Abstract
Introduction
According to the National Institute for Health and Clinical Excellence (NICE) guidelines for the treatment of schizophrenia, antipsychotic polypharmacy can only be justified in the context of cross-titration, to avoid relapse during switch over of medication [National Institute for Health and Clinical Excellence, 2009]. The addition of a second antipsychotic to clozapine is also cautiously recommended in treatment-refractory patients with an inadequate response to trials of at least two antipsychotics and clozapine.
Despite stringent guidelines and limited evidence of efficacy, antipsychotic polypharmacy regimens including and excluding clozapine have steadily increased over the past two decades [Gilmer et al. 2007]. It appears that coprescription arises out of the desire to improve symptoms and reduce adverse effect burden [Taylor et al. 2002]. Polypharmacy often inadvertently results from aborted cross-titration, possibly due to an observed improvement in symptoms or adverse effects. A recent UK audit of acute adult and psychiatric intensive care wards found that almost half of antipsychotic polypharmacy could be attributed to pro re nata (as required) prescribing, despite no evidence from randomized controlled trials (RCTs) to support this practice [Paton et al. 2008].
The paucity of empirical evidence supporting antipsychotic coprescription has led to the practice being labelled as a ‘dirty little secret’ [Stahl, 1999]. A recent meta-analysis evaluated the efficacy of antipsychotic polypharmacy versus monotherapy in individuals with schizophrenia [Correll et al. 2009]. Results from 19 parallel-design RCTs, many of which were conducted in China, demonstrated that antipsychotic cotreatment was associated with a superior therapeutic benefit compared with monotherapy. However, distinguishing between whether or not improvements observed were due to a potentiation or simple additive dosage effect is impossible. Furthermore, results were limited by significant heterogeneity and there was evidence of publication bias in support of positive studies.
Concerns regarding the safety of this practice have been raised. Indeed, there is evidence to suggest that polypharmacotherapy is associated with increased mortality [Waddington et al. 1998], adverse effects [Centorrino et al. 2004] and reduced cognitive function [Elie et al. 2010] compared with monotherapy. A cross-sectional study suggested that polypharmacy may increase the risk of metabolic syndrome [Correll et al. 2007]. However, further analysis attributed this difference to demographic and clinical risk factors.
Antipsychotic polypharmacy often results in high-dose prescribing. This could possibly reflect the treatment-refractoriness of patients coprescribed antipsychotics and may partially account for adverse effects observed with polypharmacy. Indeed, a recent inpatient audit found that almost 73% of combined antipsychotic regimens were high dose [Paton et al. 2008]. High-dose prescribing is strongly discouraged and associated with significant risks, namely QTc prolongation and sudden cardiac death [Ray et al. 2009].
In view of recent failed attempts at curtailing antipsychotic coprescription [Paton et al. 2008], it is important that prescriber reasons for initiating and continuing this practice are re-evaluated. Furthermore, studies to date have focused on short-term polypharmacy regimens. This study is one of very few to examine the clinical and adverse effects of long-term polypharmacy. We also aimed to investigate antipsychotic treatment prior to initiation of long-term polypharmacy in community and inpatients, and determined patterns of antipsychotic coprescription.
Method
Prescription charts across the South London and Maudsley (SLAM) NHS trust were reviewed during the last 2 weeks of January 2011. SLAM supplies 3,600 inpatients and community patients, of whom approximately 2880 (80%) are on antipsychotics. All patients coprescribed two or more antipsychotics excluding clozapine for at least 6 months were identified. Patient records were examined using the Electronic Patient Journey System to obtain demographic data, documented reasoning for coprescription and prior prescribing information. Clinical outcome and adverse effects were also obtained from notes. Where documented, information regarding prescriber considerations to remove one of the antipsychotics or switch to clozapine was included. Prescription of high-dose antipsychotic polypharmacy was calculated using aggregated percentages of British National Formulary (BNF) maximum doses. Patients receiving greater than 100% cumulatively were considered to be on high-dose medication.
Results
Patient demographics
Around 288 (10%) patients were receiving antipsychotic polypharmacy (versus monotherapy), of whom 38 (13%) satisfied the criteria above. A total of 31 patients (81.6%) were diagnosed with schizophrenia or schizoaffective disorder. Of the remaining 7 patients, 6 patients (15.8%) had a diagnosis of bipolar affective disorder and 1 patient (2.6%) had borderline personality disorder. Their mean age was 47.2 years (standard deviation = 10.7, range 26–71) and 26 (68%) were male. A total of 15 (39.5%) subjects were White British; 9 (23.7%) Caribbean; 7 (18.4%) Black British; and 3 (7.9%) African. The remaining 4 (10.5%) patients were classified as Black and White, Mixed Black, White and Black Caribbean and Pakistani.
