- Email: [email protected]
5-HIAA ratios in suicide attempters
European Neuropsychopharmacology 9 (1999) 399–405 www.elsevier.com / locate / euroneuro
Reduced cerebrospinal HVA concentrations and HVA / 5-HIAA ratios in suicide attempters Monoamine metabolites in 120 suicide attempters and 47 controls a ¨ a , *, Christer Alling b , Kaj Blennow c , Goran ¨ ´ a , Lil Traskman-Bendz ¨ Gunnar Engstrom Regnell a
Department of Clinical Neuroscience, Psychiatry Section, University Hospital, S-221 85 Lund, Sweden b Department of Medical Neurochemistry, Lund University Hospital, S-22185 Lund, Sweden c ¨ ¨ Hospital, S-431 80 Molndal , Institute of Clinical Neuroscience, Department of Psychiatry and Neurochemistry, Gothenburg University, Molndal Sweden Received 21 January 1999; received in revised form 22 March 1999; accepted 24 March 1999
Abstract Dysfunctions of central monoaminergic systems are important elements of the leading biological hypotheses of suicide and depression. The purpose of the present paper was to study the levels and the relationships between the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA), the dopamine metabolite homovanillic acid (HVA) and the norepinephrine metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) in the cerebrospinal fluid (CSF) in 120 hospitalised suicide attempters and 47 controls (healthy volunteers or patients admitted for minor surgery). The suicide attempters showed significantly lower HVA levels (174682 vs. 216696 nmol / L, P50.004), HVA / 5HIAA ratios (1.660.5 vs. 2.160.6, P50.0001) and HVA / MHPG ratios (4.262.1 vs. 4.861.7, P50.02) than the controls. The correlations between the monoamine metabolites were markedly lower in patients than in controls. CSF 5-HIAA showed no significant differences between patients and controls (107640 vs. 108651 nmol / L) or between violent and non-violent attempters (112658 vs. 105633 nmol / L). The monoamine metabolites showed no significant differences between survivors and patients who subsequently completed suicide, or between suicide attempters subgrouped by psychiatric diagnoses. The results suggest that low HVA levels and altered relationships between the monoamine metabolites are associated with suicidal behaviour. 1999 Elsevier Science B.V. All rights reserved. Keywords: Monoamine metabolites; Serotonin; Dopamine; Suicide; Depression
1. Introduction During the last decades, dysfunctions of central monoaminergic systems have been important elements of the leading biological hypotheses of suicide as well as depression, anxiety disorders and schizophrenia (e.g., Carlsson, 1978; Coccaro et al., 1989; Mann et al., 1992; Meltzer, 1989; Kahn et al., 1988, Potter and Manji, 1993; Westenberg and Verhoeven, 1988). Based on the assumption that *Corresponding author. Present address: Department of Community ¨ Sweden. Tel.: Medicine, Malmo¨ University Hospital, S-20502 Malmo, 146-403-3-3435; fax: 146-403-3-6215. ¨ E-mail address: [email protected] (G. Engstrom)
the concentrations of monoamine metabolites in the cerebrospinal fluid (CSF) reflect the monoaminergic activities in the brain (Stanley et al., 1985), many investigators have studied the levels of the monoamine metabolites in CSF. Suicide and suicidal behaviour are phenomenas which occur in patients with different psychiatric diagnoses (Harris and Barraclough, 1997). Ahrens and Linden (1996) even suggested that this behaviour could be regarded as a separate syndrome. Biological studies in this field have repeatedly associated low levels of the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA) in CSF with suicidal behaviour, especially with violent methods ˚ ¨ (Asberg et al., 1976; Traskman et al., 1981). Relationships between suicidality and low CSF 5-HIAA have been
0924-977X / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0924-977X( 99 )00016-4
400
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
reported in different diagnostic categories, e.g. in patients with schizophrenia, depression, alcoholism, adjustment disorders, personality disorders and in violent offenders (Linnoila et al., 1983; Banki et al., 1984; Gardner et al., 1990; Ninan et al., 1984; Lopez-Ibor et al., 1985). There are also negative findings however (Lemus et al., 1990; Roy et al., 1990; Westenberg and Verhoeven, 1988), especially in bipolar suicidal patients (Berretini et al., 1986; Roy-Byrne et al., 1983). Nevertheless, a metaanalytic review of the relationships between CSF monoamine metabolites and suicidal behaviour found strong evidence for lower CSF 5-HIAA in suicide attempters than in psychiatric controls (Lester, 1995). There are also some studies linking the dopamine metabolite homovanillic acid (HVA) to suicidal behaviour (Roy et al., 1986; Jones et al., 1990) and low CSF HVA levels is the most consistent finding when comparing CSF monoamine metabolites in depressive subjects and controls (Goodwin and Jamison, 1990). The findings of the norepinephrine metabolite 3methoxy-4-hydroxyphenylglycol (MHPG) in CSF are inconsistent in studies of suicide attempters as well as major depression (Lester, 1995; Goodwin and Jamison, 1990). The theories of the casual relationships between monoamines and psychiatric illness have gradually turned in favour of hypotheses involving interactions of several neurotransmitters rather than deficiency of one specific monoamine (Carlsson et al., 1997; Potter and Manji, 1993). Smaller correlations, which possibly reflect altered interactions, between the CSF monoamine metabolites in drug-free schizophrenics than in controls, and increasing correlations after treatment, are reported by Hsiao et al. (1993). In treatment-resistant depressives the correlations were significantly smaller than in responders (Hsiao et al., 1987). Other possible measures of the interactions between the monoamines are the ratios between the CSF monoamine metabolites (Roy et al., 1990; Hsiao et al., 1987, 1993; Reddy et al., 1992). The aim of the present study has been to investigate the CSF concentrations of 5-HIAA, HVA and MHPG and to study the relationships between the metabolites in a large sample of suicide attempters with regard to psychiatric diagnoses and suicidal behaviour, and in comparison to a control group.
2. Experimental procedures
37.4611.5, range: 19–61 years) and 65 were women (mean age 38.8614.5, range 19–72 years). The mean body height was 172.368.9 cm. The patients were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R; American Psychiatric Association, 1987) by two independent psychiatrists. The patients were categorised according to suicide method into violent and non-violent attempters. Suicide attempters who took drug overdoses or made single wrist cuts were denoted ‘‘non-violent’’, and those who used other methods or combinations of two or more methods ¨ were denoted ‘‘violent’’ (Traskman et al., 1981). The patients were also categorised into ‘‘repeaters’’ who had made previous suicide attempts, and ‘‘non-repeaters’’ without previous suicide attempts. Information on subsequently completed suicides was retrieved from the unit for Forensic Medicine in Lund and from the National Cause of Death Register.
2.2. Controls The control group consisted of 47 volunteers from Gothenburg, Sweden, (30 men, 17 women) with a mean age of 44.1615.9 years (range 18–70 years) (Blennow et al., 1993a). Eighteen subjects (14 men, 4 women, mean age 61.265.4 years) were patients who had been admitted for minor surgery (e.g. coxarthrosis and prostate hyperplasia) under spinal anaesthesia. They were examined the day before the surgery. The CSF sample was drawn in connection with the lumbar puncture for the spinal anaesthesia. Twenty-nine subjects (16 men, 13 women, mean age: 33.569.2 years) were healthy volunteers that accepted a medical and psychiatric examination, this group was primarily recruited among ward staff. The mean body height of the controls was 174.668.6 cm. A careful medical history was taken from each subject and they all went through a physical and neurological clinical examination. A structured interview including specific questions on different psychiatric symptoms was undertaken. Subjects with history, symptoms or signs of psychiatric, neurological disorders, Type I diabetes, severe hypertension, malignant disease, or autoimmune infections were excluded. By definition, no subject received medication with CNS active drugs (Blennow et al., 1993a).
