Impaired serotonergic neurotransmission has been implicated in the pathogenesis of depression and schizophrenia. Blood platelets have been used for years as a peripheral model of neuronal serotonin dynamics. The objective was to investigate platelet count and serotonin concentration in patients with depression and schizophrenia, in an attempt to ascertain their clinical usefulness.
953 participants were included in the study, 329 patients with depression, 339 patients with schizophrenia and 285 healthy controls. ELISA was used to assess platelet serotonin concentrations.
There were no statistically significant differences among groups regarding age, total platelet count and serotonin concentration. Linear regression analyses revealed inverse correlations between platelet serotonin concentration and age of patients with depression and healthy individuals, as well as between platelet serotonin concentration and illness duration in patients with schizophrenia. In other words, longer illness duration in patients with schizophrenia, and higher age in patients with depression and healthy individuals was associated with lower platelet serotonin concentrations.
Platelet count and serotonin concentration did not prove to be of diagnostic value in differentiating patients and healthy individuals. However, illness duration in patients with schizophrenia may be associated with reduced concentrations of platelet serotonin.
The search for specific diagnostic markers in psychiatry still has not provided that many clinically useful findings and that is especially true in the case of depression and schizophrenia [
Serotonergic system has long been implicated in the pathogenesis of depression, while its proposed effects within the schizophrenic process are closely related to the major implicated pathway of dopamine [
Among other factors that have been shown to influence platelet count and possibly serotonin concentration [
Available findings describe significantly lower platelet serotonin concentrations in depressive patients compared to healthy individuals [
Studies on platelet serotonin in schizophrenia repeatedly reported significantly higher concentrations, especially in untreated patients [
These findings indicate that the platelet serotonin concentrations might have a functional significance in pathophysiology and/or treatment of depression and schizophrenia, regardless significant differences between investigations. In this research we focused on two platelet variables, platelet count and platelet serotonin concentration, in an attempt to gain insight into the question if they could be used as a clinically useful diagnostic test in schizophrenia and/or depression.
In total, 953 participants were enrolled in this research. The schizophrenia group consisted of 339 patients, while the depression group consisted of 329 patients. Control group consisted of 285 healthy participants who volunteered to participate in the research. The schizophrenia group included 221 males and 118 females, while the group of patients suffering from depression included 200 males and 129 females. Group of healthy individuals included 160 males and 125 females. Sociodemographic and clinical parameters of all included subjects are presented in
The study included all patients that had been admitted for inpatient treatment with the diagnosis of schizophrenia or depression in the period from June 2014 to September 2019. All patients with depression and schizophrenia presenting the above described exclusion criteria were excluded from the study. Structured clinical interview was performed by a psychiatrist, who made the diagnosis based on the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) criteria for schizophrenia. The severity of schizophrenia was assessed by the Positive and negative syndrome scale (PANSS) [
Blood samples were taken from patients for the analysis to determine platelet serotonin levels. Blood samples were taken from the cubital vein in one vacuumed eprouvette with EDTA anticoagulant for the determination of platelet serotonin in the morning, after 12 h of fasting and after a 30 minute pause. The platelet serotonin was determined with the enzyme-linked immunosorbent assay (ELISA) procedure using commercial reagents (DRG Diagnostics, Germany). In our laboratory, the value of the inter-assay CV was 6% for platelet serotonin.
The platelet number was determined on hematology analyzer (Beckman Coulter DxH 800, USA), with prior verification of a blank sample. The PRP was prepared a maximum of 2 h after taking the blood sample. The PRP was obtained by centrifuging the sample at 200×g for 10 min at room temperature. The supernatant was separated and the number of platelets within was also determined. Platelets were then lysed by separating 200 µL PRP in 800 µL 0.9% sodium chloride solution and centrifuging the suspended cells at 4,500×g for 10 min at 4°C. After separation of the supernatant, 200 µL of redistilled water was added to the sediment containing platelets and the mixture was mechanically mixed on a vortex mixer to fully lyse the platelets. This prepared platelet lysate was stored at -20°C (for a maximum of three weeks). To determine the concentration of serotonin in the stored platelet lysate, the samples were thawed and centrifuged at 10,000×g for 2 min at room temperature.
