Hyperprolactinemia is a common and well-known side effect of antipsychotic treatment. However, prolactin levels are elevated in drug-naïve first-episode psychosis (FEP) patients, which suggests that the association between prolactin and psychosis may be more complex than previously thought. One possible hypothesis to explain this association is that prolactin has a neuroprotective effect mediated by BDNF.
Methods50 drug- naïve FEP were included and clinical characteristics as DUP; tobacco and cannabis use; and BMI were assessed. The PANSS and GAF questionnaires were administered. Fasting blood samples were obtained to determine prolactin, TSH, and BDNF levels. All statistical analyses were performed in the whole sample and stratified by sex.
ResultsOf the 50 patients, 42 % had hyperprolactinemia (33.3 % of males and 55 % of females). Prolactin levels were negatively correlated with baseline GAF scores in the whole sample and in females but not in males. Prolactin levels were positively and independently associated (regardless of sex, age, BMI, TSH level, tobacco or cannabis use) with BDNF levels.
ConclusionsProlactin secretion may be dysregulated in drug-naïve FEP, which implies that hyperprolactinemia—a common condition in FEP —may not be solely attributable to antipsychotic treatment. Elevated prolactin levels could reflect a physiological response designed to protect the central nervous system in certain critical situations. This neuroprotective effect could be mediated by increasing BDNF levels. Hyperprolactinemia is common in antipsychotic-naïve FEP. This study shows that higher prolactin levels are associated with elevated BDNF levels, suggesting a neuroprotective effect mediated through BDNF in this patient population.
Numerous studies have been carried out to evaluate the association between prolactin and schizophrenia.1 Although hyperprolactinemia is generally considered to be a side effect of antipsychotic treatment, data from some studies2,3 suggest that the association between hyperprolactinemia and schizophrenia might be more complex than previously thought, and that factors other than antipsychotics could also affect serum prolactin concentration.
Prolactin is a polypeptide hormone mainly synthesised by specialised cells in the anterior pituitary gland. This hormone is involved in the regulation of nervous system–related processes, such as stress and trauma response, energy balance, food intake, anxiety, neurogenesis, and pain. Prolactin is controlled by the inhibitory effects of dopamine, which is produced in the tuberoinfundibular dopaminergic neurons.4 It has been shown that prolactin can enter the brain.5 Although the precise mechanisms are not well understood, it is generally believed that prolactin enters the brain through prolactin receptors in the choroid plexus and endothelial cells in the brain capillaries.6 Once inside the brain, this hormone can exert important behavioural effects, mainly related to reproductive and maternal behaviour.
The studies carried out to date to assess prolactin levels in drug–naïve first episode psychosis (FEP) have reported contradictory results. While some early studies found that these patients have normal or low serum prolactin levels,7,8 more recent studies have found higher prolactin levels.2,9,10 In one study, individuals at high risk for psychotic disorder had elevated levels of prolactin.11 Prolactin levels may be associated with positive and negative symptomatology in a sex-dependent manner. For example, a study in patients with FEP12 found a negative correlation between prolactin levels and negative symptomatology and a positive correlation between prolactin and positive symptoms at baseline and at one-year of follow-up, but only in males. By contrast, Delgado-Alvarado et al. found that higher prolactin levels in female FEP patients were negatively correlated with the severity of positive symptoms,13 which suggests that prolactin could be a protective factor in females. Given these findings, it seems clear that, in addition to antipsychotic drugs, other factors may also influence prolactin levels, especially sex,2,14 but also thyroid stimulating hormone (THS), body mass index (BMI), and tobacco and/or cannabis use.15,16 For example, one study found that cannabis frequent users had significantly lower baseline prolactin levels than healthy controls.17
Brain derived neurotrophic factor (BDNF) is the most widely distributed neurotrophin in the central nervous system (CNS) and it is highly involved in the etiopathogenesis of psychotic disorders.18 Data from systematic reviews and meta-analyses show that patients with schizophrenia19 and FEP18 have lower peripheral BDNF levels than healthy controls. Nevertheless, various studies included in these meta-analysis were highly heterogeneous or excluded some potential confounding variables as cannabis use. Taking all these into account, BDNF could be considered an important potential biomarker due to the close correlation between peripheral and brain BDNF.20
The association between prolactin and BDNF has received relatively little attention to date. However, one study found that prolactin upregulates BDNF and other neuroprotective factors.21 A study conducted in mice with chronic mild stress-induced depression22 found that prolactin receptor gene silencing inhibits hippocampal neuron apoptosis by inactivating the JAK2-STAT5 signalling pathway and elevating BDNF expression. By contrast, a study in antidepressant treatment-free depressed patients23 found no association between prolactin and BDNF levels. To our knowledge, no studies have been carried out to date to specifically explore the association between BDNF and prolactin levels in patients with schizophrenia or FEP.