2.3. Lumbar punctures and monoamine metabolite assays
2.1. Patients The patients (n5120) were evaluated by a psychiatrist and a social worker with a structured psychiatric interview after a suicide attempt and then hospitalised at the Lund University Hospital in a psychiatric ward specialised in suicidal behaviour. They all gave informed consent to participate in the study. Mean age was 38.2613.2 years (range: 19–72 years). Fifty-five were men (mean age
The lumbar punctures were taken in the morning after at least 8 h. of fasting and bed rest since midnight. Neither patients nor controls had any specific diet prior to the examination. Samples were taken with the subjects in a sitting position. With exception of oral contraceptives and occasional benzodiazepines the patients were drug-free prior to the lumbar puncture. The mean drug free wash-out period was 14.867.4 days (range 4–57 days). Twelve ml
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
of spinal fluid was taken and gently mixed to avoid gradients, and then stored in 2808C in 2 ml aliquots. The CSF monoamine metabolite concentrations in the suicide attempters were analysed with gas chromatography–mass spectrometry (GC–MS) (Gjerris et al., 1987). Both GC–MS and high performance liquid chromatography (HPLC) were used for analyses of the monoamine metabolite concentrations in the controls (described in Blennow et al., 1993a, 1993b). The correlations between the values obtained by GC–MS and HPLC were above 99% (r.0.99) for all of HVA, 5-HIAA and MHPG (Blennow et al., 1993b). The inter- and intra-assay coefficients were 2–4% for all metabolites with both methods.
401
Table 2 Spearman correlation coefficients between monoamine metabolites, age and height in patients (upper right half, bold types), and controls (lower left half) HVA HVA 5-HIAA MHPG Age Height
5HIAA
MHPG
0.64*** 0.81*** 0.43** 0.35* 20.39*
0.47*** 0.52*** 20.51***
0.12 0.21* 0.31* 20.34*
Age
Height
20.01 0.08 0.06
20.30*** 20.18 20.13 20.07
20.22
***P,0.001, **P,0.01, *P,0.05.
ences between controls admitted for minor surgery and other healthy controls.
3.1. Age, height, and gender
2.4. Statistics Correlations between variables were calculated by means of Spearman correlation coefficients. Mann–Whitney U-test or Kruskal–Wallis non-parametric analysis of variance was used for comparisons of groups. As the monoamine metabolites were correlating with age and body height, we also performed analyses of covariance (ANCOVAs) with age and height as covariates in all comparisons of two or more groups.
In both patients and controls, all three monoamine metabolites were negatively correlated with height (Table 2). All three monoamine metabolites correlated significantly with age in the controls, while no significant association between the monoamine metabolites and age was found among the patients (Table 2). The monoamine metabolites showed no significant differences between men and women in patients or controls (data not shown).
3.2. Seasonal variation 3. Results The patients showed significantly lower levels of HVA (median: 172 vs. 204 nmol / L, Z52.87, P50.004) and lower HVA / 5-HIAA (median: 1.75 vs. 2.12, Z54.7, P, 0.0001) and HVA / MHPG (median: 4.35 vs. 4.98, Z52.4, P50.02) ratios than the controls (Table 1). Although not significantly, the patients also had lower MHPG (median 42 vs. 43 nmol / L, Z51.7, P50.08). After adjustments for age and body height, the results were essentially the same with significant differences between suicide attempters and controls in HVA (F514.5, P,0.001), HVA / 5-HIAA ratio (F543.0, P,0.0001) and HVA / MHPG ratio (F56.4, P5 0.013). When age and height had been taken into account, the monoamine metabolites showed no significant differ-
The suicide attempters were categorised into four groups according to time of the year: Fifty-three subjects were lumbar punctured in December–March, 22 in April or May, 10 in June–August, and finally 35 patients were lumbar punctured in September–November. The monoamine metabolites showed no significant differences between the groups (data not shown). After adjustments for age and height, there were similarly no seasonal variations among the controls.
3.3. Suicidality Eight patients (6 women, 2 men, mean age at examination: 40.9617 years) subsequently committed suicide. In this group 3 had major depression, 2 had dysthymia, 2 had
Table 1 Mean6SD of HVA, 5-HIAA, MHPG and HVA / 5HIAA and HVA / MHPG ratios in different categories of suicide attempters and controls HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
Patients (n5120) Controls (n547)
173.6682** 216.4696
106.7640 107.7651
42.469 45.169
1.660.5*** 2.160.6
4.262.1* 4.861.7
Violent (n526) Non-violent (n594)
188.66104 169.5675
112.1658 105.2633
42.067 42.5610
1.760.5 1.660.5
4.662.9 4.161.8
Repeaters (n558) Non-repeaters (n562) Suicides (n58)
167.0692 179.9693 196.56106
106.2634 107.2645 111.4646.1
41.6610 43.169 43.9611
1.660.5 1.760.5 1.760.5
4.161.8 4.362.4 4.562.3
***P,0.0001, **P,0.01, *P,0.05. Patients compared with controls.
402
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
Table 3 Monoamine metabolite concentrations (6SD) (nmol / L) in suicide attempters who took overdoses of neuroleptics or antidepressives, overdoses of other substances or who used other methods than intoxications
Neuroleptics or antidepressives (n524) Other substances (n565) No intoxication (n531)
HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
140661 181679 184699
104632 106634 110655
42610 43610 4169
1.460.5* 1.760.5* 1.760.5*
3.461.5 4.361.9 4.662.7
*P50.02 (Kruskal–Wallis non-parametric ANOVA).
adjustment disorders and one had depression not otherwise specified (NOS) at the baseline examination. No significant monoamine metabolite differences was observed between subsequently completed suicides and other suicide attempters (Table 1). The monoamine metabolites showed no significant differences between violent and non-violent attempters or between repeaters and non-repeaters (Table 1). Twenty-four patients had taken overdoses with neuroleptics (n54) or antidepressive drugs (n520), which may influence the concentrations of monoamine metabolites. The wash-out period in this group was 16.566 days. This group showed significantly lower HVA / 5HIAA ratio (p5 0.02) than patients who took overdoses of other drugs or who used other methods. The HVA levels were also lower among these patients, although the differences did not reach significance (Table 3). The 5-HIAA and MHPG concentrations showed only small differences.
3.4. Psychiatric diagnoses Thirty-six patients were diagnosed with major depression (MDD), 19 with dysthymia, 23 with adjustment disorders, 15 with depression not otherwise specified (NOS), 5 with alcohol dependence, 5 with alcohol abuse, two with schizophrenia, disorganised, two with schizoaffective disorder, two with panic disorder without agoraphobia, one each with schizophrenia, undifferentiated, psychotic disorder NOS, anxiety disorder NOS, panic disorder with agoraphobia, obsessive compulsive disorder, cyclothymia, polysubstance dependence, anorexia nervosa, bulimia nervosa, and finally, two patients did not receive any axis I diagnosis (both had borderline personality disorder). The patients were categorised according to their axis I diagnosis into MDD (n536), dysthymia (n5
19), adjustment disorders (n523), and finally into a group with other diagnoses or without any axis I diagnosis (‘‘other diagnoses’’ n542). Twenty-eight patients received two or three axis I diagnoses, eighteen of them had an alcohol- or substance abuse disorder. The monoamine metabolite levels did not differ significantly between the diagnostic groups (Table 4). The differences remained non-significant after adjustments for age and height.
3.5. Relationships between monoamine metabolites The HVA / 5-HIAA and HVA / MHPG ratios were significantly different between patients and controls (Table 1). The correlation coefficients between the monoamine metabolites were markedly lower in the patients than controls (rsp for HVA / 5-HIAA: 0.64 vs. 0.81; for HVA / MHPG: 0.12 vs. 0.43; for 5-HIAA / MHPG: 0.21 vs. 0.47) (Table 2).