Prior antipsychotic prescribing patterns
Figure 1 illustrates the total number of antipsychotics (excluding clozapine) patients had been trialled on before initiation of polypharmacy regimes.
Total number of antipsychotics previously prescribed (excluding clozapine).
Prior to initiation of antipsychotic polypharmacy, 9 patients (24%) had been prescribed two or more antipsychotics concurrently and 16 (48%) had been trialled on clozapine monotherapy. A total of 15 patients (39%) had received monotherapy trials of at least three antipsychotics, one of which included clozapine. Of the remaining 22 subjects (52%) who had never used clozapine, the drug was considered in 4 subjects (11%), 3 of whom refused due to concerns over potential adverse effects and the requirement of regular blood testing. No documented reason was found for the remaining subject.
Documented reasons for coprescribing
The reasons for coprescribing (as shown in Table 1) were documented for 29 subjects (76.3%). For 9 patients (23.7%), no reason was found.
Documented reasons for coprescribing.
Antipsychotic polypharmacy combinations
The duration that subjects had been maintained on antipsychotic polypharmacy ranged from 6 months to greater than 9 years. Typical–atypical and atypical–atypical antipsychotic combinations were prescribed equally to 34 patients (89%).
The sequencing of antipsychotic combination therapy was documented in 32 patients (76%). In three subjects, two antipsychotics were initiated concurrently. For 20 of the other 29 known combinations (69%), an atypical antipsychotic had been prescribed first and another antipsychotic added to it. A total of 34% of antipsychotics first prescribed were depots and an atypical antipsychotic was added in 79% of cases. Further details are given in Figure 2. It was unclear which antipsychotic was prescribed first in 6 subjects.
Sequence of prescribing.
High-dose antipsychotic prescribing
A total of 21 patients (55%) were on a high-dose antipsychotic regime. In 7 patients, one of the antipsychotic medications was already prescribed at high dose.
Clinical outcome
Clinical outcome of coprescribing was documented for 26 subjects (68%). An improvement in hyperprolactinaemia, tardive dyskinesia and sedation was observed in three patients (8%). Further details are provided in Figure 3.
Clinical outcome (n).
Adverse effects
A total of 25 patients (66%) experienced one or more adverse effects associated with antipsychotic polypharmacy (see Figure 4) and 12 of these patients (48%) were prescribed high-dose antipsychotic combinations.
Adverse effects.
Prescriber considerations to stopping polypharmacy
Prescribers considered discontinuing antipsychotic coprescribing in 23 patients (61%). For four of these patients, definite plans were made to switch to a clozapine trial (n = 2) or to gradually remove one of the antipsychotics (n = 2). For the remaining 19 subjects, it was decided that coprescribing should continue. Reasons for this were documented for 17 individuals (see Table 2). During the course of polypharmacy, the prescriber had gradually removed the original antipsychotic (olanzapine) in two patients, but it had been restarted following a deterioration in mental state.
Documented reasons for continuing to coprescribe antipsychotics.
Discussion
The main reason for initiating antipsychotic polypharmacy was to improve symptoms and clinical outcome, a finding that is in concordance with our earlier study [Taylor et al. 2002]. Patients’ prior antipsychotic prescribing histories varied amongst the sample. A considerable number had only tried 0–1 antipsychotics before initiating polypharmacy and less than half of individuals had been trialled on clozapine. These findings suggest that antipsychotic polypharmacy is not always used as the last resort when all other options have been exhausted. Furthermore, less than half of patients were trialled on at least two antipsychotics plus clozapine, implying that treatment resistance was not established for all other patients.
The reluctance to use clozapine has been mirrored by various studies, one of which found that clozapine treatment was theoretically being delayed by an average of 5 years [Taylor et al. 2003]. Indeed 65% of patients were prescribed concurrent antipsychotics before their first trial of clozapine.
High-dose prescribing was common in this study, with more than half of patients receiving cumulative doses of antipsychotic medication above BNF defined limits. Four patients were prescribed doses 100% above the licensed maximum, all of which involved amisulpride and high-dose quetiapine combinations. The use of high-dose quetiapine is not an isolated finding. Recently the suggestion that high doses of quetiapine is necessary for therapeutic effect has led to prescribers disregarding guidelines, despite the fact that the only available evidence to support this theory is from case reports [Sparshatt et al. 2008].