Aims and hypothesisThe main aim of the present study was to determine serum prolactin levels in a cohort of adult FEP patients of both sexes and to determine whether there are sex-related differences in terms of the presence of hyperprolactinemia. Secondary aims were to determine the association between prolactin levels and clinical characteristics and between peripheral BDNF and prolactin levels, adjusted for potential confounders.
We hypothesised the following: 1) drug-naïve FEP patients would present altered prolactin levels; 2) prolactin levels would be associated with clinical symptomatology; and 3) higher prolactin levels would be associated with higher levels of peripheral BDNF.
Material and methodsStudy populationA total of 50 drug-naïve FEP patients treated at the ETEP (Study and Treatment Program for First Episode Psychosis) at the Hospital del Mar (Barcelona, Spain) between April 2013 and July 2017 were consecutively included. The ETEP is a specialized early intervention service for young adults (aged 18 to 35 years) diagnosed with FEP. The program involves a multimodal intervention, including comprehensive assessment and intensive medical and psychosocial treatment.24
The inclusion criteria were as follows: 1) age 18–35 years; 2) fulfilment of DSM-IV-TR criteria for any of the following: brief psychotic disorder, schizophreniform disorder, schizophrenia with < one year of symptoms, or unspecified psychosis; 3) no prior history of severe neurological medical conditions or severe traumatic brain injury; 4) presumed IQ level > 80 based on clinical records, and 5) no substance abuse or dependence disorders except for cannabis and/or nicotine use.
To homogenize the sample, only non-affective psychosis patients were included, patients with affective features were excluded in our sample.
All patients were antipsychotic and antidepressant treatment-naïve. Treatment with benzodiazepines was allowed. To carry this out, blood analyses were done at admission as soon as possible, always between 8 a.m. and 12 pm. If the patient had symptoms of anxiety or agitation, benzodiazepines were administrated. In case that patients need antipsychotic medication before blood sample obtention, these patients were excluded of the study.
This study was approved by the local ethics committee at the Hospital del Mar (approval code: 2021/100,93). All participants provided written informed consent. The study protocol complies with the ethics criteria established by the Declaration of Helsinki and with all local laws on patient confidentiality and data protection.
Clinical assessment and demographic dataAll patients underwent a comprehensive evaluation at baseline by two experienced psychiatrists (A.M., D.B.). Sociodemographic variables (age, sex) were recorded. Weight and height measurements were made to calculate the BMI. Substance use was assessed, including tobacco (cigarettes per day) and cannabis use (‘joints’ per week, after dichotomization between users and non-users). The Structured Clinical Interview for DSM-IV-TR Axis I disorders was administered to establish the diagnosis. Three scales were administered, as follows: the Positive and Negative Syndrome Scale (PANSS) for symptoms related to psychosis25; the Global Assessment of Functioning (GAF)26 for functionality; and the Calgary Depression Scale for Schizophrenia (CDSS) for depressive symptoms.27
Collection of blood samples and determination of serum BDNF, TSH, and prolactin levels at baselineFasting blood samples were obtained at admission (i.e., at baseline) before any medications (except for benzodiazepines) were administered. All blood samples were obtained in the morning (between 8 a.m. and 12 pm.) to avoid circadian-related fluctuations in BDNF levels, which have been reported to occur in men but not in women.28 Blood samples were collected in glass K3– EDTA blood-drawing tubes for whole blood. The serum was isolated by centrifugation at 300 x g for 15 min, removed and stored at – 80 °C until analysis.
BDNF levels were measured using the ChemiKine sandwich enzyme-linked immunosorbent assay kit (Chemicon, Temecula, California; USA) according to the manufacturer's instructions. The Wallac Victor 2 micro-plate reader (wavelength, 450 nm) was used to determine absorbance. BDNF concentrations were determined according to the standard curve, which was constructed from duplicate samples containing appropriate concentrations. All samples were analysed in duplicate. The calculated overall intra– and inter-assay variation coefficients were 3.7 % and 8.5 %, respectively. The detection limit of the BDNF assay was 15 pg/mL.