4. Discussion In the present study, the levels of CSF HVA and the HVA / 5-HIAA and HVA / MHPG ratios were significantly lower among the patients than the controls. The levels of CSF 5-HIAA showed no significant differences between suicide attempters and controls. Previous studies of cerebrospinal monoamine metabolites in suicide attempters have associated low CSF 5-HIAA with suicidality, the findings of HVA in previous studies are however inconsistent (Lester, 1995). As suicide attempters are clinically very heterogeneous, it is possible that the sample of the
Table 4 Mean6SD of HVA, 5-HIAA, MHPG and HVA / 5HIAA and HVA / MHPG ratios in suicide attempters with different psychiatric diagnoses
Major depression (n536) Dysthymia (n519) Adjustment disorders (n523) Other diagnoses (n542)
HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
168.26101 191.7681 166.3649 174.1682
105.8654 114.6642 108.8625 102.7631
41.2610 42.868 42.966 42.9611
1.660.5 1.760.4 1.660.4 1.760.6
4.262.7 4.561.7 4.061.4 4.262.0
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
present study could explain the differences between our results and those by others. A majority of the patients in the present study had mood disorders or mood-related disorders. This could be the major explanation for the low HVA levels and HVA / 5-HIAA ratios rather than the suicidality. Low HVA levels is the most consistent finding in studies of CSF monoamine metabolites in depressed patients (Goodwin and Jamison, 1990), and previous studies which have associated low HVA levels with suicidal behaviour have studied samples of depressed suicide attempters (Roy et al., 1986; Jones et al., 1990; ¨ Traskman et al., 1981). Thus, low HVA levels and HVA / 5HIAA ratios might be markers of depression rather than suicidality. We could not find any differences in CSF 5-HIAA levels between suicide attempters and controls or between violent or nonviolent attempters. It has been suggested that individuals who volunteer in lumbar punctures are more impulsive than other healthy controls (Gustavsson et al., 1997). If that is the case, the controls would have lower CSF 5-HIAA than expected and the differences between patients and controls would be reduced. However, since the 5-HIAA levels of the controls in the present study are at the same level as in control groups from other studies ˚ ¨ (e.g., Traskman et al., 1981; Asberg et al., 1984) this explanation seems to be unlikely. We could not replicate the associations between low CSF 5-HIAA and subsequently completed suicide reported from previous follow-up studies of suicide attempters (Lester, 1995; ¨ et al., 1994). Relationships between low CSF Nordstrom 5-HIAA and high lethality of suicidal behaviour has also been reported (Mann and Malone, 1997). Most patients in the present study were non-violent attempters and few attempts were very serious. It is possible that our way of categorising the attempts into ‘‘violent’’ or ‘‘non-violent’’ didn’t reflect the violence or intention of the attempts well enough or that the variance in violence or lethality was too small. However, these explanations are challenged by the fact that the patients who later died from suicide also had CSF 5-HIAA at the same level as the non-violent attempters. Several investigators have discussed the influence of various confounders, such as age, height, gender or seasonal variation, and the importance of adjusting the monoamine metabolite levels for these factors (e.g. Blen˚ now et al., 1993a,b; Asberg et al., 1984; Nordin et al., 1995; Brewerton et al., 1988). When all subject were taken together, the monoamine metabolites were associated with age (positively) and height (negatively) and no significant difference between men and women was found. The correlations between age, height and the monoamine metabolites differed between patients and controls (Table 2), and the relationships between the monoamine metabolites and age, height and sex were weak among the suicide attempters. The results were essentially the same after statistical adjustments for age and height. There were no
403
significant seasonal variation of the monoamine metabolites in patients or controls. A relevant issue is however whether the results could have been confounded by the fact that patients and controls were from different cities. Whether there are regional variations of the monoamine metabolite levels in the south of Sweden is unknown and this remains a possible confounder. It is noteworthy however that the most significant difference between patients and controls was the HVA / 5-HIAA ratio, a measure which is relatively independent of different confounding factors, and a measure which in contrast to the different monoamine metabolite levels is independent of accuracy in monoamine measurements. The relationships between the different monoamine metabolites, rather than the metabolites themselves, have received increasing interest over the last years. Altered monoaminergic interactions have previously been associated with low response to antidepressive treatment (Hsiao et al., 1987) and with psychiatric symptoms in schizophrenics (Kahn et al., 1993). It is therefore noteworthy that in the present study, the HVA / 5-HIAA ratio showed bigger differences between patients and controls than the HVA levels alone. The subjects who had taken overdoses of neuroleptics or antidepressive drugs had lower HVA concentrations than other suicide attempters. However, as suicide attempters who used other methods than intoxications also had significantly lower HVA / 5HIAA ratio than controls, the intoxications could not explain the difference between patients and controls. The HVA / 5HIAA ratio was significantly different between patients and controls also when age and height were covaried out. It is also noteworthy that the controls showed considerably higher correlations between the monoamine metabolites than the patients. Both the ratios and correlations between CSF monoamine metabolites are measures that possibly reflect the interactions between the monoamines in CNS. These constructs are however not validated and so far it is necessary to interpret these findings with care. Smaller correlations between the metabolites in patients than in controls could reflect independence, or ‘‘impaired cooperation’’ between the monoamine systems, but also a biological heterogeneity either among patients or controls. Nevertheless, further study of the interactions between the monoamine systems might be important in order to improve our understanding of psychiatric disorders and suicidal behaviour. In summary, the results of the present study suggest that HVA and the relationships between HVA and 5-HIAA might be interesting markers in the study of suicidal behaviour and mood disorders. The markedly lower correlations coefficients between the monoamine metabolites in patients than controls suggest that the interactions between the monoamine systems differ between patients and controls. Our negative findings regarding the CSF 5-HIAA levels suggest that serotonergic dysfunctions are not present in all groups of suicide attempters. This puts
404
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
forward the need of studies comparing well-defined groups of suicidal patients.
Acknowledgements ¨ This study was supported by the Malmohus County, the Swedish Medical Research Council Nos. 08319, 714185, ¨ OM Perssons foundation, the Sjobring Foundation and funds from the University of Lund.
