Suicidal Thoughts Are Associated with Platelet Counts in Adolescent Inpatients

Abstract

Objective:

Platelets (PLT), which serve as the primary hemostatic indicator, can be used as a peripheral model for studying monoamine turnover in the brain. Therefore, they are attractive targets as circulatory biomarkers for the detection of psychiatric disorders. However, PLT counts have not been utilized as a peripheral biomarker of psychopathology.

Methods:

This study was a retrospective analysis of PLT counts upon admission of 108drug-naïve adolescents hospitalized in an inpatient psychiatric department. PLT counts of patients with suicidal ideation (SI) were compared with those of nonsuicidal in patients (NSI) and those of 77 healthy adolescents, serving as a control group. The patients' disorders were diagnosed and classified by one of four American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM IV) diagnoses, that is, unipolar depression, bipolar depression, schizophrenia, and a pooled group of conduct and borderline personality disorders.

Results:

Significantly higher PLT counts were observed in SI patients, as compared with NSI patients (300,200±53.3/mL vs. 253,900±53.2/mL, respectively; p=0.0001). A significant difference in PLT counts in SI patients, relative to the control group, was also noted (300,200±53.3/mL vs. 254,000±52/mL, respectively; p=<10⁻²⁶). Finally, a significant difference in PLT counts was observed between conduct/borderline personality disorders patient with and without suicidal ideation (292,000±55/mL vs. 246,000±64/mL, respectively; p=0.001).

Conclusions:

PLT counts are higher in suicidal hospitalized adolescents than in nonsuicidal inpatients, as well as than in controls.

Introduction

S uicide is the third most common cause of death among adolescents and young adults in the United Kingdom and the United States (McClure 2001; Dervic et al. 2008; Sadock et al. 2009)Various preventive measures, such as crisis hotlines, method restrictions, educational programs, and screening/case-finding efforts (Eaton et al. 2006) have been explored by various researchers. Despite such programs, as many as 200,000 adolescents and young adults attempt suicide annually in the United States (Greydanus and Calles 2007). Risk factors for suicidal behavior have been identified and include early onset of psychiatric disorders and counter gender behaviors, such as feminine behavior in boys (Reinherz et al. 1995); familial history; and present comorbidity with major depression, bipolar disorder, panic disorder, and drug abuse (Pfeffer et al. 1979; Brent 2001; Pfeffer 2002; Apter and King 2006).Whereas attempts to identify biological and/or genetic markers for suicidality have been reported (Laje et al. 2007),are reliable indicator for patients at risk has yet to be found.

Suicidal ideation is defined as thoughts or wishes to be dead or to kill or harm oneself (O'Carroll et al. 1996) and is perceived not only as a major risk factor for transition to suicidal behavior (Kessler et al. 1999) but also as part of suicidal behavior (Pfeffer and Plutchik 1982). Adolescents who have severe suicidal ideation (i.e., experience active, specific ideation with an intent and/or plan to commit suicide) have an up to 60% probability of attempting suicide within 1 year of declaring or confessing these thoughts (Kessler et al. 1999). Unfortunately, juveniles rarely volunteer information concerning suicidal ideation, making it difficult to ascertain risk via psychiatric evaluation and questioners in the community (Abram et al. 2008). Therefore, the availability of a laboratory aid for the detection of suicidal ideation in adolescents may assist in suicide risk assessment and, subsequently, its prevention.

For >30 years, thrombocytes have served as targets for neuropathological research. Platelets (PLT) have been shown to represent a good cellular model for studying the uptake, storage, and release of monoamines in postsynaptic brain neurons (Sneddon 1973; Longenecker 1985; Stahl 1985; Rehavi and Weizman 1990), processes, which are altered in several psychopathologies (Briley et al. 1980; Stahl et al. 1983; Hitzemann et al. 1984; Geller et al. 1988; Hrdina et al. 1995). In hospitalized patients, PLT serotonin concentration and receptor activities have been found to be significantly higher in suicidal patients than in nonsuicidal hospitalized patients (Rao et al. 1998) and differences related to specific symptoms have been demonstrated. Specifically, aggressiveness and the impulsivity of suicidal acts were correlated with serotonin content and the number of serotonin-2-binding sites on the PLT surface (Pandey et al. 1990; Spreux-Varoquaux et al. 2001).