Prolactin and TSH levels were determined under routine conditions on the same day of the blood draw. The hormonal assay was performed in the testing laboratory at the Hospital del Mar. Prolactin and TSH levels were determined by electrochemiluminescence immunoassay. TSH levels between 0.30 and 4.20 mcUI/mL and prolactin levels < 530 mcIU/mL were considered normal for both sexes. Prolactin and TSH inter- and intra-assay coefficients of variability were < 5 %.
Hyperprolactinemia was defined in accordance with the hospital laboratory reference range as a serum prolactin level > 410 mlU/L (≈ 19.3 ng/mL) for males > 510 mlU/L (≈ 24.1. ng/mL) for females.
Statistical analysisData normality was assessed with the Kolmogorov-Smirnov test. A descriptive analysis of the sample was performed, including the percentage of patients with prolactin alterations. The descriptive analysis was performed for the whole group and according to sex. Pearson's correlation coefficient was estimated to determine the correlation between prolactin levels and clinical symptomatology (PANSS total score and sub-scale scores, CDSS, and GAF) overall and stratified by sex. The same test was used to determine the correlation between prolactin and BDNF levels.
A linear regression model (enter method) was performed, with prolactin levels as the dependent variable. Based on previous studies,2,29 we included the following independent variables in the model: age; sex; duration of untreated psychosis (DUP); BMI; nicotine use; cannabis use; BDNF level; and TSH level.
All statistical analyses were performed with the IBM-SPSS Statistics for Windows, v. 29.0.2.0 (IBM Corp.; Armonk, NY, USA). P values ≤ 0.05 were considered statistically significant.
ResultsPatient characteristicsA total of 50 patients (30 males; 60 %) were included study. The median (interquartile range [IQR]) age was 26 (24 - 30.25) years. In terms of diagnosis, the 56 % (n = 28) were unspecified psychosis, the 36 % (n = 18) were schizophreniform disorder and the 4 % (n = 8) were brief psychotic disorder.
In the full cohort, the mean (standard deviation [SD]) prolactin level was 707.4 (863.0) mcIU/mL; by sex, mean levels were 988.1 (1101.2) mclU/mL in females vs. 401.2 (326.9) mclU/mL in males. Overall, the mean (SD) BDNF value was 45.3 (27.1) ng/mL; by sex, the mean levels were 51.4 (30.7) in females and 41.2 (24.1) in males. No significant sex-related differences were observed for prolactin (p = 0.054) or BDNF levels (p = 0.164). However, there was a non-significant trend towards higher prolactin levels in females.
The other sociodemographic, clinical, and biochemical characteristics of the patients at baseline are shown in Table 1.
Baseline sociodemographic, clinical, and biochemical variables.
Abbreviations: N = sample size; IQR = interquartile range; m = mean; SD = standard deviation; DUP = duration of untreated psychosis; BMI = body mass index; BDNF = brain-derived neurotrophic factor; TSH = thyroid-stimulating hormone; PANSS P = Positive and Negative Syndrome Scale positive; PANSS N = Positive and Negative Syndrome Scale negative; PANSS GP = Positive and Negative Syndrome Scale general pathology; PANSS T = Positive and Negative Syndrome Scale total; CDSS = Calgary Depression Scale for Schizophrenia; GAF = Global Assessment of Functioning; mL = milliliter.
In this sample of 50 drug-naïve patients, 42 % (n = 21) presented baseline hyperprolactinemia. Hyperprolactinemia was detected in 55 % (n = 11/20) of females and 33.3 % (n = 10/30) of males, a non-significant difference (p = 0.407).
In terms of clinical symptoms, the only significant correlation in the sample was a negative correlation between prolactin levels and the GAF score at baseline. This correlation was statistically significant in females but not in males. No other significant correlations between prolactin levels and clinical symptomatology were observed (Table 2).
