You are on page 1of 5

Objective: An emerging body of evidence supports a role for dysfunctional purinergic neurotransmission in mood

disorders. Adenosine agonists have been shown to have properties similar to those of dopamine antagonists; there is a
well-characterized interaction between adenosine and dopamine receptors in the ventral striatum, and increasing
adenosinergic transmission has been demonstrated to reduce the affinity of dopamine agonists for dopamine receptors.
Allopurinol increases adenosine levels in the brain, and hence is hypothesized to reduce the symptoms of mania. Two
randomized, placebo-controlled trials administering add-on allopurinol to manic patients showed significantly greater
improvements in Young Mania Rating Scale (YMRS) scores for drug compared to placebo, while a more recent, relatively
small, add-on study showed negative results. Based on these data, our objective was to examine the efficacy of allopurinol
as addon treatment to mood stabilizers and/or antipsychotic agents in manic patients with bipolar disorder.
Methods: We performed a large, well-powered, multicenter, six-week, randomized, placebo-controlled trial of allopurinol
added to mood stabilizers and/or antipsychotic agents in 180 patients with bipolar disorder in an acute manic episode.
Results: Both groups showed improvement on the YMRS (effect size of 1.5 for placebo and 1.6 for allopurinol), with no
difference observed between groups on YMRS scores (t = 0.28, p = 0.78). There was no difference in the proportion of
patients who responded to treatment (defined as showing at least 50% improvement in YMRS score) between the two
groups (p = 0.92), or in dropout rates (p = 0.84).
Limitations: None of our patients received lithium. However, the side effects of lithium and its narrow therapeutic index
made the use of lithium less common and, therefore, our study results reflect common current clinical practice. In the
present study, we used a variety of antipsychotic and/or mood stabilizing treatments, to which we added allopurinol; one
might hypothesize that add-on allopurinol has a different effect in combination with different antipsychotic agents or
mood stabilizers.
Conclusions: The findings of this large, well-powered study do not support add-on allopurinol as a treatment for acute
mania. This study did not test the efficacy of allopurinol as monotherapy.

Bipolar disorder is associated with significant functional impairment, and accounts for a large fraction
of society’s expenditure for mental health. The World Health Organization identified bipolar disorder as the
sixth leading cause of disability adjusted life years worldwide among people 15–44 years of age (1), and bipolar
disorder is associated with both considerable functional impairment and mental health expenditure. Its
management – although sometimes effective – suffers from shortcomings including side effects, narrow
therapeutic indices, the need for regular serum level monitoring, and high financial costs, and is not sufficiently
effective in a considerable percentage of patients (2).
In 1991, Ferre et al. (3) showed that stimulation of A2a receptors potently reduces the affinity of D2
agonist binding sites within the plasma membrane of striatal neurons. They suggested that this A2a–D2
interaction may underlie the neurolepticlike actions of adenosine agonists and the enhancing effects of
adenosine antagonists, such as caffeine, on locomotor activity. There are data supporting the modulation of
dopamine activity through A2a receptors (4). The use of adenosine receptor antagonists was suggested as
having a beneficial influence in treating psychiatric disorders (4, 5).
Allopurinol inhibits xanthine oxidase, a key enzyme in purine degradation (6). Studies conducted on
animals suggest that allopurinol increases brain adenosine levels (7). Adenosine A2a receptors are found in the
cortex, hippocampus, and striatum, and they co-localize with dopamine receptors and regulate dopaminergic
neurotransmission, while adenosine A1 receptors decrease neuronal excitability and inhibit dopamine and
glutamate release (8). Moreover, an adenosine analog demonstrated antipsychotic-like effects in a mouse
model (9). There is evidence suggesting that adenosine has a pervasive inhibitory effect on neuronal activity
(10, 11), and evidence suggesting that both dopamine and adenosine present in the striatum and hippocampus
augment the effect of antipsychotic agents, which are known to be effective treatment for mania.
In humans, two double-blind, placebo-controlled trials showed an improvement in Young Mania
Rating Scale (YMRS) scores in patients suffering from acute mania treated with allopurinol compared to placebo
(12, 13). These studies had either relatively small sample sizes, and/or were limited to patients who were not
treated with neuroleptic agents. One relatively small (n = 27), recently published, double-blind, placebo-
controlled study showed no significant improvement. with allopurinol compared to placebo in attenuating
manic symptoms (14).
Based on these initial studies, we conducted a relatively large, double-blind, placebo-controlled,
multicenter trial comparing allopurinol to placebo added on to antipsychotic agents and/or mood stabilizers in
both hospitalized patients and outpatients suffering from acute mania. This study was designed as an acute
enhancement study, the hypothesis being that patients with acute mania receiving add-on allopurinol would
show a better response than patients receiving add-on placebo.