Adverse effects were common and occurred equally amongst patients prescribed medication within daily defined limits and those on high-dose combinations. Both serious adverse events (QTc prolongation and arrhythmia) occurred in high-dose patients, which is not altogether unexpected considering risk of cardiac events is dose related [Taylor et al. 2009]. Extrapyramidal side effects (EPSEs) affected the greatest number of patients in this sample, all of whom were receiving two atypical antipsychotics or atypical–typical combinations. This finding supports the theory that combining atypical with typical antipsychotics results in loss of atypicality, probably due to additional inhibition of striatal dopamine receptors [Kapur et al. 2001].
Clinical outcome was variable where documented, with the majority of benefit being observed for nonpsychotic symptoms. A reduction in psychotic symptoms was documented in a quarter of all patients. However, differentiating between whether the effect observed is due to the combination of drugs or the addition of the second drug is a major problem inherent to polypharmacy studies. In this study, the original antipsychotic was gradually removed to observe the effects of the added antipsychotic in two patients only. In both cases, trial discontinuation did result in a significant increase in psychotic symptoms, suggesting the effectiveness of the polypharmacy regimen as opposed to the added drug only. However, these findings must clearly be interpreted with caution.
Documented improvements in adverse effects were uncommon. In a patient with risperidone-induced hyperprolactinaemia and galactorrhoea, the addition of aripiprazole resulted in a reduction in prolactin levels (and associated symptoms) without necessitating a reduction in risperidone dosage. This positive finding has also been reflected in a large RCT [Kane et al. 2009], indicating the potential value of aripiprazole in a subset of patients suffering from antipsychotic- induced hyperprolactinaemia.
A significant proportion of prescribers considered stopping polypharmacy, possibly reflecting lack of efficacy, tolerability or concerns over long-term safety implications of coprescription. However, the practice continued in the majority of cases, mainly because of patient refusal to medication changes. Some prescribers also expressed fears that reverting to monotherapy would result in relapse and the worsening of psychotic symptoms. Indeed, rationalizing combined antipsychotic therapeutic regimens to antipsychotic monotherapy was associated with deterioration in almost a quarter of patients in a small study [Suzuki et al. 2004].
This study was limited by its retrospective design and small sample size. Documentation was often of poor quality and for many patients, information regarding clinical outcome was not recorded. When outcome and adverse effects were documented, validated rating scales were seldom used. The retrospective design dictated that there was no randomization or blinding of treatment, increasing the possibility of bias and confounding. SLAM’s catchment area has one of the highest incidences of psychiatric disease in England. Its nonrepresentative nature therefore prevents these findings from being fully extrapolated to the general population.
Electronic records only included patient notes since the year 2001. High variability in dosage and antipsychotic combinations, all of which have differing pharmacological profiles, meant that assessing the effects of antipsychotic polypharmacy as a group was flawed. The absence of a comparator monotherapy group was a significant limitation to this study. It is therefore not possible to infer that clinical and safety outcomes recorded were due to effects of polypharmacy and not related to uncontrolled factors.
It is fundamental that the long-term efficacy and safety of polypharmacy is established. However, it is not possible to assess the almost infinite number of antipsychotic combinations prescribed. Instead, specific combinations most commonly used or already possessing some empirical backing should be evaluated in double-blind RCTs.
When antipsychotic polypharmacy is used, it should be the last resort after adequate monotherapy trials of at least two antipsychotics and clozapine have failed and the prescriber is confident that adherence to medication has been satisfactory. The second antipsychotic should be gradually introduced, whilst frequently monitoring the patient for benefit and adverse effects. The patient should be regularly reviewed, using validated rating scales to assess psychopathology and neuroleptic adverse effects.
Conclusion
Antipsychotics were coprescribed largely in an attempt to improve symptoms and adverse effects in patients with inadequate response to monotherapy, despite a lack of empirical evidence to support this practice. Prior to the initiation of polypharmacy, many patients received one antipsychotic only, suggesting that coprescription is not always used as the last resort when all other therapeutic options have been exhausted. Some patients prescribed antipsychotic polypharmacy did appear to benefit, although the majority of improvements were attributed to nonpsychotic symptoms. Adverse effects were also common. Prospective RCTs of specific antipsychotic combinations are required to assess long-term efficacy and safety implications, resolving some of the controversies surrounding antipsychotic polypharmacy.
Footnotes
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Professor Taylor has received consultancies fees, lecturing honoraria and/or research funding from AstraZeneca, Janssen-Cilag, Servier, Sanofi-aventis, Lundbeck, Bristol-Myers Squibb, Novartis, Eli Lilly and Wyeth. Ms Grech has no conflicts of interest.