References Ahrens, B., Linden, M., 1996. Is there a suicidality syndrome independent of specific major psychiatric disorders? Results from a split half multiple regression analysis. Acta Psychiatr. Scand. 94, 79–86. American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd Ed. (Revised). APA, Washington, DC. ˚ ¨ ´ P., 1976. 5-HIAA in the cerebrospinal Asberg, M., Traskman, L., Thoren, fluid. A biochemical suicide predictor? Arch. Gen. Psychiatry 33, 1193–1197. ˚ ˚ ´ Asberg, M., Bertilsson, L., Martensson, B., Scalia-Tomba, G.-P., Thoren, ¨ P., Traskman-Bendz, L., 1984. CSF monoamine metabolites in melancholia. Acta Psychiatr. Scand. 69, 201–219. ´ M., Papp, Z., Kurcz, M., 1984. Biochemical markers Banki, C.M., Arato, in suicidal patients. Investigations with cerebrospinal fluid monoamine metabolites and neuroendocrine tests. J. Affect. Dis. 6, 341–350. Berretini, W.H., Nurnberger, J.I., Narrow, W., Simmons-Alling, S., Gershorn, E.S., 1986. Cerebrospinal fluid studies of bipolar patients with and without a history of suicide attempts. Ann. NY Acad. Sci. 487, 197–201. ˚ Blennow, K., Wallin, A., Gottfries, C.G., Karlsson, I., Mansson, J.E., ¨ C., Svennerholm, L., 1993a. Cerebrospinal fluid Skoog, I., Wikkelso, monoamine metabolites in 114 healthy individuals 18–81 years of age. Eur. Neuropsychopharmacol. 3, 55–61. ˚ Blennow, K., Wallin, A., Gottfries, C.G., Mansson, J.E., Svennerholm, L., 1993b. Concentration gradients for monoamine metabolites in lumbar cerebrospinal fluid. J. Neural. Transm. (P-D Sect) 5, 5–15. Brewerton, T.D., Berrettini, W.H., Nurnberger, J.I., Linnoila, M., 1988. Analysis of seasonal fluctuations of CSF monoamine metabolites in normal controls: findings with 5HIAA and HVA. Psychiatry Res. 23, 257–265. Carlsson, A., 1978. Antipsychotic drugs, neurotransmitters, and schizophrenia. Am. J. Psychiatry 135, 164–173. Carlsson, A., Hansson, L.O., Waters, N., Carlsson, M., 1997. Neurotransmitter aberrations in schizophrenia: new perspectives and therapeutic implications. Life Sci. 61, 75–94. Coccaro, E.F., Siever, L.J., Klar, H.M., Maurer, G., Cochrane, K., Cooper, T.B., Mohns, R.C., Davis, K.L., 1989. Serotonergic studies in patients with affective and personality disorders. Arch. Gen. Psychiatry 46, 587–599. Gardner, D.L., Lucas, P.B., Cowdry, R.W., 1990. CSF metabolites in borderline personality disorder compared with normal controls. Biol. Psychiatry 28, 247–254. ¨ Gjerris, A., Werdelin, L., Gjerris, F., Sorensen, P.S., Rafaelsen, O.J., Alling, C., 1987. CSF-amine metabolites in depression, dementia and in controls. Acta Psychiatr. Scand. 75, 619–628. Goodwin, F.K., Jamison, K.R., 1990. In: Manic-depressive Illness, Oxford University Press, New York and Oxford, pp. 416–502. ˚ Gustavsson, J.P., Asberg, M., Schalling, D., 1997. The healthy control subject in psychiatric resaerch: Impulsiveness and volunteer bias. Acta Psychiatr. Scand. 96, 325–328.
Harris, E.C., Barraclough, B., 1997. Suicide as an outcome for mental disorders. A meta-analysis. Br. J. Psychiatry 170, 205–228. ˚ Hsiao, J.K., Agren, H., Bartko, J.J., Rudorfer, M.V., Linnoila, M., Potter, W.Z., 1987. Monoamine neurotransmittor interactions and the prediction of antidepressant response. Arch. Gen. Psychiatry 44, 1078–1083. Hsiao, J.K., Colison, J., Bartko, J.J., Doran, A.R., Konicki, P.E., Potter, W.Z., Pickar, D., 1993. Monoamine neurotransmittor interactions in drug-free and neuroleptic-treated schizophrenics. Arch. Gen. Psychiatry 50, 606–614. ¨ Jones, J.S., Stanley, B., Mann, J.J., Frances, A.J., Guido, J.R., TraskmanBendz, L., Winchel, R., Brown, R.P., Stanley, M., 1990. CSF 5-HIAA and HVA concentrations in elderly depressed patients who attempted suicide. Am. J. Psychiatry 147, 1225–1227. Kahn, R.S., Davidson, M., Knott, P., Stern, R.G., Apter, S., Davis, K.L., 1993. Effect of neuroleptic medication on cerbrospinal fluid metabolite concentrations in schizophrenia. Arch. Gen. Psychiatry 50, 599–605. Kahn, R.S., van Praag, H.M., Wetzler, S., Asnis, G., Barr, G., 1988. Serotonin and anxiety revisited. Biol. Psychiatry 23, 189–208. Lemus, C.Z., Lieberman, J.A., Johns, C.A., Pollack, S., Bookstein, P., Cooper, T.B., 1990. CSF 5-hydroxyindoleacetic acid levels and suicide attempts in schizophrenia. Biol. Psychiatry 27, 926–929. Lester, D., 1995. The concentration of neurotransmitter metabolites in the cerebrospinal fluid of suicidal individuals: a meta-analysis. Pharmacopsychiatry 28, 45–50. Linnoila, M., Virkkunen, M., Scheinin, M., Nuutila, A., Rimon, R., Goodwin, F., 1983. Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sci. 33, 2609–2614. Lopez-Ibor, J.J., Saiz-Ruiz, J., Perez de los Cobos, J.C., 1985. Biological correlations of suicide and aggressivity in major depressions (with melancholia): 5-HIAA and cortisol in CSF, dexamethasone suppression test and therapeutic response to 5-hydroxytryptophan. Neuropsychobiology 14, 67–74. Mann, J.J., McBride, P.A., Brown, R.P., Linnoila, M., Leon, A.C., DeMeo, M., Mieczkowski, T., Myers, J.E., Stanley, M., 1992. Relationships between central and peripheral serotonin indexes in depressed and suicidal psychiatric inpatients. Arch. Gen. Psychiatry 49, 442–446. Mann, J.J., Malone, K.M., 1997. Cerebrospinal fluid amines and higherlethality suicide attempts in depressed inpatients. Biol. Psychiatry 41, 162–171. Meltzer, H.Y., 1989. Serotonergic dysfunction in depression. Br. J. Psychiatry 155 (S8), 25–31. Ninan, P.T., van Kammen, D.P., Scheinin, M., Linnoila, M., Bunney, W.E., Goodwin, F.K., 1984. CSF 5-hydroxyindoleacetic acid in suicidal schizophrenic patients. Am. J. Psychiatry 141, 566–569. ¨ V., Swedin, A., Zachau, A.C., 1995. Nordin, C., Eklund, T., Fernstrom, Gradients of CSF monoamine metabolites: a comparison between male and female volonteers. J. Psychiatr. Res. 29, 133–140. ˚ ˚ ¨ P., Samuelsson, M., Asberg, ¨ Nordstrom, M., Traskman-Bendz, L., AbergWistedt, A., Nordin, C., Bertilsson, L., 1994. CSF 5-HIAA predicts suicide risk after attempted suicide. Suicide Life-Threat. Behav. 24, 1–9. Potter, W.Z., Manji, H.K., 1993. Are monoamine metabolites worth measuring? Arch. Gen. Psychiatry 50, 653–656. Reddy, P.L., Khanna, S., Subhash, M.N., Channabasavanna, S.M., Rao, B.S., 1992. CSF amine metabolites in depression. Biol. Psychiatry 31, 112–118. ˚ Roy, A., Agren, H., Pickar, D., Linnoila, M., Doran, A.R., Cutler, N.R., Paul, S.M., 1986. Reduced CSF concentrations of HVA and HVA to 5-HIAA ratios in depressed patients: relationships to suicidal behavior and dexamethasone nonsuppression. Am. J. Psychiatry 143, 1539– 1545. Roy, A., Lamparski, D., De Jong, J., Adinoff, B., Ravitz, B., George, D.T., Nutt, D., Linnoila, M., 1990. CSF monoamine metabolites in alcoholic patients who attempt suicide. Acta Psychiatr. Scand. 81, 58–61.
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom Roy-Byrne, P., Post, R.M., Rubinow, D.R., Linnoila, M., Savard, R., Davis, D., 1983. CSF 5HIAA and personal and family history of suicide in affectively ill patients: a negative study. Psychiatry Res. 10, 263–274. ¨ Stanley, M., Traskman-Bendz, L., Dorovini-Zis, K., 1985. Correlations between aminergic metabolites simultaneously obtained from human CSF and brain. Life Sci. 37, 1279–1286.
405
˚ ¨ ¨ Traskman, L., Asberg, M., Bertilsson, L., Sjostrand, L., 1981. Monoamine metabolites in CSF and suicidal behavior. Arch. Gen. Psychiatry 38, 631–636. Westenberg, H.G.M., Verhoeven, W.M.A., 1988. CSF monoamine metabolites in patients and controls: support for a bimodal distribution in major affective disorders. Acta Psychiatr. Scand. 78, 541–549.
