Low HDL cholesterol is associated with suicide attempt among young healthy women: the Third National Health and Nutrition Examination Survey

Journal of Affective Disorders 89 (2005) 25 – 33 www.elsevier.com/locate/jad

Research report

Low HDL cholesterol is associated with suicide attempt among young healthy women: the Third National Health and Nutrition Examination Survey Jian Zhang a,b,*, Robert E. McKeown b, James R. Hussey b, Shirley J. Thompson b, John R. Woods c,d, Barbara E. Ainsworth b,e a Division of Health and Family Studies, Institute for Families in Society, University of South Carolina, United States Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, United States c Center for Health Services and Policy Research, Arnold School of Public Health, University of South Carolina, United States d Department of Family and Preventive Medicine, School of Medicine, University of South Carolina, Columbia, SC, United States e Department of Exercise and Nutritional Sciences, San Diego State University, United States b

Received 15 December 2004; accepted 11 May 2005 Available online 2 November 2005

Abstract Background: Serum cholesterol is reported to be associated with suicidality, but studies conducted among general healthy population are rare. We examined the association between serum cholesterol and suicidality in a national sample of the general population of US. Methods: We used the data of 3237 adults aged 17 to 39 years, who completed a mental disorder diagnostic interview and had blood specimens collected after a 12-h fast, as a part of the Third National Health and Nutrition Examination Survey, 1988– 1994. The serum concentrations of total cholesterol, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were dichotomized according to the recommended levels of the National Cholesterol Education Program. A polytomous logistic regression was employed to control for covariates. Results: Independent of socio-demographic variables, health risks and nutrition status, and a history of medical and psychiatric illness (including depression), a significant association between low HDL-C (V 40 mg/dl) and increased prevalence of suicide attempts was observed in women (OR = 2.93, 95% CI = 1.07–8.00). No significant evidence was found to support an association between cholesterol and suicide ideation in women. Serum cholesterol was unrelated with either suicide ideation or attempts in men. Limitation: The inherent limitation of cross-sectional design prevented the authors from investigating causality.

* Corresponding author. 4770 Buford Hwy, Mailstop K24, Atlanta GA 30341, United States. Tel.: +1 770 488 5433; fax: +1 770 488 6500. E-mail address: [email protected] (J. Zhang). 0165-0327/$ - see front matter D 2005 Published by Elsevier B.V. doi:10.1016/j.jad.2005.05.021

26

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

Conclusions: Low HDL-C is significantly associated with suicide attempts in women. Further studies are necessary to explore the clinical application of serum cholesterol as an indicator for suicide attempts among high risk population. D 2005 Published by Elsevier B.V. Keywords: Serum cholesterol; Suicide ideation; Suicide attempt; Young and middle age; NHANES III

1. Introduction Due to its potential importance, the relation between suicide and serum cholesterol has been the subject of much debate since the publication of several reports (Muldoon et al., 1990; Jacobs et al., 1992) in the early 1990s suggesting the possible relevance of serum cholesterol to suicides. Increased incidence or prevalence of completed and attempted suicides among people with low serum cholesterol was observed from cross-sectional (Ozer et al., 2004; Repo-Tiihonen et al., 2002; Tripodianakis et al., 2002; Atmaca et al., 2003), retrospective case–control (Atmaca et al., 2002a,b, 2003; Garland et al., 2000; Kim et al., 2002; Kim and Myint, 2004; Lee and Kim, 2003; Lindberg et al., 1992; Vevera et al., 2003), cohort studies (Partonen et al., 1999; Zureik et al., 1996; Neaton et al., 1992; Lindberg et al., 1992; Ellison and Morrison, 2001) as well as a genetic study (Lalovic et al., 2004). However, other studies were not able to detect the association (Tanskanen et al., 2000a,b; Knox et al., 1996; Deisenhammer et al., 2004). The inconsistent findings may be related to study designs, in that most of the previous studies used convenient samples drawn from psychiatric patients, among which serum samples were generally collected on admission or within 48 h following admission. Measurement on admission might confound the results, as psychotropic medications given to psychiatric patients, such as selective serotonin uptake inhibitors, medications taken in the suicide attempts or the medical treatment itself, might affect cholesterol profiles. The risks of suicide behaviors are drawn from multiple domains and serum cholesterol levels are sensitive to age, gender, nutritional status, as well as many factors still unknown. The relatively small sample sizes precluded previous studies from controlling for covariates comprehensively. These limitations may jeopardize the validity of the studies, as well as the generalizability to the broader population. To address these limitations, we analyzed the

data from a national survey of the general population with major confounding factors controlled for. The aim of the current study is to test the hypothesis that serum cholesterol was associated with suicide ideation and suicide attempts among relatively healthy young adults from a community survey.