Several reports suggest that cellular or molecular properties of PLT change in depressed or suicidal patients. Patients with major depression were reported to have higher mean PLT volume (MPV) values than controls (Canan et al. 2012), whereas treatment with escitalopram resulted in a significant reduction in MPV of depressed patients (Ataoglu and Canan 2009). Prepubertal children with major depression and/or suicidal behavior have been reported to have lowered PLT serotonin concentrations (Pfeffer et al. 1998). Finally, elevated PLT counts have been reported to be associated with psychological stress (Van Ischoot et al. 1982). Others have reported that differences at the molecular level (e.g., receptor concentrations) might signify suicidality. However, small sample size, insufficient matching criteria of control samples, use of inadequate ligands in binding experiments, and failure to consider comorbidity have rendered these studies as suggestive, at best (Muller-Oerlinghausen et al. 2004). Moreover, differences in the standard laboratory measures of PLT have not been investigated in adolescents.

In this study, we report analysis of the differences in PLT counts between hospitalized adolescents with various psychiatric disorders and a control group of adolescents without psychiatric complaints.

Methods

Subjects

The patient population consisted of 821 adolescents hospitalized between November 1, 1995 and October 31, 2004 in the inpatient psychiatric ward at the Beer Sheva Mental Health Center. Patient data were extracted from the medical records. Records were screened for the diagnosis made at admission and at the time of discharge, as well as for laboratory test results. Inclusion criteria consisted of age (12.5–18 years) and a psychiatric diagnosis of depression, schizophrenia, borderline personality disorder, or conduct disorder. A patient's data were excluded if any of the following exclusion criteria were met: 1) Use of alcohol (self-reported or reported by parents, social workers, and/or teachers) or drugs (either psychoactive, nonpsychoactive, or recreational) during the month prior to blood examination; 2) Evidence of infection upon admission or a white blood cell count of >12,000 cells/mL; 3) Evidence of hormonal disorders, diabetes, thyroid dysfunction, heart disease, trauma, blood disease, or any infection diagnosed upon or during hospitalization; and 4) Evidence of any neurological disorder, except for attention-deficit/hyperactivity disorder (ADHD). A total of 108 patients met all these criteria and were considered in subsequent analyses. Control subjects (n=79) were randomly selected from the medical records of the Soroka University Medical Center, using the pre-operational blood tests of otherwise healthy patients designated to undergo elective surgery, mostly plastic surgery. Inclusion criterion for the control group was age (12.5–18 years). The exclusion criteria were: 1) A record of neurological or infectious diseases or 2) A record of drug use or abuse. Approval to analyze the data of patients and controls was obtained from the internal review board of the appropriate medical institutions.

Psychiatric classification and background variables

Patients were categorized into two major groups, suicidal and nonsuicidal patients, based on the presence of suicidal ideation at any point during the week prior to the psychiatric examination using questions 6 and 8 from the Brief Psychiatric Rating Scale for Children and Adolescents. The assessment was conducted by raters blinded to the blood test results and psychiatric diagnoses (Overall and Pfefferbaum 1982). Control individuals were all assumed to be nonsuicidal. In addition, patients were classified into four groups according to their major diagnosis (Table 1) as defined in the axis-I American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, 4th ed., Text Revision (DSM-IV-TR) criteria (American Psychiatric Association 2000): 1) Schizophrenia, including schizophreniform and schizoaffective disorders; 2) major depression; 3) bipolar disorder; and 4) others, including conduct and borderline personality disorder, as the main diagnoses. Age, gender, smoking habits, and PLT count were assessed for each individual. Smoking habits were not recorded for control individuals. All PLT count measurements were generated in computerized hematology laboratories at the Beer-Sheva Mental Health Center or the Soroka Medical Center.

Table 1.

The Distribution of Major Diagnoses

	Number of patients
Major Diagnosis	Total	Suicidal Thoughts	No suicidal Thoughts
Major depression (incl. bipolar in depressive state)	36	27	7
Of these, bipolar in depressive state	12	8	4
Bipolar manic/hypomanic (including cyclothimia)	11	1	10
Schizophrenia (including schizoaffective)	19	4	16
Conduct/borderline	38	18	20
Other (PTSD/No diagnosis)	6	2	4
TOTAL	108	50	58

The number of patients with each disorder according to DSM axis-I are shown, as well as the number of patients with or without suicidal thoughts for each disorder. Bipolar patients in the depressed state were grouped with patients suffering from major depression.