Reduced cerebrospinal HVA concentrations and HVA / 5-HIAA ratios in suicide attempters Monoamine metabolites in 120 suicide attempters and 47 controls a ¨ a , *, Christer Alling b , Kaj Blennow c , Goran ¨ ´ a , Lil Traskman-Bendz ¨ Gunnar Engstrom Regnell a
Department of Clinical Neuroscience, Psychiatry Section, University Hospital, S-221 85 Lund, Sweden b Department of Medical Neurochemistry, Lund University Hospital, S-22185 Lund, Sweden c ¨ ¨ Hospital, S-431 80 Molndal , Institute of Clinical Neuroscience, Department of Psychiatry and Neurochemistry, Gothenburg University, Molndal Sweden Received 21 January 1999; received in revised form 22 March 1999; accepted 24 March 1999
Abstract Dysfunctions of central monoaminergic systems are important elements of the leading biological hypotheses of suicide and depression. The purpose of the present paper was to study the levels and the relationships between the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA), the dopamine metabolite homovanillic acid (HVA) and the norepinephrine metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) in the cerebrospinal fluid (CSF) in 120 hospitalised suicide attempters and 47 controls (healthy volunteers or patients admitted for minor surgery). The suicide attempters showed significantly lower HVA levels (174682 vs. 216696 nmol / L, P50.004), HVA / 5HIAA ratios (1.660.5 vs. 2.160.6, P50.0001) and HVA / MHPG ratios (4.262.1 vs. 4.861.7, P50.02) than the controls. The correlations between the monoamine metabolites were markedly lower in patients than in controls. CSF 5-HIAA showed no significant differences between patients and controls (107640 vs. 108651 nmol / L) or between violent and non-violent attempters (112658 vs. 105633 nmol / L). The monoamine metabolites showed no significant differences between survivors and patients who subsequently completed suicide, or between suicide attempters subgrouped by psychiatric diagnoses. The results suggest that low HVA levels and altered relationships between the monoamine metabolites are associated with suicidal behaviour. 1999 Elsevier Science B.V. All rights reserved. Keywords: Monoamine metabolites; Serotonin; Dopamine; Suicide; Depression
1. Introduction During the last decades, dysfunctions of central monoaminergic systems have been important elements of the leading biological hypotheses of suicide as well as depression, anxiety disorders and schizophrenia (e.g., Carlsson, 1978; Coccaro et al., 1989; Mann et al., 1992; Meltzer, 1989; Kahn et al., 1988, Potter and Manji, 1993; Westenberg and Verhoeven, 1988). Based on the assumption that *Corresponding author. Present address: Department of Community ¨ Sweden. Tel.: Medicine, Malmo¨ University Hospital, S-20502 Malmo, 146-403-3-3435; fax: 146-403-3-6215. ¨ E-mail address: [email protected] (G. Engstrom)
the concentrations of monoamine metabolites in the cerebrospinal fluid (CSF) reflect the monoaminergic activities in the brain (Stanley et al., 1985), many investigators have studied the levels of the monoamine metabolites in CSF. Suicide and suicidal behaviour are phenomenas which occur in patients with different psychiatric diagnoses (Harris and Barraclough, 1997). Ahrens and Linden (1996) even suggested that this behaviour could be regarded as a separate syndrome. Biological studies in this field have repeatedly associated low levels of the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA) in CSF with suicidal behaviour, especially with violent methods ˚ ¨ (Asberg et al., 1976; Traskman et al., 1981). Relationships between suicidality and low CSF 5-HIAA have been
0924-977X / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0924-977X( 99 )00016-4
400
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
reported in different diagnostic categories, e.g. in patients with schizophrenia, depression, alcoholism, adjustment disorders, personality disorders and in violent offenders (Linnoila et al., 1983; Banki et al., 1984; Gardner et al., 1990; Ninan et al., 1984; Lopez-Ibor et al., 1985). There are also negative findings however (Lemus et al., 1990; Roy et al., 1990; Westenberg and Verhoeven, 1988), especially in bipolar suicidal patients (Berretini et al., 1986; Roy-Byrne et al., 1983). Nevertheless, a metaanalytic review of the relationships between CSF monoamine metabolites and suicidal behaviour found strong evidence for lower CSF 5-HIAA in suicide attempters than in psychiatric controls (Lester, 1995). There are also some studies linking the dopamine metabolite homovanillic acid (HVA) to suicidal behaviour (Roy et al., 1986; Jones et al., 1990) and low CSF HVA levels is the most consistent finding when comparing CSF monoamine metabolites in depressive subjects and controls (Goodwin and Jamison, 1990). The findings of the norepinephrine metabolite 3methoxy-4-hydroxyphenylglycol (MHPG) in CSF are inconsistent in studies of suicide attempters as well as major depression (Lester, 1995; Goodwin and Jamison, 1990). The theories of the casual relationships between monoamines and psychiatric illness have gradually turned in favour of hypotheses involving interactions of several neurotransmitters rather than deficiency of one specific monoamine (Carlsson et al., 1997; Potter and Manji, 1993). Smaller correlations, which possibly reflect altered interactions, between the CSF monoamine metabolites in drug-free schizophrenics than in controls, and increasing correlations after treatment, are reported by Hsiao et al. (1993). In treatment-resistant depressives the correlations were significantly smaller than in responders (Hsiao et al., 1987). Other possible measures of the interactions between the monoamines are the ratios between the CSF monoamine metabolites (Roy et al., 1990; Hsiao et al., 1987, 1993; Reddy et al., 1992). The aim of the present study has been to investigate the CSF concentrations of 5-HIAA, HVA and MHPG and to study the relationships between the metabolites in a large sample of suicide attempters with regard to psychiatric diagnoses and suicidal behaviour, and in comparison to a control group.
2. Experimental procedures
37.4611.5, range: 19–61 years) and 65 were women (mean age 38.8614.5, range 19–72 years). The mean body height was 172.368.9 cm. The patients were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R; American Psychiatric Association, 1987) by two independent psychiatrists. The patients were categorised according to suicide method into violent and non-violent attempters. Suicide attempters who took drug overdoses or made single wrist cuts were denoted ‘‘non-violent’’, and those who used other methods or combinations of two or more methods ¨ were denoted ‘‘violent’’ (Traskman et al., 1981). The patients were also categorised into ‘‘repeaters’’ who had made previous suicide attempts, and ‘‘non-repeaters’’ without previous suicide attempts. Information on subsequently completed suicides was retrieved from the unit for Forensic Medicine in Lund and from the National Cause of Death Register.
2.2. Controls The control group consisted of 47 volunteers from Gothenburg, Sweden, (30 men, 17 women) with a mean age of 44.1615.9 years (range 18–70 years) (Blennow et al., 1993a). Eighteen subjects (14 men, 4 women, mean age 61.265.4 years) were patients who had been admitted for minor surgery (e.g. coxarthrosis and prostate hyperplasia) under spinal anaesthesia. They were examined the day before the surgery. The CSF sample was drawn in connection with the lumbar puncture for the spinal anaesthesia. Twenty-nine subjects (16 men, 13 women, mean age: 33.569.2 years) were healthy volunteers that accepted a medical and psychiatric examination, this group was primarily recruited among ward staff. The mean body height of the controls was 174.668.6 cm. A careful medical history was taken from each subject and they all went through a physical and neurological clinical examination. A structured interview including specific questions on different psychiatric symptoms was undertaken. Subjects with history, symptoms or signs of psychiatric, neurological disorders, Type I diabetes, severe hypertension, malignant disease, or autoimmune infections were excluded. By definition, no subject received medication with CNS active drugs (Blennow et al., 1993a).