2. Material and methods 2.1. Study population The Third National Health and Nutrition Examination Survey (NHANES III) was a cross-sectional survey of the US non-institutionalized civilian population conducted from 1988 to 1994. Detailed descriptions of the survey have been published elsewhere (National Center for Health Statistics, 1996). A total of 7968 adults aged 17–39 years participated in the survey, and 3559 were randomly assigned to a morning physical examination group and had a 12-h fast before blood was collected. Two participants were excluded due to missing information on mental health: 134 participants identifying themselves as American Indians, Asians or Pacific Islanders were also excluded due to relative small sample size. Additional 186 participants were excluded because the data for serum total cholesterol (TC) was not available. The data of the remaining 3237 participants were used for the current study. 2.2. Variable definitions and measurements 2.2.1. Suicide ideation and attempts As a part of the Diagnostic Interview Schedule, participants were asked several questions specifically related to suicide, including bHave you ever felt so low that you thought of committing suicide (suicide ideation)?Q and bHave you ever attempted suicide (suicide attempt)?Q The time frame for these questions was lifetime. We generated a new variable from these

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

two questions to classify participants into three unoverlapped groups by the degree of suicidality: (1) had attempted suicide, (2) had suicide ideation only, or (3) had neither suicide ideation nor suicide attempts. 2.2.2. Cholesterol measures A single phlebotomy was performed during the medical examination at any time between 8 a.m. and noon after a 12-h fast for each participant. Serum total cholesterol (TC), and high-density-lipoprotein cholesterol (HDL-C) were measured enzymatically, and low-density lipoprotein cholesterol (LDL-C) was calculated using the Friedwald equation. The value of LDL-C was missing when serum triglycerides exceeded 400 mg/dl due to the limitation of Friedwald equation. We categorized cholesterol concentrations into two strata: high and low. The concentrations at or above the desirable (TC), optimal (LDL-C), or low (HDL-C) levels recommended by National Cholesterol Education Program (NCEP, 2001) were categorized as high, otherwise as low. Specifically, cutoffs were TC z 200 mg/dl (32% of weighted male population and 30% of weighted female population), LDLC z 100 mg/dl (66% and 59%, respectively) and HDL-C z 40 mg/dl (70% and 86%, respectively). The same cutoffs were employed for both genders. 2.2.3. Covariates 2.2.3.1. Socio-economic status, social support, health behavior and nutrition indicators. Ethnicity was coded as non-Hispanic White, non-Hispanic Black, or Mexican-American. The poverty index was calculated from the previous year family income and family size. A poverty index less than one was classified as below the poverty line, otherwise as at or above the poverty line. We classified those legally married and those living as married as cohabiting, all others as single. Three variables were generated as social support indicators: social contact, church attending, and club affiliation. Social contact was defined as a total frequency of visiting relatives, friends and neighbors during the entire year before the interview. Current cigarette smokers were defined as those who had smoked cigarettes in the past 5 days. Current alcohol drinkers were defined as those who had consumed at least 12 drinks in the last 12 months. Leisure-time