Materials and methods


Participants were inpatients or outpatients aged 18–65 years with a current DSM-IV-TR diagnosis of
bipolar I disorder in a manic or mixed episode, confirmed by the Structured Clinical Interview for DSM-IV (SCID)
(15). Participants were required to be treated with mood stabilizers and/or neuroleptic medications for a
period of between three days and two weeks. After giving informed consent, participants were administered
either 300 mg allopurinol or an identical cellulose placebo capsule daily for 42 days in a double-blind fashion,
in addition to their mood stabilizer and/or antipsychotic medication. Study medication (allopurinol/ placebo)
was administered orally twice daily (150 mg or matching placebo at each administration). The preparation of
the study medication and placebo was done centrally, and study sites were equipped with each patient’s
medication in individual bottles labeled with a code number, so study personnel were blinded to the treatment
assigned. The primary outcome measure was the total YMRS (16) score; the Clinical Global Impression for
Bipolar Disorder (CGI-BP) (17), Positive and Negative Syndrome Scale (PANSS) (18), and its subscales were used
as secondary outcome measures.
Sample size was estimated prior to the study based on a meta-analysis of the results of the two
published randomized controlled trials (RCTs) showing an effect size of 0.86 for improvement in total YMRS.
As it is unlikely that a large multicentre trial will replicate the large effect sizes seen in the two small RCTs, it
was more reasonable to assume a smaller effect size of 0.6 in the present trial. In order to achieve 90% power
with an effect size of 0.6, we needed 120 completers. As the subjects were in an acute manic state, we assumed
a 30% dropout, and therefore we recruited 180 patients. The number of patients screened, randomized, and
who dropped out of the study are shown in Figure 1.
Changes within each group over time were assessed using two-tailed paired t-tests. Between group
comparisons employed two-sample t-tests, the primary analysis being of the drug/placebo difference in change
score [baseline – last-observation- carried-forward (LOCF) endpoint], except for discontinuous demographic
characteristics, which were compared using chi-square analysis. Secondary analysis of the change from
baseline to LOCF endpoint were assessed by analysis of covariance of the LOCF endpoint, with baseline as
covariate and with a mixed model linear regression with baseline as intercept, and the slope modeled by a
linearized time using SAS 9.2 (Statistical Analysis System Institute Inc., Cary, NC, USA) – that is, baseline, drug
(drug versus placebo) + time 9 drug. The intention-to-treat population, defined as all subjects who received at
least one dose of study medication, was analyzed. Dropouts were accounted for using an LOCF method. As
LOCF can be over-conservative, a mixed models analysis was also made, in order to confirm the LOCF with a
different statistical method of estimating the missing values.
The trial was conducted at 25 centers in Romania according to good clinical practice after receiving
approval from local authorities regulating experiments in human subjects.

Results
Of the 180 participants enrolled in the study, 90 were randomized to allopurinol and 90 to placebo.
The two groups did not differ significantly in demographic or baseline clinical or treatment variables, including
concomitant medications (Tables 1 and 2). Mean baseline YMRS scores were 24.6 in the placebo group
[standard deviation (SD) = 5.6] and 25.4 in the allopurinol group (SD = 5.3) (t = _1.02, p = 0.31). Completion
rates were similar, with 74 out of 90 (82.2%) completing the six-week study in the allopurinol group and 75 out
of 90 (83.3%) in the placebo group (v2 = 0.04, p = 0.84). There were no significant differences in the reasons for
premature termination between the two groups (v2 = 0.59, p = 0.9).
The primary analysis was based on the intent-to treat population using an LOCF analysis to account for
non-completers. Overall, both the treatment and placebo groups improved significantly. Mean total YMRS
score improved by 11.0 points in the allopurinol group (SD = 8.65, t = _11.83, p < 0.001, effect size 1.612) and
10.6 points in the placebo group (SD = 8.56, t = _11.48, p < 0.001, effect size 1.497) by the end of the study.
Other clinical outcome measures such as CGI-BP and PANSS showed similar improvements. However, no
significant between-group differences were detected in the change in YMRS score (t = 0.28, p = 0.78) or in the
secondary outcome measures at the end of the study (see Table 2). There was no significant difference in the
rate of improvement in any of the outcome measures between the two groups, as measured by the decline in
YMRS, CGI-BP, and mean total PANSS scores (Fig. 2).
There was no difference in the percentage of patients who responded to treatment (defined as
showing at least 50% improvement in YMRS score) between the two groups, with 37.8% of responders in the
allopurinol group and 38.6% of responders in the placebo group (p = 0.92). The frequency of adverse events
was also similar between the two groups (38.9% for the placebo group and 41.1% for the allopurinol group, p
= 0.88). A detailed list of side effects for each group is shown in Table 3.