Kolmogorov-Smirnov test was used to assess normal distribution for all measures and for each group. Sociodemographic and clinical characteristics of patients and healthy control subjects were compared in frequencies using the chi-square test and for continuous sociodemographic variables and platelet count and serotonin concentration using the independent samples t-test (comparison between two groups) or ANOVA (comparison between three groups). We performed linear regression analysis in order to test the predictions of symptom severity of schizophrenia and depression (PANSS and HDRS), number of episodes and overall illness duration for both disorders on platelet count and serotonin concentration. We also used linear regression analysis to test the predictions of patient’s age on total platelet count and serotonin concentration. Optimal sensitivity and specificity of the platelet serotonin concentration for the diagnosis of depression and schizophrenia were determined with the receiver operating characteristic (ROC) curve analysis and utilizing a nonparametric approach. The Youden index was calculated for each cut-off value as corresponding [(sensitivity+specificity)-1] to find the cut-off values that maximize discriminating power of the test. Alpha values were conservatively considered significant at less than 0.01. Statistical analyses were conducted using SPSS software (SPSS for Windows 20.0, IBM Corp., Armonk, NY, USA).
We established statistically significant differences between patients with depression and patients with schizophrenia and healthy individuals regarding several sociodemographic and clinical characteristics. However, there were no statistically significant differences among groups regarding age, total platelet count and serotonin concentration. Two groups consisting of patients with depression and schizophrenia did not statistically differ in terms of number of illness episodes and illness duration (
In order to establish if age influences total platelet count, we performed linear regression analysis in which platelet count was a dependent variable and age (separately for all three groups) was a predictor variable. Linear regression analysis established that the variance of the dependent variable is explained by 0.1% (r2=0.001) in depression, 0.2% (r2=0.002) in schizophrenia and 0.3% (r2=0.003) in healthy individuals. Analysis of variance was performed as the next phase of linear regression. It did not reveal statistical significance of the model (F=0.278; p=0.598 in depression; F=0.514; p=0.474 in schizophrenia and F=0.786; p=0.376 in healthy individuals), which means that platelet count does not depend on age (
Next step was to test if number of episodes and illness duration have an influence on the total platelet count. Therefore, we performed linear regression analysis in which platelet count was a dependent variable and number of episodes and illness duration (separately for groups with depression and schizophrenia) were predictor variables. Linear regression analysis established that the variance of the dependent variable is explained by 1.9% (r2=0.019) in depression and 1.6% (r2=0.016) in schizophrenia. Analysis of variance was performed as the next phase of linear regression. It did not reveal statistical significance (F=1.217; p=0.300 in depression; F=2.249; p=0.107 in schizophrenia), which means that platelet count does not depend on the two included predictor variables (
And finally, we tested if clinical severity scores (HDRS and PANSS) have an influence on the total platelet count. Linear regression analysis was performed with platelet count as a dependent variable and clinical severity scores as predictor variables. In the end, variance of the dependent variable is explained by 2.6% (r2=0.026) of the HDRS score in depression and 1.7% (r2=0.017) of the PANSS score in schizophrenia. Analysis of variance was performed as the next phase of linear regression. It did reveal statistical significance (F=5.382; p=0.021 for HDRS and F=1.095; p=0.360 for PANSS), which means that platelet count does not depend on PANSS scores in schizophrenia, but it does on HDRS score in depression (
Linear regression analyses were also performed in order to establish the influence of the same set of variables on platelet serotonin concentration. In the first step platelet serotonin concentration was the dependent variable and age (separately for all three groups) was a predictor variable. Linear regression analysis established that the variance of the dependent variable is explained by 1.5% (r2=0.015) in depression, 0.3% (r2=0.003) in schizophrenia and 1.8% (r2=0.018) in healthy individuals. Analysis of variance was performed as the next phase of linear regression. It did reveal statistical significance in patients with depression (F=5.038; p=0.025) and healthy individuals (F=4.464; p=0.036), but not in patients with schizophrenia (F=1.080; p=0.300), which means that platelet serotonin concentration does depend on age in patients with depression and healthy individuals (
Next step was to test if number of episodes and illness duration have an influence on platelet serotonin concentration. In this linear regression analysis number of episodes and illness duration (separately for groups with depression and schizophrenia) were predictor variables. Linear regression analysis established that the variance of the dependent variable is explained by 0.2% (r2=0.002) in depression and 2.6% (r2=0.026) in schizophrenia. Analysis of variance revealed statistical significance for the sample of patients with schizophrenia (F=3.972; p=0.020), but not for the sample of patients with depression (F=0.174; p=0.840), which means that platelet serotonin concentration does depend on the illness duration variable in patients with schizophrenia (