Correlation between prolactin levels and clinical symptomatology overall and by sex.
| PANSS P | PANSS N | PANSS GP | PANSS T | CDSS | GAF | |
|---|---|---|---|---|---|---|
| Total (r, p) | 0.31 (0.161) | 0.28 (0.188) | 0.25 (0.255) | 0.02 (0.920) | 0.04 (0.866) | −0.47 (0.025*) |
| Female (r, p) | 0.39 (0.212) | 0.25 (0.433) | 0.29 (0.355) | 0.04 (0.914) | 0.09 (0.768) | −0.69 (0.013*) |
| Male (r, p) | 0.30 (0.365) | 0.48 (0.137) | 0.22 (0.508) | 0.17 (0.615) | 0.05 (0.879) | −0.09 (0.802) |
Significant correlations are given in bold text.
Abbreviations: r = correlation coefficient; PANSS P = Positive and Negative Syndrome Scale positive; PANSS N = Positive and Negative Syndrome Scale negative; PANSS GP = Positive and Negative Syndrome Scale general pathology; PANSS T = Positive and Negative Syndrome Scale total; CDSS = Calgary Depression Scale for schizophrenia; GAF = Global Assessment of Functioning.
Prolactin levels were significantly and positively correlated with BDNF levels at baseline (r = 0.521; p = 0.011) (Fig. 1).
On the multivariate analysis, the model that best predicted prolactin levels was a single variable model (BDNF level), with an adjusted R2 of 0.541. At baseline, higher BDNF levels were significantly associated (95 % confidence interval [CI]: 4.18 - 28.35; p = 0.011) with higher prolactin levels. None of the following variables were associated with prolactin levels: age, sex, DUP, BMI, nicotine use, cannabis use, or TSH (Table 3).
Lineal regression model (enter method) of predictors of prolactin levels at baseline.
Abbreviations: B = regression coefficient, CI = confidence interval, DUP = duration untreated psychosis; BMI = body mass index; BDNF = brain-derived neurotrophic factor; TSH = thyroid- stimulating hormone.
In this sample of 50 drug-naïve FEP patients, hyperprolactinemia was present in 42 % of patients (n = 21) (33.3 % of males and 55 % of females). Prolactin levels were negatively correlated with the baseline GAF score in the whole sample and in females but not in males. Prolactin levels were positively associated with BDNF levels and this association could not be attributed to sex, age, BMI, TSH level, tobacco or cannabis use.
The reported prevalence of hyperprolactinemia in psychotic patients is highly variable, ranging from 30 % to 70 % in drug-naïve FEP patients.9,30 A study conducted in antipsychotic-naïve patients at clinical high risk for psychosis found higher prolactin levels in those patients compared to healthy controls,31 suggesting that elevated levels of prolactin could be due to dysregulation of prolactin secretion in drug-naïve FEP, a finding that seems to indicate that hyperprolactinemia in FEP patients may be attributable to other factors, not just to antipsychotic treatment alone.
Although we do not find any sex-related differences in prolactin levels in this sample, the literature shows that sex can influence these levels and must therefore be considering when determining prolactin levels. For example, it has long been known—even before the introduction of antipsychotics—that hypoestrogenism (which can cause irregular menstrual cycles and alter menopause periods) is common in women with schizophrenia.32 The cause of hypoestrogenism in these patients is believed to be undiagnosed hyperprolactinemia in which elevated levels of prolactin suppress gonadal function and physiological oestrogen production.
Several studies have examined the association between prolactin levels and psychotic symptomatology. One study carried out in antipsychotic-naïve FEP patients found that negative symptoms were associated with lower prolactin levels in women but higher levels in men.30 Two other studies33,34 found a positive correlation between plasma prolactin levels and negative PANSS subscale scores in male patients with chronic schizophrenia. Another study carried out in FEP patients found that men with higher levels of prolactin had more positive symptoms.35 The findings of those studies suggest that high prolactin levels may influence FEP severity in male patients but not in females. This hypothesis is consistent with the findings of several previous studies that found a protective role for prolactin in female FEP patients.13,36
Riecher-Rossler et al. hypothesised that hyperprolactinemia could trigger psychotic symptoms by enhancing dopamine levels through a feedback mechanism in predisposed individuals.30 This hypothesis, if correct, would explain the emergence of psychotic symptoms in some patients during times of stress or during periods of life (such as adolescence and menopause) in which sex hormones and prolactin levels fluctuate widely.