27

physical activity was measured by the frequency of exercise, sports, or physically active hobbies in the past month. The participants were classified as physically inactive (16% of the weighted population), irregularly active (43%), and regularly active (41%). Selfevaluated health was classified as bexcellentQ, bgoodQ and bpoorQ. Serum vitamins selected were vitamin E, and carotenoids (alpha-carotene, beta-carotene). We also used Body Mass Index (BMI), serum total protein, serum albumin concentration and daily dietary energy intake obtained from a 24-h dietary recall as indicators of malnutrition or energy deprivation. 2.2.3.2. History of medical and psychiatric illnesses. In the NHANES III interview, the participants were asked bHave you ever been told by a doctor that you had one or more of the following general medical illnesses: asthma, arthritis, cancer, chronic bronchitis, diabetes, hypertension, gout, lupus, stroke, or thyroid disease?Q. This list of conditions was drawn from the National Health Interview Survey and was chosen to represent conditions that are both prevalent and associated with substantial morbidity in the US population (Druss and Pincus, 2000). Since most of these illnesses are relatively rare in young adults, and have a very weak association with suicide ideation or attempts in the general population, having a history of cancer (including skin cancer), asthma, or chronic bronchitis were coded dyesT, having other illnesses or no illness as dnoT. The diagnoses of lifetime major depression were generated by a computerized algorithm according to standardized criteria established in the Diagnostic and Statistical Manual of Mental Disorders III (DSM-III), a valid method to obtain diagnoses for community-based psychiatric study. 2.3. Statistical analysis SUDAAN software (Shah et al., 1997) (SAS-callable version, 8.2) with appropriate weighting and nesting variables was employed. We used a polytomous logistic regression model with generalized logit as the link function and a three-level variable of suicidality as the dependent variable. Participants with neither a history of suicide ideation nor suicide attempts were used as the controls. Contrasts were made between the controls and those with suicide

28

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

ideation and between the controls and suicide attempters. Odds ratios (OR) and corresponding 95% confidence intervals (95% CI) were calculated to estimate the associations. Wald tests were used to determine whether the prevalence of lifetime suicide attempts or suicide ideation among those with low cholesterol differed significantly from that among the individuals with high cholesterol (reference). For data reduction, we examined the associations between suicide ideation, attempts and each covariate before multivariable modeling, only those covariates with p values of bivariate relations with suicide ideation or attempts less than 0.10 were kept as significant covariates for the multivariable regression. Multivariable models were built in a stepwise fashion for each cholesterol measure. In the first step, the main effect and the significant covariates were entered into the model. In the second step, the covariates with p values greater than 0.05 were dropped. Decisions about exclusion of contrasts from the model were based on two-sided p values (b 0.05) of the Wald tests alone. It has been well documented that suicidal behaviors differ between genders in terms of neurobiology and sociology (Moscicki, 1994), therefore, separate models were built for each gender. To avoid over-specifying the models, no interaction terms were included.

Table 1 Selected characteristics of weighted study population Characteristica

Age, mean Ethnicity, White Education, below high school Income, below poverty lineb Marital status, single Current smoker, yes Alcohol drinker, yes Serum TC, mean (mg/dl) Serum LDL-C, mean (mg/dl) Serum HDL-C, mean (mg/dl) A history of Cancer/pulmonary disease Life time major depression Life time suicidal ideation Life time suicide attempt

Men (SE)

Women (SE)

(unweighted sample size = 1467)

(unweighted sample size = 1770)

28.80 79.03 21.55 14.06 44.57 39.50 33.44 189.10 119.41 46.33

(0.32) (1.53) (1.33) (1.60) (2.65) (2.02) (2.68) (1.62) (1.53) (0.68)

28.95 78.34 18.23 17.25 41.76 33.82 20.10 185.64 110.62 55.06

(0.25) (1.14) (1.42) (1.69) (2.03) (2.51)* (1.76)* (2.02) (1.76)* (0.60)*

9.14 4.77 13.16 4.13

(1.20) (0.90) (1.63) (0.92)

15.47 11.15 13.97 7.04

(1.31)* (1.30)* (1.65) (0.91)*

Sample of 3237 adults aged 17–39 years, NHANES III, 1988–1994. TC=total cholesterol; HDL=High Density Lipoprotein; LDL=Low Density Lipoprotein; SE=standard error; NHANES III=The Third National Health Examination and Nutrition Survey. a Presented as percentage (standard error) unless otherwise specified. b Poverty index was a calculated variable based on family income and family size. A poverty index less than 1 is usually categorized as below poverty line. * p b 0.05.

