipsychotics possessing M₁ antagonist activity.⁴⁷ A clear underlying mechanism for the variable response rates associated with different clozapine:norclozapine ratios has yet to be fully explained and could provide an interesting avenue of research with practical application.

Clinical Practice Implications and Recommendations

Looking toward the future, more prospective trials are necessary to assess the relationship between serum clozapine and norclozapine levels and therapeutic response and side effects. Routine monitoring of clozapine/norclozapine levels and response/adverse effects data in large collaborative studies or using central databases could help make therapeutic monitoring more clinically relevant. Cost may be a major impediment to this practice, as prices for the levels vary between centers. On the positive side, patients would not be inconvenienced by additional laboratory work as blood is already routinely drawn at least every 4 weeks for white blood cell and neutrophil monitoring. Also useful, as alluded to previously, would be obtaining and recording efficacy and adverse effect ratings on standard nonproprietary rating scales to make a repository of information that could be used for further study.

Routine monitoring of serum clozapine levels may lead to debate over the proper course of action to undertake when a level comes back questionably elevated. With no clearly defined toxic range for clozapine or norclozapine serum levels, if clinical signs of toxicity are absent, overreliance on the clozapine serum level could lead to poorer outcomes for patients.^12,21 However, with proper utilization these levels can potentially become a helpful tool. As it stands currently, the clear indications for the TDM of clozapine are signs of (1) central nervous system toxicity, which can be linked to serum levels; (2) the addition of a medication that can inhibit CYP1A2 or 3A4; (3) the withdrawal of a medication or compound (polycyclic aromatic hydrocarbons) that can induce clozapine metabolism; (4) concerns of patient medication nonadherence; or (5) a recurrence of schizophrenic symptomatology while maintained on a previously effective dose.^1,12,21,22 Psychiatric pharmacists can play a crucial role in interpreting clozapine levels. Their expertise in the pharmacokinetics of psychopharmacologic agents can help to maximize clozapine's full potential in a population desperate for effective treatments. If these criteria are followed, along with the comprehensive monitoring for signs and symptoms of side effects, clozapine serum levels can serve a valuable role in routine clinical practice.