All antipsychotic medications have the potential to cause hyperprolactinaemia, as the inhibition of dopamine release effectively removes the negative feedback loop for prolactin secretion from the anterior pituitary gland [1]. Elevated serum prolactin levels have been shown to have profound effects on reproductive health and sexual function, including hypogonadism, decreased libido in both sexes, amenorrhoea and infertility in women and low sperm count and reduced muscle mass in men [2]. In addition, there is evidence to suggest a link between hyperprolactinaemia and reduced bone mineral density – a study of women with hyperprolactinaemia showed a 15–30% reduction in bone mineral density compared with healthy women [3]. There is also a suspected link between elevated plasma prolactin levels and breast cancer, although the causal relationship in this link remains to be established [1].

Although psychotic illness can be associated with sexual problems there have been considerable reports in the literature of sexual dysfunction related to antipsychotic use being a significant issue of patients, which may have a negative impact on the patient's quality of life and adherence to treatment [4, 5]. Clinical studies have demonstrated that with the exception of risperidone, atypical antipsychotics have significantly less impact on serum prolactin compared with conventional antipsychotics [6–10], and are also less likely to cause sexual dysfunction [11–13]. In addition Gopalakrishnan et al [5] has shown that it is possible to successfully treat antipsychotic sexual dysfunction with sildenafil including patients exhibiting raised prolactin levels. In this study 74% (23 patients) of the sildenafil treatment group felt that in future they would use this treatment compared to 26% of the placebo group. Whilst short-term clinical studies with aripiprazole have shown comparable serum prolactin levels to placebo [14], the effects of aripiprazole on serum prolactin and sexual function over a longer period of time have yet to be investigated.

The Schizophrenia Trial of Aripiprazole (STAR) study is a naturalistic effectiveness study conducted with the aim of providing critical information on the relationships between the condition, the medication, and the patients in their everyday care setting. Following an initial review of atypical antipsychotics in 2002, the UK National Institute of Health and Clinical Excellence (NICE) found a lack of 'real-world' data such as patient-reported outcomes, patient preferences, and treatment regimens based on the clinician's judgment [15]. For subsequent reviews, NICE has acknowledged that whilst randomised clinical trials are of restricted applicability to the general patient population due to their limited duration and number of variables measured, naturalistic studies such as STAR allow evaluations of treatment effectiveness using less selective patient criteria and a wider spectrum of effectiveness variables (such as safety, tolerability, symptom relief or worsening, patient preference for medication, quality of life, and cost-effectiveness) [16, 17].

The primary objective of the STAR study was to evaluate the effectiveness of a 26-week treatment of aripiprazole versus standard of care (SOC), which was a limited choice of three widely used and available antipsychotics (olanzapine, quetiapine or risperidone), in patients treated for schizophrenia in a community health or hospital-based outpatient setting. Outcomes research variables included quality of life and sexual function as secondary outcomes, and the safety evaluation included monitoring of serum prolactin levels. This paper will discuss specifically the findings with respect to sexual function in light of the serum prolactin analysis.

The demographics, primary efficacy results and safety findings of the STAR study have been presented in detail elsewhere [18]. This paper will present the results of the patient-reported secondary endpoint ASEX and the safety endpoint of serum prolactin levels.

Clinical effectiveness and outcomes

Primary efficacy outcomes

The results of the primary effectiveness analysis (IAQ, GCI-I and CGI-S scores) in the STAR study have been described in detail elsewhere [18]. However in summary the aripiprazole treatment group (n = 268) resulted in significantly better effectiveness than SOC (n = 254; P < 0.001; week 26 LOCF) as evidenced by the IAQ total score beginning at the first assessment point (week 4) and this was sustained through to week 26.

Patient-reported secondary outcome variables: sexual function (ASEX)

Both treatment groups experienced improvements in sexual function from baseline as assessed by ASEX; the improvement was significantly greater in the aripiprazole group compared with the SOC group (Figure 1). The difference in favour of aripiprazole was statistically significant from Week 8 (first assessment point) onwards. At 26 weeks, the ASEX Total Score had dropped by 1.88 points from 19.27 at baseline, compared with a drop of 0.92 from 19.7 at baseline in the SOC group (p = 0.032; OC).