In terms of functionality, we found a significative negative correlation between GAF scores and prolactin levels in the whole sample and in females but not in men. To date, only a few studies have explored the association between prolactin and functioning, with contradictory results. Shrivastava et al. found that higher baseline levels of serum prolactin in schizophrenia patients were associated with better functional outcomes at 5 years of follow-up.37 Another study found that FEP patients with hyperprolactinemia had more severe clinical symptoms (significantly lower GAF scores at baseline) ,3 which could be due to the negative effects of hyperprolactinemia on cognitive function, as suggested by studies carried out in animal models,38 non-psychiatric populations,38 and patients with early psychosis.39
In our study, higher prolactin levels were significantly associated with higher BDNF levels in drug- naïve FEP patients. Moreover, this association could not be attributed to potential confounders (e.g., sex, age, BMI, TSH level, tobacco or cannabis use).
It is not entirely clear why prolactin levels tend to be elevated in drug- naïve FEP patients, although the cause is likely multifactorial. Studies show that even minor stress can increase prolactin secretion40 and severe stress is associated with onset of FEP.41 Nevertheless, some studies have ruled out this cause, as hyperprolactinemia does not appear to be associated with cortisol levels.2,10 Moreover, some studies have suggested that prolactin levels could be a good biomarker of acute stress but not of chronic stress, since prolactin can increase rapidly in response to acute stress, but these levels can also normalize rapidly once the stressful stimulus ceases.42,43
An alternative hypothesis is that inflammation may play a role in elevated prolactin levels.44 Similarly, the tendency for patients with FEP to present with hyperprolactinemia may indicate a more general vulnerability of those patients, which affects multiple components of the hypothalamic- pituitary-adrenal-gonadal axis, even prior to any antipsychotic treatment.45
Another hypothesis that could explain hyperprolactinemia in FEP patients, which is supported by our data, is that prolactin has a neuroprotective effect mediated by an increase in BDNF. Prolactin may have protective effects on the brain under stressful conditions,46 and one study found that prolactin may protect the hippocampus during pregnancy and lactation against high concentrations of glucocorticoids.47 Prolactin has also been shown to prevent a decrease in hippocampal neurogenesis induced by chronic stress.48 Another study22 found that prolactin could inhibit hippocampal neuron apoptosis through activation of the JAK/STAT signalling pathway in a mouse model of induced depression. The JAK/STAT signalling pathway is involved in key processes such as immunity, cell division, and cell death, and is also activated by BDNF.49
Considering the totality of the available evidence, together with the findings of our study, it appears that the presence of elevated prolactin levels in drug–naïve FEP could reflect a physiological response designed to protect the hippocampus or other brain structures. This neuroprotective effect could be mediated through an increase in BDNF levels.
Strengths and limitationsThe main limitation of this study is the cross-sectional design, which does not allow us to infer causality. Another limitation is that we did not clinically assess the menstrual cycle status in female patients, which could have influenced the results.50 Moreover, BDNF levels were determined at peripheral level, not at central. Although peripheral blood BDNF levels are commonly used as an indirect indicator of its concentration in the brain, the exact relationship between both is not fully established. In fact, some studies suggest that peripheral BDNF levels may be influenced by factors unrelated to brain activity, such as inflammation or exercise.51 In relation to this, genetic variations must also be taken into account, such as the Val66Met polymorphism, which has established pleiotropic effects on schizophrenia incidence and morphologic alterations in the illness.52
By contrast, the main strength of this study is the large sample size of drug-naïve FEP patients, a clinical population that is particularly difficult to recruit. Importantly, by excluding patients treated with antipsychotics, we were able to rule out the potential influence of those drugs on prolactin and BDNF levels.18,53 Another important strength is the multifactorial evaluation, which allowed us to control for several important potential confounding factors (e.g., cannabis use, tobacco use, TSH, and BMI).
ConclusionsOur findings show that hyperprolactinemia is common in antipsychotic-naïve FEP patients, thus suggesting that elevated levels of prolactin are not solely due to the effects of neuroleptics (D2– receptor blockade). We also found that higher prolactin levels were associated with higher BDNF levels in this sample of FEP patients, which indicates that prolactin could have a neuroprotective effect mediated by an increase in BDNF levels.
Ethical considerationsAll study procedures were carried out in accordance with the Declaration of Helsinki and approved by the Hospital del Mar Clinical Research Ethics Committee.
FoundingThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Dr. Mané and Dr. Bergé have both received financial support to attend meetings, travel support, and served as speakers for Otsuka, Angelini and Janssen Cilag. The other authors of this manuscript have no conflicts of interest to report.
We would like to thank all of the patients who made this study possible. We also thank Bradley Londres for professional English-language editing.