Institutional review board approval

The study was approved by the Hospital Beer-Sheva Mental Health Center Review Board and the Israeli Ministry of Health. The need for informed consent was waived because of the naturalistic retrospective nature of the study.

Statistical methods

Data were analyzed using the t-test for two independent samples, assuming equal variance, and one-way ANOVA to compare several groups. In addition, the equivalent nonparametric Wilcox test and Chi squared test were also used as appropriate. Statistical analysis was carried out using the R programming language (version 2.10.1). In all statistical tests, a significant cutoff of 5% (p=0.05) was used.

Results

We reviewed the records of 821 adolescents hospitalized between November 1, 1995 and October 31, 2004 at the Beer Sheva Mental Health Center. PLT counts of 108 drug-free patients with no evidence of relevant hematological, endocrine, infectious, or neurological disorder upon admission were identified. A single patient with a recent and significant suicidal act (i.e., low-height jump without any indication of trauma), and five patients with superficial cuts in the wrists were not excluded from the study group. For controls, as mentioned in the Methods section, data on PLT counts of 79 individuals were collected from records in the Soroka Medical Center (see Methods section for detailed description of the inclusion/exclusion criteria).

The background variables of the different subgroups revealed few significant differences (Table 2). The average age of suicidal patients was slightly lower than that of the control group (14.7±1.5 years, as compared with15.5±1.6 years; p=0.005, Student's t-test) and of the nonsuicidal patients group (14.7±1.5years, as compared to 15.5±1.6 years; p=0.01, Student's t-test). These differences are not likely to introduce significant differences in PLT levels among these groups, as no significant correlation was observed between PLT and age when the entire population was considered (p=0.9, Pearson's test). However, to further reduce the possibility that age differences contribute to differences in PLT counts, the groups were balanced in terms of patients' ages. This process only marginally changed the results, such that all of the differences detected without age balancing were also found to be significant with such balancing and vice versa.

Table 2.

A Comparison of Background Variables in Patients Grouped According to Diagnosis or Symptoms

Compared groups		N		Age±SD (Years)		Gender (% females)		Smoking (#)
I	II	I	II	I	II	I	II	I	II
All	—	187	n.a.	15.3±1.6	n.a.	51	n.a.	12	n.a.
Patients	Controls	108	79	15.1±1.6	15.6±1.6	49	48	12	n.d.
Suicidal	Controls
without age correction		50	79	14.7±1.5	15.6±1.6	56	48	10	n.d.
with age correction		50	63^*	14.7±1.5	15.2±1.4	56	48	10	n.d.
Non-suicidal	Controls	58	79	15.4±1.6	15.6±1.6	43	48	2	n.d.
Suicidal	Other cases
without age correction		50	58	14.7±1.5	15.4±1.6	56	43	10	2
with age correction		50	46^*	14.7±1.5	15.0±1.4	56	43	10	2
Suicidal (Not depressed)	Other cases (Not depressed)
without age correction		23	51	14.6±1.6	15.4±1.5	48	37	3	2
with age correction		23	41^*	14.6±1.6	15.0±1.4	48	39	3	2
Non-depressed suicidal	Controls
without age correction		23	79	14.6±1.6	15.6±1.6	48	48	3	n.d.
with age correction		23	63^*	14.6±1.6	15.2±1.4	48	48	3	n.d.

The background variables are compared for different groups of subjects. The number of subjects included in each group (N), the mean and standard deviation of their ages (Age±SD), the % females and the number of smokers is given for the two groups (arbitrarily denoted as I or II). Statistically significant differences (p<0.05, Student's t-test) are marked in bold. For comparison, where a significant difference in patients' age was found, the analysis was repeated with age correction by resampling one of the groups to reduce the age difference. The group that was resampled is marked with an asterisk (^*).

n.a.=not applicable; n.d.=not determined.

No significant differences were found in the number of smokers in those groups for which smoking status was known (i.e., suicidal and nonsuicidal patients) or in the fraction of females in the different groups.

PLT counts were found to be significantly higher in the group of suicidal patients (Table 3). The mean PLT count in this group was 298±55 cells/mL×1000. In comparison, the mean PLT count of healthy controls was 262±47 cells/mL×1000 (p=0.0005, Student's t-test), and 260±60 for inpatients without suicidal tendencies (p=0.0004, Student's t-test).

