Dose requirement and prolactin elevation of antipsychotics in male and female patients with schizophrenia or related psychoses

Abstract

Aims

The aim of this study was to investigate the prolactin (PRL) secretion and the growth hormone (GH)-insulin-like growth factor I (IGF-I) axis in relation to gender and side-effects and dose of antipsychotic drugs during long-term treatment.

Methods

Forty-seven patients (21 men and 26 women), diagnosed with schizophrenia or related psychoses according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria and treated with different classical antipsychotics, were studied. Prolactin, GH and IGF-I were measured, as well as the serum concentration of the antipsychotics. In addition, body mass index (BMI) was calculated.

Results

The median daily, as well as the median body weight, adjusted daily dose of antipsychotic drugs was twofold higher in male compared with female patients. Antipsychotic-induced hyperprolactinaemia was more frequent and occurred at a lower daily dose of antipsychotics in women. Irrespective of sex, more than half of the patients had elevated BMI. Two patients had a slight increment in IGF-I levels, whereas the GH concentration, as assessed on a single occasion, was normal in all patients.

Conclusions

Patients on long-term antipsychotic therapy, with doses adjusted according to therapeutic efficiency, exhibited hyperprolactinaemia and elevated BMI, but no obvious influence on the GH-IGF-I axis. Furthermore, it appeared that the males required twice the dose of antipsychotic compared with females.

Introduction

It is well known that treatment with antipsychotics can cause hyperprolactinaemia [1, 2]. Through blockade of the dopamine D₂-receptors on the lactotrophs in the pituitary, the dopamine inhibiting effect on the prolactin (PRL) release is more or less abolished [3], leading to hyperprolactinaemia and hormonal side-effects such as menstrual disturbances, galactorrhoea and impotence [4–7]. In addition, previous studies show that hyperprolactinaemia due to antipsychotic therapy is more common in women than in men, despite lower dosages in women [8–10].

Antipsychotics are also expected to block hypothalamic postsynaptic dopamine receptors, and may thereby interact with the growth hormone (GH)-regulating systems in hypothalamus [11]. As a consequence, dysregulation in these systems may arise, resulting in a decreased hypophyseal GH secretion and a reduced production of the GH-dependent, insulin-like growth factor I (IGF-I) [12]. Deficiency of GH per se causes weight gain and changed body constitution with central obesity [13]. Therefore, GH deficiency caused by antipsychotics may contribute to weight gain during antipsychotic therapy.

Although antipsychotics block predominantly dopamine receptors of the D₂ class, these drugs also block other cell receptors, such as adrenergic and cholinergic receptors, to varying degrees [14, 15]. GH secretion in turn is regulated by complicated interactions between suprahypothalamic and hypothalamic neurotransmitters and neuropeptides, where dopamine, noradrenaline and acetylcholine play an important role [11, 16]. Thus, antipsychotic agents may have a complex influence on GH regulation.

Earlier investigators have studied short-term, but not long-term antipsychotic treatments, and found gender differences in dose requirement of antipsychotics [17], which in turn might lead to differences between men and women in side-effects.

The aim of the present study was to evaluate gender differences in hormone levels in patients on long-term treatment with classical antipsychotics. We studied the PRL secretion and the GH–IGF-I axis in relation to side-effects and dose of antipsychotic drugs.

Methods

Patients

The investigation included 47 patients, 21 men and 26 women, with diagnoses of schizophrenia, schizophreniform-, schizoaffective- or delusional disorder according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria [18] (Table 1). The median age of the patients was 42 (range 27–80) years, with no age difference found between the sexes (Table 2). The patients had no physical illness or medication except for the psychopharmacological drugs. They were chronically ill or remitted outpatients on treatment with different types of classical antipsychotics, either with a single or combined drug regimen. The types of antipsychotics used were haloperidol (n = 2), perphenazine (n = 18), remoxipride (n = 5), thioridazine (n = 3), zuclopenthixol (n = 9) and concomitant therapy (n = 10). All patients had been on medication with antipsychotics for more than 6 months and the median treatment period was 6.3 (range 0.5–39.5) years with no significant difference found between men and women. The antipsychotic doses were adjusted to obtain therapeutic efficiency and the patients had been on stable doses for at least the last 2 months before the beginning of the study. In addition, four of the patients (numbers 4, 12, 30, 36, Table 2) were on concomitant medication with the antidepressant agent clomipramine at a daily dose of 25, 25, 75 and 100 mg, respectively, and one patient (number 7, Table 2) with trimipramine, 50 mg daily. The hormonal analyses included PRL, LH, FSH, oestradiol, testosterone (bioactive) and GH together with IGF-I. Clinical signs and symptoms regarded as hormonal-related side-effects were evaluated.

