Drug treatment patterns in bipolar disorder: analysis of long-term self-reported data

Drug treatment of bipolar disorder is a difficult challenge due to the recurrent, episodic, and heterogeneous nature of the disease. In recent studies, less than one third of patients in the USA and about half of those in Europe received monotherapy, while the balance received polypharmacy (Azorin et al. 2009; Baldessarini et al. 2008a, b; Goldberg et al. 2009; Hayes et al. 2011; Quante et al. 2010). Many factors have contributed to the increasing use of polypharmacy for bipolar disorder. A substantial proportion of patients do not respond adequately to monotherapy, and additional drugs are required to achieve acute stability and reduce relapses (Goldberg et al. 2009; Frye et al. 2000; Lin et al. 2006; Malhi et al. 2012; Post et al. 2010). There is increased recognition that even while euthymic, patients experience frequent subsyndromal symptoms (Bauer et al. 2010) and an impaired quality of life (Michalak et al. 2005). The drugs available for long-term maintenance have different efficacy for preventing mania and depression, suggesting the need for complimentary agents (Smith et al. 2007). Additionally, with the expansion of drugs available for bipolar disorder such as new-generation antipsychotics and antidepressants, and with psychiatric indications for anticonvulsants, there are numerous choices for the prescriber. Prior research on drug treatment patterns is primarily based on health plan claims (Baldessarini et al. 2007, 2008a) or physician prescribing records (Azorin et al. 2009; Goldberg et al. 2009; Hayes et al. 2011; Quante et al. 2010). This study explored drug treatment patterns in bipolar disorder using 6 months of daily self-reported data on psychotropic drug intake.

Data were obtained from an ongoing, long-term naturalistic study in which patients with bipolar disorder recorded mood, sleep, and drugs taken daily (Bauer et al. 2012). All study participants have a diagnosis of bipolar disorder by DSM-IV criteria, are at least 18 years old, receive pharmacological treatment, and agreed to use self-reporting software daily for 6 months. The study has minimal inclusion criteria to better represent the patient heterogeneity of clinical practice. The diagnosis of bipolar disorder was made by the prescribing psychiatrist in a clinical interview. The demographic profile of the patients in this study is similar to those who participated in five large studies of bipolar disorder (Bauer et al. 2012). All patients received pharmaceutical treatment as usual throughout the study. All participants were volunteers and provided written informed consent. The study was approved by each local institutional review board.

Data were collected from 513 patients with bipolar disorder, of which 450 (87.7%) returned sufficient data to be included in the study. The 450 patients returned a total of 99,895 days of data (mean 222.0 days). Of the 450 patients, 315 (70%) resided in the USA, 53 (11.8%) resided in Germany, 45 (10%) resided in Canada, and 37 (8.2%) resided elsewhere. The patients had a diagnosis of bipolar I disorder (272, 60%), bipolar II disorder (157, 35%), or bipolar NOS (17, 4%). With a mean age of 40 ± 11.3 years, 313 (70%) were female and 137 (30%) were male. The patients reported 2.4 ± 4.0 lifetime hospitalizations for bipolar disorder and an age of onset of 22.0 ± 10.3 years. For the 450 patients, Table 1 shows the number and percent of days taking any drug within a medication class. Of the 242 patients taking an antipsychotic, only 31 (12.8%) of patients were taking a typical antipsychotic for a total of 2.5% of days. Within each medication class, most of the 450 patients took a small number of specific drugs. As shown in Table 2, only eight antidepressants, five antipsychotics, and three benzodiazepines were taken by at least 5% of the patients.

The drug load analysis revealed that the more drugs taken within a medication class, the larger the mean PDD/DDD ratio (Table 3). Of the 450 patients, 236 were taking an antidepressant for ≥50% of days; 36% of their drug regimens included two or more antidepressants and had a PDD/DDD ratio of >1.5. Of the 450 patients, 176 were taking an antipsychotic for ≥50% of days; 23% of their drug regimens included two or more antipsychotics and had a PDD/DDD ratio of >1.5.

Of the 450 patients, 353 (78.4%) took a stable drug combination for ≥50% of the days. There were no significant differences between the demographic characteristics of the patients who did or did not take a stable combination. Considering the stable combinations, there were 52 unique combinations when categorized by medication class (Table 4) and 231 unique combinations when categorized by specific drugs (Table 5). The characteristics of the stable drug combinations are summarized in Table 6. Of the 353 patients with a stable combination, 293 (83%) were taking a mood stabilizer and 82.4% took three or less drugs. The stable drug combinations involved more than one medication class in 248 (70.3%) of the 353 patients. Of the 353 patients, those taking antidepressants were older (42.7 versus 37.2 years, p < 0.001) and more likely female (58.5% of females versus 43.8% of males, p = 0.010) than those not taking antidepressants. Those taking benzodiazepines were more likely female (27.4% of females versus 12.5% of males, p = 0.002) than those not taking benzodiazepines.

Patients were more likely to take a mood stabilizer than any other drug. About 80% of the 450 patients were taking a stable drug combination for 50% or more days. Consistent with prior research (Baldessarini et al. 2008a, b; Goldberg et al. 2009), the majority of patients were taking polypharmacy including 75% of those with a stable combination. About half the patients with a stable combination were taking an antidepressant, despite the ongoing controversy. There was no predominant treatment regimen, as the stable combinations consisted of 52 unique combinations by medication class and 231 unique combinations by specific drugs.