Table 3.

PLT Counts in Adolescents with or without Suicidal Ideation

Compared groups	N	Group I Average±SD	N	Group II Average±SD	p-value(t-statistic)
Suicidal vs. controls	50	300±53	79	254±52	<10⁻²⁶ (4.8)
With age correction	50	300±53	63^*	254±54	<10⁻²⁶ (4.5)
Nonsuicidal cases vs. controls	58	254±67	79	254±52	0.96 (−0.04)
Suicidal vs. nonsuicidal cases	50	300.2±53.3	58	253.9±67	1e–04 (4.0)
With age correction	50	300±53	46^*	255±70	7e–4 (3.5)
Non-depressed suicidal vs. other non-depressed cases	23	292.5±55	51	243.5±58.7	0.0011 (3.4774)
With age correction	23	292±55	41^*	246±64	0.004 (3.0)
Suicidal, not depressed vs. controls	23	292±55	79	254±55	0.006 (3.0)

The PLT values in specific sub-groups of hospitalized adolescents were compared. The statistical significance of each comparison is provided (p-value and Student's t test statistics). For comparison, where a significant difference in patients' age was found, the analysis was repeated with age correction by resampling one of the groups to reduce the age difference. The group that was resampled is marked with an asterisk (^*).

A large fraction of the patients with suicidal tendencies in this study had major depression (25 of the 51 patients considered; see Table 1). This raises the possibility that the observed differences in PLT levels are associated with the depression, rather than with the suicidal thoughts. To test this hypothesis, we compared the PLT levels of nondepressed suicidal patients with nondepressed, nonsuicidal patients and with healthy controls. Significantly higher PLT counts were observed in the nondepressed suicidal patients than in either control group (Table 3).

The differences between PLT levels in suicidal patients, nonsuicidal patients, and healthy controls reported previously were also detected when the three groups were compared using the ANOVA test (p=0.0001).

Discussion

In this retrospective study, we compared the platelet counts of patients with suicidal thoughts, patients without such thoughts, and nonsuicidal, healthy controls. To the best of our knowledge, this study is the first to compare PLT counts among adolescents admitted with or without suicidal ideation. Suicidal patients were found to have significantly higher PLT than did nonsuicidal patients or healthy controls. No statistically significant differences were found among the groups with respect to the number of smokers or gender distribution. The stress associated with psychiatric hospitalization (Beattie et al. 2009), which could affect PLT levels (Van Ischoot et al. 1982), might also be a confounder. Therefore, we compared subjects with different psychopathologies, as well as healthy adolescents, a comparison that may have emphasized differences within the population of the hospitalized patients. Both groups of patients, that is, suicidal and nonsuicidal, are expected to experience similar stress. We therefore propose that the design presented here, namely, comparing inpatients with different complaints and diagnosis, may be more sensitive and robust in identifying markers that differ among diseases than would comparing patients with healthy nonhospitalized controls.

Another possible confounder is the large overlap between suicidality and depression. One cannot ignore the possibility that the elevation in PLT is mainly caused by the state of depression. This hypothesis was, however, rejected by showing differences in PLT counts between non depressed patients with and without suicidal ideation. As such, it is possible that depression contributes to the elevation of PLT, even if it does not explain it. Follow-up studies with larger populations are required to evaluate the independent contribution of depression to PLT levels. Finally, age could be a confounder. A small but significant difference was observed between the age of patients with or without suicidal thoughts (Table 2). However, controlling for these differences by resampling did not eliminate the difference in PLT counts.

These findings suggest a possible association between suicidal thoughts and elevated PLT counts. In addition, they suggest that depressed suicidal adolescents constitute a specific group of adolescents. Alternatively, this difference might reflect depression in progression that was misdiagnosed at this stage of the disorder as conduct or borderline personality disorder (Valevski et al. 2001; Farbstein et al. 2002) leading to the mis-assignment of such patients to the borderline personality/conduct disorder group. Further investigation is required to clarify these points.