3. Results Table 1 presents the selected characteristics of the weighted study population by gender. There were significant differences between men and women for most characteristics selected. Men were more likely to be smokers at the time of interview, and women had a more favorable cholesterol profile (i.e. lower LDL-C and higher HDL-C). More women, however, had a history of psychiatric or medical illness than men did. About 15% of women vs. 9% of men had a history of cancer/pulmonary diseases. The prevalence of a lifetime history of suicide attempts was twice in women as in men, 7.04% vs. 4.13%, respectively. There was no significant difference of the prevalence of lifetime suicide ideation between two genders. Table 2 shows the unadjusted associations between suicide ideation, attempts and the risks selected. For both men and women, the individuals with suicide ideation were more similar to the controls than to the

attempters. For example, among men, 45.17% of the attempters as opposed to 15.91% of the individuals with suicidal ideation and 21.26% of the controls had education below high school, and 66.75% of the attempters were living under the poverty line, while only 11.51% of the individuals with ideation and 11.84% of the controls did so. More than 80% of the attempters, while less than half of the individuals with ideation and 36% of the controls, were smokers. Among female attempters, serum vitamin E, alphacarotene, and beta-carotene were significantly lower than that of the controls, while there was no significant difference between the controls and the individuals with ideation for serum vitamins selected. The prevalence of lifetime major depression, however, was similar between individuals with ideation and attempters, and significantly lower among the controls for both genders. No significant difference of daily dietary energy intakes was observed between attempters,

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

29

Table 2 Selected characteristics of the suicide attempters, the individuals with suicide ideation, and the controlsa Men (N = 1467) Control (n = 1269) Categorical variables, % (SE) Education: below high school Income: below poverty line Current smoker: yes Self-evaluated health: poor Try to lose weight: no Obese (BMI z 30 mg/m2): yes Cancer/pulmonary disease: yes Lost appetite symptomb: yes Depression: yes

21.26 11.84 36.13 0.33 28.19 17.30 8.26 7.63 1.03

Continuous variables, Mean(SE) Daily dietary energy intake, 3090.31 kcal/day Vitamin E, Ag/dl 961.13 a-Carotene, Ag/dl 3.37 h-Carotene, Ag/dl 13.5 Total Cholesterol, mg/dl 189.34 LDL-cholesterol, mg/dl 119.56 HDL-cholesterol, mg/dl 46.54

(1.34) (1.62) (1.89) (0.15) (1.60) (1.44) (1.31) (1.34) (0.35)

Women (N = 1770) Ideation (n = 151) 15.91 11.53 46.30 0.15 27.80 16.52 5.53 27.48 20.48

Attempt (n = 47) (3.38) (3.14) (5.74) (0.15) (5.23) (5.14) (2.14) (5.71)* (5.47)*

45.17 66.75 85.20 6.29 21.81 2.61 38.25 39.44 29.53

Control (n = 1419) (10.19)* (9.60)* (6.07)* (6.03)* (7.78)* (1.35)* (9.89)* (10.17) * (8.65)*

(54.41) 3137.37 (111.54) 2963.57 (486.67) (14.71) (0.14) (0.46) (1.84) (1.75) (0.75)

972.50 2.97 14.31 189.61 121.33 43.85

(33.61) (0.30) (1.47) (3.40) (3.28) (1.32)

896.58 2.07 11.20 182.77 109.42 49.91

(20.59)** (0.28)* (1.32)** (4.88) (5.63) (3.46)

16.35 15.81 31.11 0.46 50.53 17.67 13.38 20.69 5.99

(1.48) (1.59) (2.61) (0.21) (2.20) (1.77) (1.47) (2.19) (1.41)

Ideation (n = 224) 15.58 18.03 38.00 0.80 50.29 15.48 21.74 33.20 30.01

Attempt (n = 127) (2.84) (4.40) (4.10) (0.37) (4.68) (2.94) (4.05) (4.13) (4.54)b

44.37 32.13 55.97 1.18 68.09 20.98 26.55 35.29 32.06

(6.66)* (6.47)* (8.69)* (0.77)* (6.40)* (5.93)* (7.79) (4.76)* (5.85)*

1926.83 (42.76) 2050.34 (117.91) 2050.30 (185.76) 961.43 4.08 17.46 185.93 110.76 55.73

(16.83) (0.19) (0.73) (2.14) (1.88) (0.63)

944.43 4.29 18.09 185.50 112.68 53.63

(17.51) (0.59) (2.18) (3.92) (4.15) (1.27)

934.27 3.25 14.18 182.09 105.25 50.61

(38.22) (0.52) (2.13) (5.20) (4.10) (2.68)

Sample of 3237 adults aged 17–39 years, NHANES III, 1988–1994. BMI=body mass index; HDL=high density lipoprotein; LDL=low density lipoprotein; SE=standard error; NHANES III=The Third National Health Examination and Nutrition Survey. *p b 0.05; **p b 0.01. a The individuals with neither suicide ideation nor suicide attempts. b Last month only.