2.3. Lumbar punctures and monoamine metabolite assays
2.1. Patients The patients (n5120) were evaluated by a psychiatrist and a social worker with a structured psychiatric interview after a suicide attempt and then hospitalised at the Lund University Hospital in a psychiatric ward specialised in suicidal behaviour. They all gave informed consent to participate in the study. Mean age was 38.2613.2 years (range: 19–72 years). Fifty-five were men (mean age
The lumbar punctures were taken in the morning after at least 8 h. of fasting and bed rest since midnight. Neither patients nor controls had any specific diet prior to the examination. Samples were taken with the subjects in a sitting position. With exception of oral contraceptives and occasional benzodiazepines the patients were drug-free prior to the lumbar puncture. The mean drug free wash-out period was 14.867.4 days (range 4–57 days). Twelve ml
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
of spinal fluid was taken and gently mixed to avoid gradients, and then stored in 2808C in 2 ml aliquots. The CSF monoamine metabolite concentrations in the suicide attempters were analysed with gas chromatography–mass spectrometry (GC–MS) (Gjerris et al., 1987). Both GC–MS and high performance liquid chromatography (HPLC) were used for analyses of the monoamine metabolite concentrations in the controls (described in Blennow et al., 1993a, 1993b). The correlations between the values obtained by GC–MS and HPLC were above 99% (r.0.99) for all of HVA, 5-HIAA and MHPG (Blennow et al., 1993b). The inter- and intra-assay coefficients were 2–4% for all metabolites with both methods.
401
Table 2 Spearman correlation coefficients between monoamine metabolites, age and height in patients (upper right half, bold types), and controls (lower left half) HVA HVA 5-HIAA MHPG Age Height
5HIAA
MHPG
0.64*** 0.81*** 0.43** 0.35* 20.39*
0.47*** 0.52*** 20.51***
0.12 0.21* 0.31* 20.34*
Age
Height
20.01 0.08 0.06
20.30*** 20.18 20.13 20.07
20.22
***P,0.001, **P,0.01, *P,0.05.
ences between controls admitted for minor surgery and other healthy controls.
3.1. Age, height, and gender
2.4. Statistics Correlations between variables were calculated by means of Spearman correlation coefficients. Mann–Whitney U-test or Kruskal–Wallis non-parametric analysis of variance was used for comparisons of groups. As the monoamine metabolites were correlating with age and body height, we also performed analyses of covariance (ANCOVAs) with age and height as covariates in all comparisons of two or more groups.
In both patients and controls, all three monoamine metabolites were negatively correlated with height (Table 2). All three monoamine metabolites correlated significantly with age in the controls, while no significant association between the monoamine metabolites and age was found among the patients (Table 2). The monoamine metabolites showed no significant differences between men and women in patients or controls (data not shown).
3.2. Seasonal variation 3. Results The patients showed significantly lower levels of HVA (median: 172 vs. 204 nmol / L, Z52.87, P50.004) and lower HVA / 5-HIAA (median: 1.75 vs. 2.12, Z54.7, P, 0.0001) and HVA / MHPG (median: 4.35 vs. 4.98, Z52.4, P50.02) ratios than the controls (Table 1). Although not significantly, the patients also had lower MHPG (median 42 vs. 43 nmol / L, Z51.7, P50.08). After adjustments for age and body height, the results were essentially the same with significant differences between suicide attempters and controls in HVA (F514.5, P,0.001), HVA / 5-HIAA ratio (F543.0, P,0.0001) and HVA / MHPG ratio (F56.4, P5 0.013). When age and height had been taken into account, the monoamine metabolites showed no significant differ-
The suicide attempters were categorised into four groups according to time of the year: Fifty-three subjects were lumbar punctured in December–March, 22 in April or May, 10 in June–August, and finally 35 patients were lumbar punctured in September–November. The monoamine metabolites showed no significant differences between the groups (data not shown). After adjustments for age and height, there were similarly no seasonal variations among the controls.
3.3. Suicidality Eight patients (6 women, 2 men, mean age at examination: 40.9617 years) subsequently committed suicide. In this group 3 had major depression, 2 had dysthymia, 2 had
Table 1 Mean6SD of HVA, 5-HIAA, MHPG and HVA / 5HIAA and HVA / MHPG ratios in different categories of suicide attempters and controls HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
Patients (n5120) Controls (n547)
173.6682** 216.4696
106.7640 107.7651
42.469 45.169
1.660.5*** 2.160.6
4.262.1* 4.861.7
Violent (n526) Non-violent (n594)
188.66104 169.5675
112.1658 105.2633
42.067 42.5610
1.760.5 1.660.5
4.662.9 4.161.8
Repeaters (n558) Non-repeaters (n562) Suicides (n58)
167.0692 179.9693 196.56106
106.2634 107.2645 111.4646.1
41.6610 43.169 43.9611
1.660.5 1.760.5 1.760.5
4.161.8 4.362.4 4.562.3
***P,0.0001, **P,0.01, *P,0.05. Patients compared with controls.
402
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
Table 3 Monoamine metabolite concentrations (6SD) (nmol / L) in suicide attempters who took overdoses of neuroleptics or antidepressives, overdoses of other substances or who used other methods than intoxications
Neuroleptics or antidepressives (n524) Other substances (n565) No intoxication (n531)
HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
140661 181679 184699
104632 106634 110655
42610 43610 4169
1.460.5* 1.760.5* 1.760.5*
3.461.5 4.361.9 4.662.7
*P50.02 (Kruskal–Wallis non-parametric ANOVA).
adjustment disorders and one had depression not otherwise specified (NOS) at the baseline examination. No significant monoamine metabolite differences was observed between subsequently completed suicides and other suicide attempters (Table 1). The monoamine metabolites showed no significant differences between violent and non-violent attempters or between repeaters and non-repeaters (Table 1). Twenty-four patients had taken overdoses with neuroleptics (n54) or antidepressive drugs (n520), which may influence the concentrations of monoamine metabolites. The wash-out period in this group was 16.566 days. This group showed significantly lower HVA / 5HIAA ratio (p5 0.02) than patients who took overdoses of other drugs or who used other methods. The HVA levels were also lower among these patients, although the differences did not reach significance (Table 3). The 5-HIAA and MHPG concentrations showed only small differences.
3.4. Psychiatric diagnoses Thirty-six patients were diagnosed with major depression (MDD), 19 with dysthymia, 23 with adjustment disorders, 15 with depression not otherwise specified (NOS), 5 with alcohol dependence, 5 with alcohol abuse, two with schizophrenia, disorganised, two with schizoaffective disorder, two with panic disorder without agoraphobia, one each with schizophrenia, undifferentiated, psychotic disorder NOS, anxiety disorder NOS, panic disorder with agoraphobia, obsessive compulsive disorder, cyclothymia, polysubstance dependence, anorexia nervosa, bulimia nervosa, and finally, two patients did not receive any axis I diagnosis (both had borderline personality disorder). The patients were categorised according to their axis I diagnosis into MDD (n536), dysthymia (n5
19), adjustment disorders (n523), and finally into a group with other diagnoses or without any axis I diagnosis (‘‘other diagnoses’’ n542). Twenty-eight patients received two or three axis I diagnoses, eighteen of them had an alcohol- or substance abuse disorder. The monoamine metabolite levels did not differ significantly between the diagnostic groups (Table 4). The differences remained non-significant after adjustments for age and height.
3.5. Relationships between monoamine metabolites The HVA / 5-HIAA and HVA / MHPG ratios were significantly different between patients and controls (Table 1). The correlation coefficients between the monoamine metabolites were markedly lower in the patients than controls (rsp for HVA / 5-HIAA: 0.64 vs. 0.81; for HVA / MHPG: 0.12 vs. 0.43; for 5-HIAA / MHPG: 0.21 vs. 0.47) (Table 2).