Table 1.

Psychiatric diagnosis according to DSM-IV criteria [18] and confirmed by two psychiatrists, in 21 men and 26 women. The number of men and women who reported psychotic symptoms such as hallucinations, delusions or thought disorders are also given.

	Number of men		Number of women
Psychiatric diagnosis	All men	Men with psychotic symptoms	All women	Women with psychotic symptoms
Schizophrenia, disorganized type	2	0	1	1
Schizophrenia, paranoid type	7	5	8	2
Schizophrenia, residual type	0	0	2	0
Schizophrenia, undifferentiated type	11	6	10	1
Schizophreniform disorder	0	0	2	0
Schizoaffective disorder	0	0	1	0
Delusional disorder	1	0	2	1
Total number	21	11	26	5

Table 2.

Hormonal levels in 21 men (patient numbers 1–21) and 26 women, (patient numbers 22–47) on long-term treatment with antipsychotics due to schizophrenia or related psychoses. Serum levels of prolactin (PRL), growth hormone (GH) and insulin-like growth factor I (IGF-I) represent fasting morning values, analysed in the same sample.

Patient number	Age (years)	PRL (µg l−1) Ref.range men ≤ 15 Ref.range women ≤ 19	GH (µg l−1) Ref. range <14	IGF-I (µg l−1)	Age correlated s.d.-score
Men
1	27	8.1	0.01	301	0.71
2	29	8.7	0.21	265	0.36
3	31	33*	0.03	198	−0.58
4	32	7.9	0.09	211	−0.29
5	34	3.0	0.15	264	0.64
6	36	1.6	0.03	325	1.50
7	37	9.2	0.03	240	0.46
8	40	1.4	0.05	264	0.98
9	40	2.7	0.07	130	−1.58
10	41	13	0.70	314	1.67
11	42	10	0.02	152	−0.90
12	43	6.1	0.03	173	−0.37
13	43	27*	0.04	360*	2.28*
14	45	5.7	0.02	174	−0.24
15	51	9.8	0.03	207	0.74
16	55	11	0.14	215	1.10
17	56	23*	0.02	131	−0.63
18	64	14	0.70	232	1.90
19	70	9.4	2.21	147	0.60
20	71	5.5	0.22	190	1.58
21	71	5.3	2.25	149	0.70
Women
22	30	10	1.82	212	−0.39
23	31	23*	0.16	196	−0.62
24	35	11	5.48	223	0.08
25	36	21*	3.37	230	0.25
26	37	40*	0.03	185	−0.48
27	37	18	0.06	226	0.25
28	38	9.5	0.45	226	0.30
29	39	33*	0.06	197	−0.14
30	39	55*	0.45	271	1.02
31	39	14	0.96	435*	2.73*
32	41	10	0.99	139	−1.28
33	41	6.6	3.64	265	1.05
34	42	8.1	0.14	178	−0.33
35	43	21*	0.11	146	−0.99
36	46	40*	0.26	107	−1.94
37	46	7.7	3.16	194	0.21
38	50	17	0.03	204	0.63
39	51	24*	1.16	208	0.75
40	51	25*	0.06	248	1.39
41	56	8.6	0.11	180	0.52
42	56	9.8	0.19	154	−0.05
43	61	9.3	0.03	138	−0.15
44	72	37*	7.27	151	0.81
45	73	5.1	0.17	103	−0.52
46	77	22*	3.52	135	0.69
47	80	55*	2.07	158	1.43

^*= increased value.

The study was approved by the Ethics Committee of the Karolinska Institute, Stockholm, Sweden and the patients participated after giving informed consent.

Clinical and laboratory methods

Blood samples were collected in the morning, prior to breakfast and medication. Oral antipsychotic treatment was given 12–14 h and antipsychotic depot-injections 2–4 weeks before blood withdrawal. Serum was stored at −20 °C until analysis.