Limitations
The two previous studies showing a significant benefit of allopurinol in treating mania administered
allopurinol as an add-on treatment to lithium. None of our patients received lithium. However, the side effects
of lithium and its narrow therapeutic index make its use less common and therefore our study results reflect
current common clinical practice.
In the present study we used a variety of antipsychotic treatments, to which we added allopurinol.
This leads to two possible limitations. The first is the variety of the antipsychotic medications; one might
hypothesize that add-on allopurinol has a different effect in combination with different antipsychotic agents
or mood stabilizers. The second is the possibility that the use of effective antipsychotic and mood stabilizing
treatment may have caused a ceiling effect, masking the potential benefits of allopurinol as a monotherapy.
An additional limitation is that the present study was neither an acute monotherapy study nor a non-
responder study; it was an acute enhancement study, and it was negative in that regard.

Discussion
The current study was a large, six-week trial of allopurinol as add-on therapy to antipsychotic agents
and/or mood stabilizers in acute mania. YMRS, PANSS, and CGI-BP assessments showed that allopurinol
conferred no greater benefit than placebo for the clinical symptoms of mania (Fig. 2). We observed neither
adverse effects nor worsening of clinical symptoms in the treatment group (Table 3), both of which might be
important for patients with bipolar disorder suffering from gout or other diseases for which allopurinol is
indicated.
Previous studies administering allopurinol to manic patients included a study by Akhondzadeh et al.
(12) (n = 41) that examined allopurinol versus placebo as add-on treatment in patients receiving lithium and
haloperidol and found a small (effect size 0.26), statistically significant improvement for allopurinol compared
to placebo. Another study conducted by Machado-Vieira et al. (13) compared allopurinol, dipyridamole, or
placebo added on to lithium only, with no antipsychotic medications used during the study. The results showed
significantly higher remission rates in patients receiving allopurinol compared to dipyridamole or placebo (p =
0.008). The results of the current study replicate those of a third, relatively small (n = 27), double-blind,
placebo-controlled study conducted by Fan et al. (14), which showed no significant improvement with
allopurinol compared to placebo in attenuating symptoms of mania.
Our study population was much larger than the previous positive studies mentioned above and, in
addition, it turns out that none of the patients in the present study received lithium, whereas both of the
positive studies used allopurinol as an add-on to lithium.
The failure of allopurinol to separate from placebo may have risen from adequate treatment of
responsive patients with antipsychotic agents or mood stabilizers (hence, a ceiling effect). This might be a
possible explanation for the relatively high placebo response rates. In order to perform a study that would
eliminate this possible ceiling effect, a monotherapy study of allopurinol compared to placebo is required.
However, given that the current study results are clearly negative, we believe that such a study is not
appropriate at this time. Based on the fact that caffeine serves as an adenosine receptor antagonist, Fan et al.
(14) demonstrated that caffeine non-users treated with allopurinol showed greater improvement in YMRS
scores compared to caffeine users. As caffeine intake can pharmacodynamically antagonize the effects of
allopurinol, the failure to control for caffeine intake in this study might have affected the results of this current
study; future studies on allopurinol in mania should take caffeine consumption into consideration. Regarding
the dose used, Machado-Vieira et al. (13) (positive result) and Fan et al. (14) (no difference between drug and
placebo) administered allopurinol 600 mg/day. The current study, similarly to the study conducted by
Akhondzadeh et al. (12), examined the effect of 300 mg/day of allopurinol. It is possible that 300 mg/day is an
insufficient dose, although the studies administering higher doses were not definitive.
As mentioned above, the present study was a large-scale, multicenter study. It is not uncommon for
positive small, single-center, investigator-initiated exploratory trials not to be replicated in large, multicenter
studies (19, 20), and the current results underscore the concept that single-center, smaller studies need to be
replicated in large, multi-site trials.
In summary, our current findings do not support a clinical role for allopurinol, at the dose tested, as
an add-on treatment for acute mania.

You might also like