individuals with ideation and the controls within each gender. For both men and women, the means of serum TC and LDL-C of attempters were lower than that of the controls and the individuals with ideation. The associations of HDL-C with suicidal behaviors differed between genders. Among men, HDL-C was higher in attempters than in the controls and individuals with suicide ideation. In contrast, among women, HDL-C was lower in attempters than in these with suicide ideation and the controls. These unadjusted means, however, were not statistically different either in men or in women. In men, neither unadjusted nor adjusted OR of low HDL-C significantly differed from the null value (Table 3). LDL-C was found positively associated with ideation, but this association disappeared as a consequence of multivariable adjustment. In women, after multivariable adjustment, the only significant association ( p b 0.05) was between HDL-C and suicide attempts (OR = 2.93, 95% CI = 1.07–8.00).

Due to a relative small number of cases, the estimate was significant but lacked precision.

4. Discussion Using a national survey, we have documented a significant association between increased prevalence of life-time suicide attempts and low HDL-C in young women. No significant evidence was found to support an association between serum cholesterol and suicide ideation in young women. Serum cholesterol was unrelated with either suicide ideation or suicide attempts among young men. To the best of our knowledge, as of today, there are only two studies(Modai et al., 1994; Zale et al., 1996) examining the differences of serum cholesterol concentrations between suicide attempters, individuals with suicide ideation, and the controls simultaneously. These two studies and the current one used an iden-

30

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

Table 3 Adjusted association between low serum cholesterol, suicide ideation and attempts Men Ideation Total cholesterol High level, z200 mg/dl Na Prevalence (SE)b Low level, b200 mg/dl Na Prevalence (SE)b OR (95% CI)c

Women Attempters

Ideation

52 13.90 (2.96)

14 2.83 (1.09)

99 12.75 (1.74) 1.22 (0.62–2.40)

33 4.84 (1.23) 1.53 (0.35–3.30)

153 14.42 (2.00) 1.23 (0.75–2.00)

95 7.56 (1.06) 1.24 (0.66–3.57)

23 3.16 (0.94)

131 14.16 (1.90)

62 5.57 (1.03)

Low density lipoprotein cholesterol High level, z100 mg/dl Na 104 Prevalence (SE)b 16.06 (2.33) Low level, b100 mg/dl Na 39 8.94 (1.86) Prevalence (SE)b OR (95% CI)c 0.59 (0.33–1.07) High density lipoprotein cholesterol High level, z40 mg/dl Na 106 Prevalence (SE)b 12.44 (1.57) Low level, b40 mg/dl Na 45 Prevalence (SE)b 15.17 (3.44) OR (95% CI)c 1.36 (0.77–2.40)

17 5.51 (2.02) 1.56 (0.73–3.32)

33 4.35 (1.24) 14 3.69 (1.87) 1.27 (0.33–4.95)

71 12.97 (2.07)

Attempters

82 13.19 (2.13) 0.86 (0.53–1.37)

193 14.12 (1.70) 31 13.14 (2.75) 1.05 (0.58–1.92)

32 5.89 (1.66)

53 8.08 (1.43) 1.47 (0.87–2.49)

93 5.52 (0.90) 32 16.20 (4.76) 2.93 (1.07–8.00)

Sample of 3237 adults aged 17–39 years, NHANES III, 1988–1994 (Adjusted for social demographic variables, social/familial support factors, health risk and nutrition status, a history of medical and psychiatric illness (including depression) and other cholesterols). a Unweighted sample size. b Weighted prevalence. c Using the high level as the reference.

tical protocol to identify the individuals with suicide ideation and attempters but were carried out in substantially different settings. One (Modai et al., 1994) was conducted among psychiatric inpatients with diversified diagnoses, the other with depressed outpatients (Sullivan et al., 1994), while the current one was among a relatively young general population. In all these studies, serum TC of attempters was lower than that of the controls and the individuals with suicide ideation. The differences between groups, however, were much smaller (less than 7 mg/dl between the controls and attempters in men, and less than 4 mg/dl in women) and not statistically significant in the current study; while in the studies of Sullivan et al. and Modai et al., the differences between the controls and attempters were as much

as 35 and 15 mg/dl, respectively. Instead, we observed a significant difference in HDL-C concentration between attempters and controls among women. These discrepancies may be due to the differences of age, gender proportion and background psychiatric illness of the study participants. If impulsivity was defined as the rapidity and probability of acting on powerful feelings, the individuals with suicide ideation might be a group with low impulsivities, while the attempters have high impulsivities. Therefore, the findings of the current study may serve as indirect evidence to support an association between serum cholesterol and suicidal impulsivity. The finding that serum HDL-C was inversely and significantly associated with suicide attempts in women is consistent with previous studies. Maes et