Prolactin, LH, FSH and total testosterone were measured by commercial radioimmunoassay (r.i.a.) kits (Wallac, Sweden and Diagnostic Products Cooperation, USA) and oestradiol by a fluoroimmunometric assay kit (Delfia Oestradiol, Wallac, Inc., Turku, Finland). To determine the bioavailable testosterone level, bioactive testosterone was calculated according to the following equation; BioTesto=T₁ × [1 + (0.541 × CA)], where T₁ denotes unbound testosterone concentration (as calculated from total testosterone) and CA plasma concentration of albumin [19]. Growth hormone was measured with a commercial kit, a two-site fluoroimmunometric GH assay based on two monoclonal antibodies (Delfia hGH, Wallac, Inc., Turku, Finland). Insulin-like growth factor I was determined according to the r.i.a. method of Bang et al. [20] and expressed as age-correlated s.d.-scores, based on samples from healthy men and women [21]. The detection limit was 8 µg l⁻¹. Including the extraction step, the intra-assay and interassay coefficients of variation were 4% and 11%, respectively.

Body mass index (BMI) was calculated as kg body weight/m², where m denotes the height in metres [22].

To compare the doses of the different antipsychotics, irrespective of the use of oral medication or depot injections, the patients' daily doses of the depot drugs were transformed into oral chlorpromazine (CPZ) equivalent doses in two steps [23, 24]. The daily dose was first adjusted for bioavailability according to current data; 20% for perphenazine [25], 70% for haloperidol [26] and 50% for zuclopenthixol [27]. The dose was then transformed to CPZ equivalent doses by the CPZ table for oral equipotency [28].

Serum concentration of the antipsychotic drug remoxipride was analysed by a high-performance liquid chromatography (h.p.l.c.) method with photometry detection as described by Nilsson [29] and thioridazine by the h.p.l.c. method described by Yasui et al. [30]. For serum concentration analyses of haloperidol, perphenazine and zuclopenthixol, published h.p.l.c. methods subject to minor modifications were used [31–33].

Statistical methods

As the different variables were assumed not to be normally distributed, the data are described as median and range. Accordingly, the nonparametric Mann–Whitney's and Fisher's exact tests were used in the statistical analyses. In addition, the strength of the linear relationship between two parameters was calculated by Spearman rank correlation (r_s). A P value of less than 0.05 was considered statistically significant. The statistical analyses were performed using Statistica® for Windows (Statsoft, Inc., Tulsa, Oklahoma, USA).

Results

Gender differences in antipsychotic dose requirement

The median body weight (bw) adjusted daily dose (expressed as CPZ equivalents) was twice as high in men compared with women, 3.2 mg kg⁻¹ bw and 1.7 mg kg⁻¹ bw, respectively. Accordingly, the median daily dose was 257 (range 70–2441) mg for the men and 116 (range 27–375) mg for the women. The sex difference in the daily dose as well as in the bw adjusted daily dose was statistically significant (P = 0.0001 and P = 0.0006, respectively).

Endocrine parameters and adverse drug reactions

Hormonal levels in the patients are given in Table 2. Three of the 21 men (14%) had elevated PRL levels. In these men (patients 3, 13, 17) a moderate increment of PRL was noted 33, 27 and 23 µg l⁻¹, respectively (reference value ≤15 µg l⁻¹). Two of the male patients with hyperprolactinaemia (numbers 3, 13) complained of impotence. The bioactive testosterone values for these two men were however, within the normal range, which was also the case for the other 19 men (data not shown).

Hyperprolactinaemia was statistically more frequent in the women than in the men (12/26[46%] vs 3/21[14%], P = 0.03). Three of the 12 women with hyperprolactinaemia (patients 23, 30, 36) had menstrual disturbances, in one case (patient 23) in combination with galactorrhoea, whereas six of the 12 hyperprolactinaemic women (patients 25, 26, 29, 35, 39, 40) reported regular bleeding and three (patients 44, 46, 47) were menopausal. Of the 14 women with normal PRL values only one (patient 24) had menstrual disturbances of unknown cause, one (patient 31) had premature menopause with elevated LH-FSH, four (patients 41, 42, 43, 45) reported menopausal amenorrhoea and the other 8 had normal regular bleeding. In addition, two of the women with elevated PRL (patients 30, 46) exhibited a distinct suppression in LH, FSH and oestradiol levels, whereas these levels were within the normal range in the other 10 women with hyperprolactinaemia (data not shown).