As the last step we tested if clinical severity scores (HDRS and PANSS) have an influence on the platelet serotonin concentration. Clinical severity scores were predictor variables. Variance of the dependent variable is explained by 0.4% (r2=0.004) of HDRS score in depression and 5.2% (r2=0.052) of PANSS score in schizophrenia. Analysis of variance revealed statistical significance of the model with platelet serotonin concentration as a dependent variable and PANSS score as a predictor variable (F=3.878; p=0.004), but not when HDRS was a predictor variable (F=0.917; p=0.339), which means that platelet serotonin concentration does not depend on HDRS score in depression, but it does on PANSS score in schizophrenia. However, when analysing PANSS subscale scores and even total PANSS score independently, that significance was not present (
ROC analysis did not establish significant diagnostic value of platelet serotonin concentration when comparing patients with depression and healthy individuals. The area under the curve (AUC) was 0.517 with a 95% confidence interval (0.472–0.563; p=0.455). ROC analysis also failed to demonstrate the diagnostic value of platelet serotonin concentration when comparing patients with depression and schizophrenia. The area under the curve (AUC) was 0.515 with a 95% confidence interval (0.471–0.558; p=0.514). Finally, ROC analysis did not establish significant diagnostic value of platelet serotonin concentration when comparing patients with schizophrenia and healthy individuals. The area under the curve (AUC) was 0.510 with a 95% confidence interval (0.465–0.556; p=0.655). We did not include graphical presentations due to non-significant findings of the ROC analysis.
This study included 953 participants, which is to the best of our knowledge largest investigation of platelet serotonin count and concentration in depression and schizophrenia. There were no differences in ratio of males to females when comparing three groups of participants, and majority of patients (compared to healthy individuals) were of lower educational status, unemployed or retired. All of these findings reflect our understanding of the burden that psychiatric disorders impose on patients, especially those suffering from schizophrenia [
The groups did not differ in age, platelet count and platelet serotonin concentration. When it comes to comparing patients with depression and healthy individuals, previously there have been similar findings [
Previous reports in healthy individuals established that age was inversely associated with platelet count and platelet serotonin content [
Two groups of patients did not differ regarding number of episodes and overall illness duration. However, when examining for a possible correlation between number of episodes and overall illness duration on platelet count and serotonin concentration, we established a single positive correlation, the one between overall illness duration and platelet serotonin concentration in schizophrenia. Therefore, in patients with schizophrenia there is a significant decline of platelet serotonin concentration (in an otherwise not significantly declining platelet count) with longer duration of illness. This finding is to the best of our knowledge the first report of such an association. It also adds some weight to the hypothesized serotonergic dysfunction in schizophrenia [
There are several limitations to this study, one of those is the fact that it was cross-sectional without a prospective design. Therefore, it was not possible to study the dynamics of the influence of the clinical course and/or the influence of psychopharmacological treatment on the investigated parameters, which is important, as studies have shown that antidepressants and antipsychotics can influence platelet serotonin and scores of scales used in this study [
In conclusion, this study provides first evidence of differing platelet serotonin concentrations in schizophrenia regarding the duration of the illness in a manner that duration seems to be inversely correlated to platelet serotonin concentration. Interestingly, age was not correlated with platelet serotonin concentration in schizophrenia, suggesting that illness duration solely influences platelet serotonin homeostasis and is quite possibly a trait of schizophrenia duration. Combined with age wise significant findings in patients with depression and healthy individuals, this finding could be specific to schizophrenia. In terms of distinguishing patients with schizophrenia and healthy individuals (or patients with depression for that matter) platelet serotonin concentration did not prove to be of diagnostic value. However, this finding does not allow us to make generalizations about the concentration of intracellular serotonin. In a platelet model such as this, intracellular serotonin depends also on several other factors, which were not investigated at this point. Future studies should therefore include all relevant elements of serotonin homeostasis across the human life span in an attempt to clarify serotonin dynamics in ageing of psychiatric patients, which might provide an answer to relatively ambiguous findings regarding platelet serotonin concentrations in researches performed so far.