This study cannot distinguish whether or not patients were taking polypharmacy because they were refractory to monotherapy. In a study of US medical claims, about one third of patients with bipolar disorder received initial treatment with polypharmacy (Baldessarini et al. 2008a). Also, polypharmacy is routinely prescribed to inpatients (Greil et al. 2012; Vincenti et al. 2010; Wolfsperger et al. 2007). Recent research suggests that monotherapy with a mood stabilizer or second-generation antipsychotic is generally effective for acute mania (Goldberg et al. 2009; Grunze et al. 2009; Malhi et al. 2012; Tamayo et al. 2010), but polypharmacy is often required for acute depression, mixed states, and maintenance (Goldberg et al. 2009; Malhi et al. 2012; Tamayo et al. 2010). However, polypharmacy increases the risk of drug interactions and adverse events (Besag and Berry 2006; Mojtabai and Olfson 2010; Sandson et al. 2005). In a US national study, psychotropics were a common cause of emergency department visits for adverse drug reactions for those under age 65 (Lucado et al. 2011). The challenge of polypharmacy is that the gain in stability from adding another drug must outweigh the safety risks and side effects (Malhi et al. 2012).

In the current study, the use of more than one drug within a medication class greatly increased the drug load. Patients taking two antipsychotics, antidepressants, or benzodiazepines had about double the drug load of those taking one drug, similar to prior findings (Procyshyn et al. 2010; Tognoni 1999). This result suggests that within a medication class, increasing the dose of one drug produces a lower drug load than adding a second drug. In patients with epilepsy, anticonvulsant polypharmacy is associated with a much higher drug load than with monotherapy, and a higher drug load is associated with increased adverse events (Lammers et al. 1995; Deckers et al. 2001; Deckers 2002). In contrast to that in epilepsy, polypharmacy in bipolar disorder primarily involves drugs from more than one medication class, as with 70% of the patients taking a stable combination in this study. There are few studies of the relation between drug load as measured by the PDD/DDD ratio and adverse events for psychotropic drugs. For example, while persistent exposure to antipsychotic polypharmacy increases the risk of adverse reactions (Carnahan et al. 2006; Centorrino et al. 2004), the optimal measure to predict adverse events is unclear (Barr et al. 2010; Nosè et al. 2008). However, the drug load concept may be useful for improving tolerability of polypharmacy in bipolar disorder, and investigation of the relation between drug load and adverse reactions is indicated.

It is noteworthy that over 80% of the patients with a stable combination were taking three or less drugs. A relatively small number of drugs may facilitate long-term use, and there may be an upper limit on the number of daily drugs beyond which adherence decreases (Bauer et al. 2009; Robertson et al. 2008). Polypharmacy regimens that require frequent dosing, have dietary or time requirements, or are expensive may contribute to nonadherence (Cramer et al. 1989; Ingersoll and Cohen 2008). Patients with bipolar disorder consistently report that side effects contribute to nonadherence (Baldessarini et al. 2008b; Perlick et al. 2004). In this study, the patients without a stable combination were taking drugs, but inconsistently, with drug changes more than once a month. This finding is in agreement with prior evidence that nonadherence in bipolar disorder is typically partial or intermittent rather than complete (Baldessarini et al. 2008a; Lingam and Scott 2002).

There are many limitations to this analysis. All study patients had access to a psychiatrist and were taking medication. Furthermore, the very process of daily recording may improve medication adherence (van Berge Henegouwen et al. 1999). In clinical practice, most patients would not volunteer to keep a long-term daily record. This study may be less generalizable to other patients with bipolar disorder. With a naturalistic design, the patients in this study varied in severity, phase and course of the disease, and years of illness. The study included more females than males and more patients with bipolar I than bipolar II disorder. All data were self-reported and daily access to a computer was required. The strengths of this study include the use of prospective data from a relatively large number of patients and the entry of specific drugs and dose taken daily. However, the prescribed regimens and reasons for addition of new drugs were not known. Some drugs may be misclassified as to prescribing intent. For example, trazodone is considered an antidepressant but is frequently prescribed as a hypnotic agent. The analyses did not include other classes of psychotropic drugs such as stimulants, drugs taken for medical conditions, and over-the-counter preparations. Finally, many diverse factors that may influence medication selection such as physician prescribing habits, drug costs, regulatory influences, insurance plan formularies, pharmaceutical marketing, and patient preferences were not considered (Thistlewaite et al. 2010; Poulsen 1992).

Treatment of bipolar disorder is individualized in clinical practice as illustrated by 231 unique stable drug combinations for 353 patients. Other than the use of a mood stabilizer, no patterns were found within the stable combinations. A very large number of drug combinations may be prescribed. Considering only the most commonly taken drugs in this study (eight antidepressants, five antipsychotics, and four mood stabilizers), there are 32 possible combinations of mood stabilizer and antidepressant drugs, 20 possible combinations of mood stabilizer and antipsychotic drugs, and 160 possible combinations of mood stabilizer, antidepressant, and antipsychotic drugs. There will never be clinical trials of all possible combination therapies for bipolar disorder. Although controlled studies of polypharmacy are increasing, only limited evidence of effectiveness and safety is available to assist the clinician (Lin et al. 2006; Malhi et al. 2012; Beynon et al. 2009; Ghaemi and Ko, 2002). Further complicating drug selection, patients often have comorbid psychiatric and medical illnesses (Kupfer 2005; McIntyre et al. 2006), including those related to psychotropic drug use (McIntyre et al. 2006; Fagiolini et al. 2005). Some drug regimens in this study may appear inconsistent with current guidelines, but we assume that these are the best regimens for the circumstances of the individual patient. However, the wide variation in drug regimens and numerous possible drug combinations suggests that clinicians need more evidence to optimize treatment. While neuroscience research offers hope for a future with specifically targeted drugs (Insel 2009), the routine use of polypharmacy highlights the importance of clinical judgment in the current treatment of bipolar disorder.