A slight increment of IGF-I was observed in two patients (numbers 13 and 31, Table 2). However, the median level of IGF-I (expressed as age-correlated s.d.-scores) was normal (0.41 [range −1.94–2.73]) and the GH concentration was below the upper reference limit in all patients (Table 2). Furthermore, no correlation was found between PRL and IGF-I (expressed as age-correlated s.d.-scores) either in the whole patient group or in the men and women separately.

BMI was elevated in 57% of the patients, in 48% of the men and 65% of the women, with no statistically significant sex difference (reference for men ≤ 27, for women ≤ 25). In addition, the median value of BMI was 27 for both men and women (range 19–40 and 19–41, respectively). Moreover, neither the patients with elevated BMI, nor the whole patient group showed any correlation between IGF-I (expressed as age-correlated s.d.-scores) and BMI.

Endocrine parameters and antipsychotics

The women had elevated PRL values at lower daily reference doses than the men (Figure 1). Accordingly, the ratio between the PRL level and the daily reference dose (expressed as CPZ equivalents) was significantly higher in the women compared with the men (P < 0.001) (Figure 2). A correlation between the daily reference dose and PRL was however, found only in the men (r_s = 0.59, P = 0.004), but not in the women or in the whole patient group. The daily reference dose did not correlate with BMI or IGF-I (expressed as age-correlated s.d.-scores).

Figure 1.

Prolactin (PRL) levels related to daily reference dose, expressed as chlorpromazine (CPZ) equivalents in 21 men (•) and 26 women (•) on long-term treatment with antipsychotics. Reference value of PRL ≤ 15µg l⁻¹ for men (dashed line), ≤ 19 µg l⁻¹ for women (solid line).

Figure 2.

Ratio between prolactin (PRL) level and daily reference dose of antipsychotic drug (expressed as chlorpromazine equivalents) in 21 men and 26 women on long-term treatment. The box plots indicate the median and lower and upper quartiles, and the whiskers show 10th and 90th percentiles. Outliers are also indicated (•).

Data on median values of PRL, IGF-I and serum concentration level for the different antipsychotics are given in Table 3. The women showed higher PRL values per drug serum concentration unit compared with the men, irrespective of antipsychotic drugs. Moreover, the median concentration of IGF-I was similar for the different antipsychotics used.

Table 3.

Median and range of prolactin (PRL), insulin-like growth factor I (IGF-I expressed as age-correlated s.d.-scores) and of serum concentration for the different antipsychotics, in 37 patients with schizophrenia or related psychoses on long-term monotherapy (x = women, y = men).

	Perphenazine n = 18 (x = 13, y = 5)	Zuclopenthixol n = 9 (x = 3, y = 6)	Haloperidol n = 2 (x = 2)	Thioridazine n = 3 (x = 2, y = 1)	Remoxipride n = 5 (x = 2, y = 3)
PRL (µg l−1)
median	9.7	9.8	16.4	33	9.5
range	1.4–55	2.7–18	7.7–25	17–33	1.6–27
IGF-I
median	0.48	0.60	0.80	0.46	0.36
range	(−0.99)-2.73	(−1.58)-1.90	0.21–1.39	(−0.14)-0.63	0.08–2.28
Serum concentration
median	2.0	15	6.1	1.1	2.2
range	1–11.5	5–27	2.2–10	0.2–1.6	1.0–6.9
	(nmol l−1)	(nmol l−1)	(nmol l−1)	(µmol l−1)	(µmol l−1)

Discussion

In this study, we found a twofold difference in median dose of antipsychotics between the sexes, male patients receiving doses twice as high as in women based on CPZ equivalents. This difference may be explained by sex-related differences in metabolic clearance [17]. Considering this aspect, the sex difference in dose requirement can be ascribed to a clinical adjustment of the medication in order to obtain optimal therapeutic effect in both men and women. Since more men than women reported psychotic symptoms despite medication (Table 1), it cannot however be excluded that also the psychotic disorders were more severe in the men, which may have demanded higher doses of antipsychotics, although the different types of psychoses were represented in similar frequencies in both sexes (Table 1). Therefore, our finding of a difference between the sexes in dose of antipsychotics may be partly explained by the better prognosis of women than men in schizophrenia [34]. To compare, gender differences in antipsychotic dose requirement during long-term treatment have not been studied thoroughly. However, in accordance with the finding in our study, Chouinard et al. [35] found in a short-term study that women required approximately half the doses of the antipsychotic fluspirilene compared with men, when prescribed doses were titrated to therapeutic efficiency.