Further analyses were performed on each antipsychotic taken. However, the results of this sub-analysis should be treated with care as the investigator could determine the antipsychotic taken in the SOC group. Figure 2 shows the results for each agent at week 26 (OC data) demonstrating that of the SOC group quetiapine had a greater improvement on ASEX Total score than olanzapine or risperidone. At baseline all groups had similar ASEX Total scores.

There was no interaction found between treatment and gender in the analysis of the ASEX Total Score. However when pooling the SOC and aripiprazole group, male patients exhibited an improvement in ASEX Total score of -1.53 (SE 0.31) from a baseline score of 17.93 compared to an improvement of -0.71 (SE 0.41) in the female patients (baseline score 21.90), suggesting that for the male patients sexual dysfunction was more affected by a change of medication.

In a post-hoc analysis, the change from baseline in ASEX Total Score was also adjusted for CGI-I, and a statistically significant treatment difference was still observed (-1.47 for aripiprazole versus -0.67; p = 0.024).

Safety

Adverse events

The results of the safety analysis have been reported in detail elsewhere [18]. In summary, the incidence of treatment-emergent AEs was similar in both treatment groups (77% in the aripiprazole group versus 71% in the SOC group). The most common AEs (≥ 10%) were insomnia, anxiety, headache, and nausea in the aripiprazole group, and anxiety and somnolence in the SOC group.

Serum prolactin levels

Baseline mean serum prolactin levels were similar in the two treatment groups (43.4 mg/dL in the aripiprazole group and 42.3 mg/dL in the SOC group, p = NS). At week 26, mean decreases in serum prolactin were 34.2 mg/dL in the aripiprazole group, compared with 13.3 mg/dL in the SOC group (p < 0.001; OC) (Figure 3). The maximum on-treatment effect on serum prolactin levels in the safety sample was a mean reduction of 28.8 mg/dL in the aripiprazole, versus a mean increase of 0.3 mg/dL in the SOC group (p < 0.001).

Potentially clinically relevant elevations in prolactin levels during the 26-week period were seen in a significantly higher proportion of patients in the SOC group (136 patients [54.4%]) compared with the aripiprazole group (45 patients [16.8%]).

When interpreting the results of these sub analyses, it should be considered that the investigators could choose the antipsychotic treatment in the SOC group. As can be seen in figure 4, the patients treated with aripiprazole, olanzapine and quetiapine showed a mean decrease in serum prolactin levels whereas patients treated with risperidone showed a mean increase in serum prolactin at week 26.

No interaction between treatment and sex was seen but male patients had a greater mean decrease in prolactin at week 26 compared with the female patients (22.67 mg/dL SE 2.15 vs 8.95 mg/dL SE 2.61, respectively).

Correlation between changes from baseline in ASEX Total Score and Serum prolactin levels

There was no meaningful correlation found between the changes from baseline in ASEX Total Score and prolactin (r = 0.12, p = 0.036). Moreover, further inspection of the data revealed some influential outliers. When the two outlying observations were removed from the analyses, it resulted in a loss of statistical significance.

The STAR study was designed to allow comparison of aripiprazole with SOC treatment in a setting close to daily clinical practice, and also to assess the patient and caregiver perspective on the treatment received. The primary outcome measure in the STAR study was the IAQ total score, which takes into account both the efficacy and tolerability of an antipsychotic medication and thus is a measure of effectiveness rather than merely efficacy [21]. The results of STAR demonstrated clearly that at all time points, treatment with aripiprazole was associated with significantly better effectiveness compared with SOC treatment, in the form of one of three atypical agents chosen by the clinician as optimal treatment for the individual patient.