This work is a pilot study and subsequent studies conducted with larger samples in multiple centers are needed to substantiate our findings. Different age groups, the effects of medication and drugs, and other laboratory measures, as well as other symptoms should be investigated in patients having suicidal thoughts. Moreover, searching for additional PLT-related biomarkers may enable the discrimination between suicidal and nonsuicidal adolescents and eventually lead to the identification of the mechanism(s) underlying suicidal ideation/behavior.

Can elevated PLT counts serve as diagnostic aid? Obviously, as the differences we report are within the normal range, the discriminative power of PLT counting is limited. However, understanding the mechanism underlying the observed differences may point to new markers that could assist in diagnosis.

Conclusions

Hospitalized adolescents with suicidal thoughts are reported to have elevated PLT counts within the normal range.

Clinical Significance

Our findings suggest a possible association between suicidal thoughts and elevated PLT counts. Although the discriminative power of PLT counting is limited, understanding the mechanism underlying the observed differences may point to new markers that could assist in the diagnosis of suicidality.

Footnotes

Acknowledgments

We thank Hagit Cohen for her assistance in analyzing the results, Asher Moser and Dr. Hadas Ner Gaon for their assistance in manuscript preparation, and Yael Djorno for helping in data acquisition. Finally we thank our anonymous reviewers for their helpful suggestions.

Disclosures

No competing financial interests exist. This work was partially funded by the National Institute for Biotechnology in the Negev through a scholarship to Eitan Rubin.

References

Abram

, Choe

, Washburn

, Teplin

, King

, Dulcan

. Suicidal ideation and behaviors among youths in juvenile detention. J Am Acad Child Adolesc Psychiatry, 47:291–300. 2008.

American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th Text Revision (DSM-IV-TR) Washington, DC: American Psychiatric Association, 2000.

Apter

, King

. Management of the depressed, suicidal child or adolescent. Child Adolesc Psychiatr Clin N Am, 15:999–1013. 2006.

Ataoglu

, Canan

. Mean platelet volume in patients with major depression: Effect of escitalopram treatment. J Clin Psychopharmacol, 29:368–371. 2009.

Beattie

, Shannon

, Kavanagh

, Mulholland

. Predictors of PTSD symptoms in response to psychosis and psychiatric admission. J Nerv Ment Dis, 197:56–60. 2009.

Brent

. Assessment and treatment of the youthful suicidal patient. Ann N Y Acad Sci, 932:106–131. 2001.

Briley

, Langer

, Raisman

, Sechter

, Zarifian

. Tritiated imipramine binding sites are decreased in platelets of untreated depressed patients. Science, 209:303–305. 1980.

Canan

, Dikici

, Kutlucan

, Celbek

, Coskun

, Gungor

, Aydin

, Kocaman

. Association of mean platelet volume with DSM-IV major depression in a large community-based population: The MELEN study. J Psychiatr Res, 46:298–302. 2012.

Dervic

, Brent

, Oquendo

. Completed suicide in childhood. Psychiatr Clin N Am, 31:271–91. 2008.

10.

Eaton

, Kann

, Kinchen

, Ross

, Hawkins

, Harris

, Lowry

, McManus

, Chyen

, Shanklin

, Lim

, Grunbaum

, Wechsler

. Youth risk behavior surveillance—United States, 2005. MMWR Surveill Summ, 55:1–108. 2006.

11.

Farbstein

, Dycian

, Gothelf

, King

, Cohen

, Kron

, Apter

. A follow-up study of adolescent attempted suicide in Israel. J Am Acad Child Adolesc Psychiatry, 41:1342–1349. 2002.

12.

Geller

, Yuwiler

, Freeman

, Ritvo

. Platelet size, number, and serotonin content in blood of autistic, childhood schizophrenic, and normal children. J Aut Dev Disord, 18:119–126. 1988.

13.

Greydanus

, Calles

Jr . Suicide in children and adolescents. Prim Care, 34:259–273abstract vi2007.

14.

Hitzemann

, Sautter

, Hitzemann

, Hirschowitz

, Garver

. Platelet monoamine oxidase activity in the psychoses: Relationship to symptoms and lithium response. Psychiatry Res, 11:211–221. 1984.

15.

Hrdina

, Bakish

, Chudzik

, Ravindran

, Lapierre

. Serotonergic markers in platelets of patients with major depression: upregulation of 5-HT2 receptors. J Psychiatry Neurosci, 20:11–19. 1995.

16.