As shown in earlier studies [8–10], antipsychotic-induced hyperprolactinaemia was also more common among women compared with men in our study. We found that 46% of the women, but only 14% of the men, had slightly to moderately elevated PRL levels, even though the women were receiving notably lower antipsychotic doses. Furthermore, the present results showed that symptom producing hyperprolactinaemia appeared in 16% of the fertile women and in 10% of the men on long-term antipsychotic therapy.

Our findings of elevated PRL values at lower reference doses in female compared with male patients, as well as both higher ratios between PRL level and reference dose and higher PRL values per drug serum concentration unit in the women, all suggest a sex-related difference in the sensitivity to antipsychotics of the hypothalamic-pituitary PRL axis. Higher oestrogen levels in fertile women may be one explanation [36], but hyperprolactinaemia, despite lower antipsychotic doses, was also noted in our postmenopausal women.

Through analysing IGF-I, and using it together with GH as an indirect parameter of GH secretion, no GH deficiency was found in the patients in this study. However, in a previous report [12], three of 28 antipsychotic-treated patients showed decreased IGF-I levels (< −2 s.d.). In comparison, other investigators have studied the influence of antipsychotics on GH secretion by using the apomorphine test, and have found both a normal and blunted GH-response to apomorphine in patients on long-term antipsychotic therapy [37].

IGF-I levels in the patients in our study were not measured prior to treatment with antipsychotics. Thus, we are unable to make an intra-individual comparison of IGF-I levels. However, dysregulation of GH secretion at the supra-hypothalamic/hypothalamic level in unmedicated schizophrenic patients cannot be excluded; in an earlier study, it was shown that the GH-response in the apomorphine test was larger in patients having Schneider's first-rank symptoms of schizophrenia compared with patients without such symptoms and with healthy controls [38], and in another study the GH-response correlated significantly with psychosis ratings and negative symptom scale scores [39]. A positive correlation between the apomorphine stimulated GH-response and the occurrence of thought disorders has also been found in 138 investigated patients with schizophrenia or schizoaffective disorders [40]. Thus, the GH-IGF-I secretory pattern in our patients might have been influenced by the psychosis per se before treatment with antipsychotics.

Weight gain is a significant side-effect of antipsychotics [41, 42]. In this study BMI was elevated in 57% of the patients, which may reflect that antipsychotics induce marked weight gain [43]. However, we did not find any sex difference in the frequency of elevated BMI. Some of our patients also showed a clinical appearance of altered body constitution with central obesity (no measurements performed), where GH deficiency might be suspected [13, 44]. However, using IGF-I as a marker, GH deficiency could not be verified.

This study includes a large patient group on long-term treatment with antipsychotics. However, the study is limited since the hormone concentrations were assessed on a single occasion, and by the patients' use of several different types of antipsychotics. In the future, studies of diurnal profiles of PRL, GH and IGF-I in patients on monotherapy with either classical or newer antipsychotics would be of interest.

In conclusion, we observed a notable difference in the median, as well as in the median body weight adjusted dose of antipsychotics between the sexes, which can be explained by sex differences in metabolic clearance and possibly also by differences in severity of the psychoses between men and women. In accordance with other studies, we also found antipsychotic-induced hyperprolactinaemia, which was more frequent and occurred at a lower daily dose of antipsychotics in women compared with men, pointing to a sex-related difference in the sensitivity to antipsychotics in the hypothalamic-pituitary PRL-regulation. Furthermore, irrespective of sex, more than half of our patients had elevated BMI. Some of the patients also showed a clinical appearance, where GH deficiency might be suspected. However, assessing GH-IGF-I in serum, no obvious influence on the GH-IGF-I axis was noted. Thus, 47 patients on long-term antipsychotic treatment, where the doses were adjusted to obtain therapeutic efficiency, exhibited hyperprolactinaemia in 32% and elevated BMI in 57%, but no signs of impaired GH-IGF-I secretion.