The results of the ASEX evaluation in the STAR study showed that patients treated with aripiprazole experienced significantly less sexual dysfunction at the end of the treatment period compared with patients receiving SOC. These findings are consistent with the existing body of evidence for atypical antipsychotics, as discussed below. When considering the agents within the SOC group individually, olanzapine and quetiapine were associated with less sexual dysfunction at the end of treatment and risperidone was associated with the least improvement in sexual dysfunction, which is in agreement with the Intercontinental Schizophrenia Outpatient Health Outcomes IC-SOHO study [13].

Whilst sexual dysfunction has traditionally run the risk of being underestimated in clinical trials due to the lack of self-reporting and unwillingness to discuss the issue, there is increasing awareness amongst clinicians and allied mental health care professionals that sexual dysfunction is a common and distressing side effect of antipsychotic medication [22]. A survey carried out by the UK patient advocacy group Rethink (formerly the National Schizophrenia Fellowship) showed that 66% of patients who had experienced sexual dysfunction felt that this side effect was 'bad' or 'very bad' [23]. Men appear to suffer more than women [22]. However, in this study female patients showed less improvement in sexual function than their male counterparts who experienced greater improvement in ASEX score. There is evidence that sildenafil is able to address such antipsychotic related sexual dysfunction in men including those patients who also report raised serum prolactin levels. In the study by Gopalakrishnan et al, [4] 77% reported improved erections and 74% (24 patients) felt they would use the medication in future compared with 26% of the placebo group. However, it may be possible to improve sexual dysfunction in both male and female by changing antipsychotic medication rather than adding further medication to the treatment regimen of schizophrenia patients.

Prolactin plays a key role in the regulation of sexual behaviour and activity [27], and elevated serum prolactin levels associated with antipsychotic medication are known to cause erectile dysfunction, orgasmic difficulties, amenorrhoea and gynaecomastia [11, 28]. The serum prolactin measurements also demonstrated that patients in the aripiprazole treated group were more likely to have normal serum prolactin levels at the end of the treatment period compared with patients receiving SOC. Given previous risperidone data, it is not surprising that this was the only agent in the SOC treatment group associated with an increase in mean serum prolactin level at week 26.

The implications of sexual dysfunction for treatment compliance and the prevention of relapse, together with the potential link between primary and secondary prolactinaemia and breast cancer and/or decreased bone mineral density [29] makes serum prolactin an important endpoint in the assessment of antipsychotic therapy. A growing evidence base indicates that, with the exception of risperidone, the atypical antipsychotics are less associated with hyperprolactinaemia than conventional antipsychotics. Some agents, including quetiapine, olanzapine and clozapine, were shown previously to cause no significant or sustained increase in serum prolactin in adult patients [6–10]. More recently, a systematic review of short-term clinical trials with aripiprazole showed that the effect of aripiprazole on serum prolactin levels was comparable to that of placebo, and significantly lower than that of risperidone [14]. The IC-SOHO study demonstrated that a reduction in the incidence of elevated serum prolactin translates to a reduced potential for sexual side effects [13]. Switching patients from conventional antipsychotics or risperidone to an antipsychotic with less potential for prolactin elevation is therefore likely to be an effective measure for improving treatment compliance and reducing the risk of symptomatic relapse in patients experiencing sexual dysfunction during antipsychotic therapy.

One major limitation to the interpretation of the findings is that the STAR study employed an open-label design and both patients and physicians would have been aware as to whether they were receiving study medication or not. Moreover, physicians could choose the SOC medication. This may have introduced bias in subjective measures such as the ASEX, and caution should thus be exercised when interpreting these results.

In conclusion, the results of this study show that treatment with aripiprazole has the potential to improve not only the clinical effectiveness, but also to reduce the potential for sexual dysfunction, which is increasingly held to be one of the most distressing side effects associated with antipsychotic treatment, as well as the risk of severe side effects associated with hyperprolactinaemia such as loss of bone mineral density and osteoporosis. It is important that a holistic approach is taken to the management of schizophrenia to improve patients' quality of life and potentially compliance.