Kessler

, Borges

, Walters

. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch Gen Psychiatry, 56:617–626. 1999.

17.

Laje

, Paddock

, Manji

, Rush

, Wilson

, Charney

, McMahon

. Genetic markers of suicidal ideation emerging during citalopram treatment of major depression. Am J Psychiatry, 164:1530–1538. 2007.

18.

Longenecker

. The Platelets: Physiology, Pharmacology. Orlando, FL: Academic Press, Inc., 1985.

19.

McClure

. Suicide in children and adolescents in England and Wales 1970–1998. Br J Psychiatry, 178:469–474. 2001.

20.

Muller–Oerlinghausen

, Roggenbach

, Franke

. Serotonergic platelet markers of suicidal behavior—do they really exist? J Affect Disord, 79:13–24. 2004.

21.

O'Carroll

, Berman

, Maris

, Moscicki

, Tanney

, Silverman

. Beyond the Tower of Babel: A nomenclature for suicidology. Suicide Life Threat Behav, 26:237–252. 1996.

22.

Overall

, Pfefferbaum

. The Brief Psychiatric Rating Scale for Children. Psychopharmacol Bull, 18:10–16. 1982.

23.

Pandey

, Pandey

, Janicak

, Marks

, Davis

. Platelet serotonin-2 receptor binding sites in depression and suicide. Biol Psychiatry, 28:215–222. 1990.

24.

Pfeffer

. Suicide in mood disordered children, adolescents. Child Adolesc Psychiatr Clin N Am, 11:639–647. 2002.

25.

Pfeffer

, Conte

, Plutchik

, Jerrett

. Suicidal behavior in latency-age children: An empirical study. J Am Acad Child Psychiatry, 18:679–692. 1979.

26.

Pfeffer

, McBride

, Anderson

, Kakuma

, Fensterheim

, Khait

. Peripheral serotonin measures in prepubertal psychiatric inpatients and normal children: Associations with suicidal behavior and its risk factors. Biol Psychiatry, 44:568–577. 1998.

27.

Pfeffer

, Plutchik

. Psychopathology of latency age children. Relation to treatment planning. J Nerv Ment Dis, 170:193–197. 1982.

28.

Rao

, Hawellek

, Papassotiropoulos

, Deister

, Frahnert

. Upregulation of the platelet Serotonin2A receptor and low blood serotonin in suicidal psychiatric patients. Neuropsychobiology, 38:84–89. 1998.

29.

Rehavi

, Weizman

. The platelets as a peripheral model of serotoninergic function in child psychiatry. 141–146. 1990.

30.

Reinherz

, Giaconia

, Silverman

, Friedman

, Pakiz

, Frost

, Cohen

. Early psychosocial risks for adolescent suicidal ideation and attempts. J AmAcad Child Adolesc Psychiatry, 34:599–611. 1995.

31.

Sadock

, Sadock

, Ruiz

. Kaplan & Sadock's Comprehensive Textbook of Psychiatry, 9th. Baltimore/Boston: Lipponcott Williams & Wilkins, 2009.

32.

Sneddon

. Platelets as a model for monoamine-containg neurons. Progress in Neurobiology, 2:151–198. 1973.

33.

Spreux–Varoquaux

, Alvarez

, Berlin

, Batista

, Despierre

, Gilton

, Cremniter

. Differential abnormalities in plasma 5-HIAA and platelet serotonin concentrations in violent suicide attempters: Relationships with impulsivity and depression. Life Sci, 69:647–657. 2001.

34.

Stahl

. Platelets as pharmacologic models for the receptors, biochemistry of monoaminergic neurons. The Platelets. Physiology and Pharmacology. Longenecker

G.L.

Orlando, FL: Academic Press, 308–340. 1985.

35.

Stahl

, Woo

, Mefford

, Berger

, Ciaranello

. Hyperserotonemia and platelet serotonin uptake and release in schizophrenia and affective disorders. Am J Psychiatry, 140:26–30. 1983.

36.

Valevski

, Apter

, Wiezman

, Tyano

. Stability of diagnosis: A 20-year retrospective cohort study of Israeli psychiatric adolescents inpatients. J Adolesc, 24:625–633. 2001.

37.

Van

Ischoot K

, Liesse

, Lauwers

. Neurotic triad and change in blood platelet number in reaction to stress. Encephale, 8:501–505. 1982.