4. Discussion In the present study, the levels of CSF HVA and the HVA / 5-HIAA and HVA / MHPG ratios were significantly lower among the patients than the controls. The levels of CSF 5-HIAA showed no significant differences between suicide attempters and controls. Previous studies of cerebrospinal monoamine metabolites in suicide attempters have associated low CSF 5-HIAA with suicidality, the findings of HVA in previous studies are however inconsistent (Lester, 1995). As suicide attempters are clinically very heterogeneous, it is possible that the sample of the
Table 4 Mean6SD of HVA, 5-HIAA, MHPG and HVA / 5HIAA and HVA / MHPG ratios in suicide attempters with different psychiatric diagnoses
Major depression (n536) Dysthymia (n519) Adjustment disorders (n523) Other diagnoses (n542)
HVA
5-HIAA
MHPG
HVA / 5HIAA
HVA / MHPG
168.26101 191.7681 166.3649 174.1682
105.8654 114.6642 108.8625 102.7631
41.2610 42.868 42.966 42.9611
1.660.5 1.760.4 1.660.4 1.760.6
4.262.7 4.561.7 4.061.4 4.262.0
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
present study could explain the differences between our results and those by others. A majority of the patients in the present study had mood disorders or mood-related disorders. This could be the major explanation for the low HVA levels and HVA / 5-HIAA ratios rather than the suicidality. Low HVA levels is the most consistent finding in studies of CSF monoamine metabolites in depressed patients (Goodwin and Jamison, 1990), and previous studies which have associated low HVA levels with suicidal behaviour have studied samples of depressed suicide attempters (Roy et al., 1986; Jones et al., 1990; ¨ Traskman et al., 1981). Thus, low HVA levels and HVA / 5HIAA ratios might be markers of depression rather than suicidality. We could not find any differences in CSF 5-HIAA levels between suicide attempters and controls or between violent or nonviolent attempters. It has been suggested that individuals who volunteer in lumbar punctures are more impulsive than other healthy controls (Gustavsson et al., 1997). If that is the case, the controls would have lower CSF 5-HIAA than expected and the differences between patients and controls would be reduced. However, since the 5-HIAA levels of the controls in the present study are at the same level as in control groups from other studies ˚ ¨ (e.g., Traskman et al., 1981; Asberg et al., 1984) this explanation seems to be unlikely. We could not replicate the associations between low CSF 5-HIAA and subsequently completed suicide reported from previous follow-up studies of suicide attempters (Lester, 1995; ¨ et al., 1994). Relationships between low CSF Nordstrom 5-HIAA and high lethality of suicidal behaviour has also been reported (Mann and Malone, 1997). Most patients in the present study were non-violent attempters and few attempts were very serious. It is possible that our way of categorising the attempts into ‘‘violent’’ or ‘‘non-violent’’ didn’t reflect the violence or intention of the attempts well enough or that the variance in violence or lethality was too small. However, these explanations are challenged by the fact that the patients who later died from suicide also had CSF 5-HIAA at the same level as the non-violent attempters. Several investigators have discussed the influence of various confounders, such as age, height, gender or seasonal variation, and the importance of adjusting the monoamine metabolite levels for these factors (e.g. Blen˚ now et al., 1993a,b; Asberg et al., 1984; Nordin et al., 1995; Brewerton et al., 1988). When all subject were taken together, the monoamine metabolites were associated with age (positively) and height (negatively) and no significant difference between men and women was found. The correlations between age, height and the monoamine metabolites differed between patients and controls (Table 2), and the relationships between the monoamine metabolites and age, height and sex were weak among the suicide attempters. The results were essentially the same after statistical adjustments for age and height. There were no
403
significant seasonal variation of the monoamine metabolites in patients or controls. A relevant issue is however whether the results could have been confounded by the fact that patients and controls were from different cities. Whether there are regional variations of the monoamine metabolite levels in the south of Sweden is unknown and this remains a possible confounder. It is noteworthy however that the most significant difference between patients and controls was the HVA / 5-HIAA ratio, a measure which is relatively independent of different confounding factors, and a measure which in contrast to the different monoamine metabolite levels is independent of accuracy in monoamine measurements. The relationships between the different monoamine metabolites, rather than the metabolites themselves, have received increasing interest over the last years. Altered monoaminergic interactions have previously been associated with low response to antidepressive treatment (Hsiao et al., 1987) and with psychiatric symptoms in schizophrenics (Kahn et al., 1993). It is therefore noteworthy that in the present study, the HVA / 5-HIAA ratio showed bigger differences between patients and controls than the HVA levels alone. The subjects who had taken overdoses of neuroleptics or antidepressive drugs had lower HVA concentrations than other suicide attempters. However, as suicide attempters who used other methods than intoxications also had significantly lower HVA / 5HIAA ratio than controls, the intoxications could not explain the difference between patients and controls. The HVA / 5HIAA ratio was significantly different between patients and controls also when age and height were covaried out. It is also noteworthy that the controls showed considerably higher correlations between the monoamine metabolites than the patients. Both the ratios and correlations between CSF monoamine metabolites are measures that possibly reflect the interactions between the monoamines in CNS. These constructs are however not validated and so far it is necessary to interpret these findings with care. Smaller correlations between the metabolites in patients than in controls could reflect independence, or ‘‘impaired cooperation’’ between the monoamine systems, but also a biological heterogeneity either among patients or controls. Nevertheless, further study of the interactions between the monoamine systems might be important in order to improve our understanding of psychiatric disorders and suicidal behaviour. In summary, the results of the present study suggest that HVA and the relationships between HVA and 5-HIAA might be interesting markers in the study of suicidal behaviour and mood disorders. The markedly lower correlations coefficients between the monoamine metabolites in patients than controls suggest that the interactions between the monoamine systems differ between patients and controls. Our negative findings regarding the CSF 5-HIAA levels suggest that serotonergic dysfunctions are not present in all groups of suicide attempters. This puts
404
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom
forward the need of studies comparing well-defined groups of suicidal patients.
Acknowledgements ¨ This study was supported by the Malmohus County, the Swedish Medical Research Council Nos. 08319, 714185, ¨ OM Perssons foundation, the Sjobring Foundation and funds from the University of Lund.