This research was funded from the fund for scientific research of University of Zagreb’s School of Dental Medicine.
The authors are indebted to all of the physicians, nurses, and study coordinators who cared for the patients and who kindly provided the data necessary for our analysis.
The authors have no potential conflicts of interest to disclose.
Conceptualization: Dalibor Karlović. Data curation: Vjekoslav Peitl, Biserka Getaldić-Švarc. Formal analysis: Vjekoslav Peitl. Investigation: Vjekoslav Peitl, Biserka Getaldić-Švarc, Dalibor Karlović. Methodology: Biserka Getaldić-Švarc, Dalibor Karlović. Resources: Dalibor Karlović. Software: Dalibor Karlović. Supervision: Vjekoslav Peitl, Biserka GetaldićŠvarc, Dalibor Karlović. Validation: Vjekoslav Peitl, Dalibor Karlović. Writing—original draft: Vjekoslav Peitl, Biserka Getaldić-Švarc, Dalibor Karlović. Writing—review & editing: Vjekoslav Peitl, Dalibor Karlović.
Sociodemographic and clinical characteristics of healthy individuals and patients with depression and schizophrenia
Depression | Schizophrenia | Healthy individuals | Statistical details |
||
---|---|---|---|---|---|
N=329 | N=339 | N=285 | χ2-value | p-value | |
Gender, N (%) | 5.337 | 0.069 | |||
Male | 200 (60.8) | 221 (65.2) | 160 (56.1) | ||
Female | 129 (39.2) | 118 (34.8) | 125 (43.9) | ||
Education, N (%) | 158.124 | <0.01 | |||
Elementary | 44 (13.4) | 39 (11.5) | 4 (1.3) | ||
High school | 234 (71.1) | 240 (70.8) | 99 (34.8) | ||
University | 51 (15.5) | 60 (17.7) | 182 (63.9) | ||
Marital status, N (%) | 95.717 | <0.01 | |||
Married | 236 (71.9) | 103 (30.6) | 178 (62.4) | ||
Single | 68 (20.6) | 213 (62.7) | 96 (33.6) | ||
Divorced | 25 (7.5) | 23 (6.7) | 11 (4) | ||
Working status, N (%) | 181.615 | <0.01 | |||
Employed | 128 (39.0) | 105 (30.9) | 248 (86.9) | ||
Unemployed | 101 (30.8) | 142 (41.8) | 28 (10.0) | ||
Retired | 100 (30.2) | 92 (27.3) | 9 (3.1) | ||
Place of living, N (%) | 4.081 | 0.130 | |||
Urban | 223 (67.8) | 259 (76.3) | 216 (75.7) | ||
Rural | 106 (32.2) | 80 (23.7) | 69 (23.3) | ||
Ages in years (mean±SD) | 38.3±14.5 | 39.3±14.0 | 37.7±11.8 | F 1.026 | 0.359 |
Platelets (mean±SD) | 232.2±62.2 | 232.4±73.5 | 232.0±49.5 | F 0.005 | 0.995 |
Platelet serotonin (mean±SD) | 475.4±321.6 | 532.1±434.1 | 507.8±350.2 | F 1.936 | 0.145 |
Number of episodes (mean±SD) | 3.77±2.65 | 4.88±3.64 | N/A | ||
Illness duration (mean±SD) | 2.40±1.69 | 2.04±1.12 | N/A | ||
PANSS total (mean±SD) | N/A | 103.74±20.16 | N/A | ||
PANSS positive (mean±SD) | N/A | 20.12±6.23 | N/A | ||
PANSS negative (mean±SD) | N/A | 27.56±5.54 | N/A | ||
PANSS general psychopatology (mean±SD) | N/A | 56.00±11.24 | N/A | ||
HDRS total (mean±SD) | 18.13±6.01 | N/A | N/A |
PANSS: Positive and Negative Syndrome Scale, HDRS: Hamilton Depression Rating Scale, N/A: not applicable