References Ahrens, B., Linden, M., 1996. Is there a suicidality syndrome independent of specific major psychiatric disorders? Results from a split half multiple regression analysis. Acta Psychiatr. Scand. 94, 79–86. American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd Ed. (Revised). APA, Washington, DC. ˚ ¨ ´ P., 1976. 5-HIAA in the cerebrospinal Asberg, M., Traskman, L., Thoren, fluid. A biochemical suicide predictor? Arch. Gen. Psychiatry 33, 1193–1197. ˚ ˚ ´ Asberg, M., Bertilsson, L., Martensson, B., Scalia-Tomba, G.-P., Thoren, ¨ P., Traskman-Bendz, L., 1984. CSF monoamine metabolites in melancholia. Acta Psychiatr. Scand. 69, 201–219. ´ M., Papp, Z., Kurcz, M., 1984. Biochemical markers Banki, C.M., Arato, in suicidal patients. Investigations with cerebrospinal fluid monoamine metabolites and neuroendocrine tests. J. Affect. Dis. 6, 341–350. Berretini, W.H., Nurnberger, J.I., Narrow, W., Simmons-Alling, S., Gershorn, E.S., 1986. Cerebrospinal fluid studies of bipolar patients with and without a history of suicide attempts. Ann. NY Acad. Sci. 487, 197–201. ˚ Blennow, K., Wallin, A., Gottfries, C.G., Karlsson, I., Mansson, J.E., ¨ C., Svennerholm, L., 1993a. Cerebrospinal fluid Skoog, I., Wikkelso, monoamine metabolites in 114 healthy individuals 18–81 years of age. Eur. Neuropsychopharmacol. 3, 55–61. ˚ Blennow, K., Wallin, A., Gottfries, C.G., Mansson, J.E., Svennerholm, L., 1993b. Concentration gradients for monoamine metabolites in lumbar cerebrospinal fluid. J. Neural. Transm. (P-D Sect) 5, 5–15. Brewerton, T.D., Berrettini, W.H., Nurnberger, J.I., Linnoila, M., 1988. Analysis of seasonal fluctuations of CSF monoamine metabolites in normal controls: findings with 5HIAA and HVA. Psychiatry Res. 23, 257–265. Carlsson, A., 1978. Antipsychotic drugs, neurotransmitters, and schizophrenia. Am. J. Psychiatry 135, 164–173. Carlsson, A., Hansson, L.O., Waters, N., Carlsson, M., 1997. Neurotransmitter aberrations in schizophrenia: new perspectives and therapeutic implications. Life Sci. 61, 75–94. Coccaro, E.F., Siever, L.J., Klar, H.M., Maurer, G., Cochrane, K., Cooper, T.B., Mohns, R.C., Davis, K.L., 1989. Serotonergic studies in patients with affective and personality disorders. Arch. Gen. Psychiatry 46, 587–599. Gardner, D.L., Lucas, P.B., Cowdry, R.W., 1990. CSF metabolites in borderline personality disorder compared with normal controls. Biol. Psychiatry 28, 247–254. ¨ Gjerris, A., Werdelin, L., Gjerris, F., Sorensen, P.S., Rafaelsen, O.J., Alling, C., 1987. CSF-amine metabolites in depression, dementia and in controls. Acta Psychiatr. Scand. 75, 619–628. Goodwin, F.K., Jamison, K.R., 1990. In: Manic-depressive Illness, Oxford University Press, New York and Oxford, pp. 416–502. ˚ Gustavsson, J.P., Asberg, M., Schalling, D., 1997. The healthy control subject in psychiatric resaerch: Impulsiveness and volunteer bias. Acta Psychiatr. Scand. 96, 325–328.
Harris, E.C., Barraclough, B., 1997. Suicide as an outcome for mental disorders. A meta-analysis. Br. J. Psychiatry 170, 205–228. ˚ Hsiao, J.K., Agren, H., Bartko, J.J., Rudorfer, M.V., Linnoila, M., Potter, W.Z., 1987. Monoamine neurotransmittor interactions and the prediction of antidepressant response. Arch. Gen. Psychiatry 44, 1078–1083. Hsiao, J.K., Colison, J., Bartko, J.J., Doran, A.R., Konicki, P.E., Potter, W.Z., Pickar, D., 1993. Monoamine neurotransmittor interactions in drug-free and neuroleptic-treated schizophrenics. Arch. Gen. Psychiatry 50, 606–614. ¨ Jones, J.S., Stanley, B., Mann, J.J., Frances, A.J., Guido, J.R., TraskmanBendz, L., Winchel, R., Brown, R.P., Stanley, M., 1990. CSF 5-HIAA and HVA concentrations in elderly depressed patients who attempted suicide. Am. J. Psychiatry 147, 1225–1227. Kahn, R.S., Davidson, M., Knott, P., Stern, R.G., Apter, S., Davis, K.L., 1993. Effect of neuroleptic medication on cerbrospinal fluid metabolite concentrations in schizophrenia. Arch. Gen. Psychiatry 50, 599–605. Kahn, R.S., van Praag, H.M., Wetzler, S., Asnis, G., Barr, G., 1988. Serotonin and anxiety revisited. Biol. Psychiatry 23, 189–208. Lemus, C.Z., Lieberman, J.A., Johns, C.A., Pollack, S., Bookstein, P., Cooper, T.B., 1990. CSF 5-hydroxyindoleacetic acid levels and suicide attempts in schizophrenia. Biol. Psychiatry 27, 926–929. Lester, D., 1995. The concentration of neurotransmitter metabolites in the cerebrospinal fluid of suicidal individuals: a meta-analysis. Pharmacopsychiatry 28, 45–50. Linnoila, M., Virkkunen, M., Scheinin, M., Nuutila, A., Rimon, R., Goodwin, F., 1983. Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sci. 33, 2609–2614. Lopez-Ibor, J.J., Saiz-Ruiz, J., Perez de los Cobos, J.C., 1985. Biological correlations of suicide and aggressivity in major depressions (with melancholia): 5-HIAA and cortisol in CSF, dexamethasone suppression test and therapeutic response to 5-hydroxytryptophan. Neuropsychobiology 14, 67–74. Mann, J.J., McBride, P.A., Brown, R.P., Linnoila, M., Leon, A.C., DeMeo, M., Mieczkowski, T., Myers, J.E., Stanley, M., 1992. Relationships between central and peripheral serotonin indexes in depressed and suicidal psychiatric inpatients. Arch. Gen. Psychiatry 49, 442–446. Mann, J.J., Malone, K.M., 1997. Cerebrospinal fluid amines and higherlethality suicide attempts in depressed inpatients. Biol. Psychiatry 41, 162–171. Meltzer, H.Y., 1989. Serotonergic dysfunction in depression. Br. J. Psychiatry 155 (S8), 25–31. Ninan, P.T., van Kammen, D.P., Scheinin, M., Linnoila, M., Bunney, W.E., Goodwin, F.K., 1984. CSF 5-hydroxyindoleacetic acid in suicidal schizophrenic patients. Am. J. Psychiatry 141, 566–569. ¨ V., Swedin, A., Zachau, A.C., 1995. Nordin, C., Eklund, T., Fernstrom, Gradients of CSF monoamine metabolites: a comparison between male and female volonteers. J. Psychiatr. Res. 29, 133–140. ˚ ˚ ¨ P., Samuelsson, M., Asberg, ¨ Nordstrom, M., Traskman-Bendz, L., AbergWistedt, A., Nordin, C., Bertilsson, L., 1994. CSF 5-HIAA predicts suicide risk after attempted suicide. Suicide Life-Threat. Behav. 24, 1–9. Potter, W.Z., Manji, H.K., 1993. Are monoamine metabolites worth measuring? Arch. Gen. Psychiatry 50, 653–656. Reddy, P.L., Khanna, S., Subhash, M.N., Channabasavanna, S.M., Rao, B.S., 1992. CSF amine metabolites in depression. Biol. Psychiatry 31, 112–118. ˚ Roy, A., Agren, H., Pickar, D., Linnoila, M., Doran, A.R., Cutler, N.R., Paul, S.M., 1986. Reduced CSF concentrations of HVA and HVA to 5-HIAA ratios in depressed patients: relationships to suicidal behavior and dexamethasone nonsuppression. Am. J. Psychiatry 143, 1539– 1545. Roy, A., Lamparski, D., De Jong, J., Adinoff, B., Ravitz, B., George, D.T., Nutt, D., Linnoila, M., 1990. CSF monoamine metabolites in alcoholic patients who attempt suicide. Acta Psychiatr. Scand. 81, 58–61.
¨ et al. / European Neuropsychopharmacology 9 (1999) 399 – 405 G. Engstrom Roy-Byrne, P., Post, R.M., Rubinow, D.R., Linnoila, M., Savard, R., Davis, D., 1983. CSF 5HIAA and personal and family history of suicide in affectively ill patients: a negative study. Psychiatry Res. 10, 263–274. ¨ Stanley, M., Traskman-Bendz, L., Dorovini-Zis, K., 1985. Correlations between aminergic metabolites simultaneously obtained from human CSF and brain. Life Sci. 37, 1279–1286.
405
˚ ¨ ¨ Traskman, L., Asberg, M., Bertilsson, L., Sjostrand, L., 1981. Monoamine metabolites in CSF and suicidal behavior. Arch. Gen. Psychiatry 38, 631–636. Westenberg, H.G.M., Verhoeven, W.M.A., 1988. CSF monoamine metabolites in patients and controls: support for a bimodal distribution in major affective disorders. Acta Psychiatr. Scand. 78, 541–549.