Predictive values of age, illness duration, episode number and clinical severity scores (HDRS and PANSS) on total platelet count
Platelet count | Unstandardized coefficients |
Standardized coefficients |
t | p | 95.0% confidence interval for B |
||
---|---|---|---|---|---|---|---|
B | Std. error | β | Lower bound | Upper bound | |||
Age | |||||||
Depression | 0.165 | 0.314 | 0.037 | 0.528 | 0.598 | -0.453 | 0.784 |
Schizophrenia | 0.221 | 0.309 | 0.043 | 0.717 | 0.474 | -0.387 | 0.830 |
Healthy | -0.244 | 0.276 | -0.059 | -0.886 | 0.376 | -0.788 | 0.299 |
Depression | |||||||
Episode number | 0.307 | 2.127 | 0.013 | 0.144 | 0.885 | -3.902 | 4.516 |
Illness duration | -5.225 | 3.349 | -0.138 | -1.560 | 0.121 | -11.853 | 1.404 |
HDRS | 1.675 | 0.722 | 0.161 | 2.320 | 0.021 | 0.251 | 3.098 |
Schizophrenia | |||||||
Episode number | -2.006 | 1.168 | -0.104 | -1.718 | 0.087 | -4.306 | 0.293 |
Illness duration | -5.659 | 4.827 | -0.071 | -1.172 | 0.242 | -15.162 | 3.844 |
PANSS total | 2.635 | 1.961 | 0.722 | 1.344 | 0.180 | -1.226 | 6.497 |
PANSS positive | -2.001 | 2.090 | -0.170 | -0.957 | 0.339 | -6.117 | 2.116 |
PANSS negative | -1.809 | 2.313 | -0.138 | -0.782 | 0.435 | -6.364 | 2.747 |
PANSS general | -3.585 | 2.065 | -0.545 | -1.736 | 0.084 | -7.652 | 0.481 |
HDRS: Hamilton Depression Rating Scale, PANSS: Positive and Negative Syndrome Scale
Predictive values of age, illness duration, episode number and clinical severity scores (HDRS and PANSS) on platelet serotonin concentration
Platelet serotonin concentration | Unstandardized coefficients |
Standardized coefficients |
t | p | 95.0% confidence interval for B |
||
---|---|---|---|---|---|---|---|
B | Std. error | β | Lower bound | Upper bound | |||
Age | |||||||
Depression | -2.912 | 1.297 | -0.123 | -2.244 | 0.025 | -5.465 | -0.360 |
Schizophrenia | -1.755 | 1.690 | -0.057 | -1.039 | 0.300 | -5.079 | 1.568 |
Healthy | -4.178 | 1.977 | -0.136 | -2.113 | 0.036 | -8.074 | -0.283 |
Depression | |||||||
Episode number | 0.228 | 10.388 | 0.002 | 0.022 | 0.982 | -20.306 | 20.763 |
Illness duration | 9.613 | 16.360 | 0.049 | 0.588 | 0.558 | -22.728 | 41.954 |
HDRS | 3.333 | 3.481 | 0.064 | 0.957 | 0.339 | -3.527 | 10.193 |
Schizophrenia | |||||||
Episode number | -0.357 | 6.882 | -0.003 | -0.052 | 0.959 | -13.900 | 13.186 |
Illness duration | -62.869 | 22.306 | -0.160 | -2.819 | 0.005 | -106.763 | -18.975 |
PANSS total | -3.718 | 11.638 | -0.168 | -0.320 | 0.750 | -26.626 | 19.189 |
PANSS positive | -2.148 | 12.354 | -0.030 | -0.174 | 0.862 | -26.465 | 22.169 |
PANSS negative | 6.377 | 13.642 | 0.079 | 0.467 | 0.641 | -20.474 | 33.229 |
PANSS general | -3.837 | 12.223 | -0.097 | -0.314 | 0.754 | -27.897 | 20.222 |
HDRS: Hamilton Depression Rating Scale, PANSS: Positive and Negative Syndrome Scale