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

al.(Maes et al., 1997) found a significantly lower serum HDL level in the depressed suicide attempters than in the depressed non-attempters. Buydens-Branchey et al. (Buydens-Branch et al., 2000) observed that in psychiatric patients, low HDL-C was significantly associated with blunted cortisol responses to meta-chlorophenylpiperazine, which has been interpreted as an indicator of impulsivity alterations in a variety of psychiatric conditions. The current study supports the hypothesis that the most important changes in serum cholesterol composition in suicide attempters occurred in HDL-C rather than TC or LDL-C. A perplexing question around the association between serum cholesterol and suicides is the tripod relationship among cholesterol, depression and cardiovascular diseases. Hypercholesterolemia has been well known as a predisposing factor of cardiovascular diseases. A high comorbidity between major depression and coronary heart diseases and myocardial infarction has been observed (Ferketich et al., 2000). If low TC was linked to increased depression, and further, to suicidality, it would be difficult to interpret the correlations between depression and coronary artery diseases (Su et al., 2000). The findings of the current study are in agreement with the expected changes in cholesterol metabolism in the comorbidity between depression and cardiovascular diseases. Low HDL-C may be a common cause for depression, suicide attempts, and cardiovascular diseases among women. A plausible mechanism for an association between cholesterol and suicide attempts has been proposed involving reduction of central serotonin activity (Kaplan et al., 1997). Other possible mechanisms have described a central role for fatty acids, particularly docosahexaenoic acid (Hibbeln and Salem, 1995). Inflammatory responses are increasingly recognized as being important in serotonergic activity and the development of depression, HDL-C has antiinflammatory properties as well (Cockerill et al., 2001). In each case, the proximal neurobiological mechanism remains to be identified. Gender difference is one of the key findings of the current study. Previous experimental studies revealed this difference. Goodwin et al. (Goodwin et al., 1987) and Anderson et al. (Anderson et al., 1990) have shown that moderate dieting increased the prolactin response to l-

31

tryptophan, the 5-hydroxytryptamine (5-HT) precursor, in women but not men. Since the prolactin response to l-tryptophan is probably mediated via brain 5-HT pathways (Cowen et al., 1992), this suggests that women may be particularly vulnerable to dieting-induced changes in brain 5-HT function. Perhaps the greatest limitation of the current study is its cross-sectional design. This inherent limitation prevented us from investigating causality. The ability to detect an association between cholesterol and suicidality was constrained by not sampling the individuals who may have been institutionalized due to severe medical consequences of suicide attempts, or the individuals who completed suicides. Serum samples were collected only once, which may misrepresent the true serum cholesterol before attempting suicide. All these limitations led to measurement errors with ensuing loss of precision, resulting in an attenuation of the true relationship, rather than a production of spurious ones. The current study has strengths as well. Richness of NHANES III allows us to delineate the relation between serum cholesterol and suicidality after adjustment for a wide array of covariates. To the best of our knowledge, no study to date has investigated the association by controlling for serum vitamins. A frequent criticism of the studies reporting an association between low cholesterol and suicidal behaviors has been related to the fact that depression could conceivably result in low cholesterol from depression-induced anorexia, and may account for the association observed. Extra efforts have been exercised in the current study to control for possible confounding effects from this pathway by using five variables: BMI, loss of weight in the last month, the history of reduced appetites, daily dietary energy intake and life-time depression. Additionally, compared with previous studies, the cut-points of cholesterol we used were uniquely based on the recommendations of NCEP. The findings, therefore, are much more practically relevant than using statistical judgment to categorize the cholesterol concentrations. The participants of NHANES III were randomly selected from the community-dwelling population; our conclusion should be more generalizable. In summary, the current study found that low serum HDL-C was significantly associated with suicide attempts, but not with suicide ideation in women. Although we cannot support or reject a causal asso-

32

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33

ciation between cholesterol and suicidal behaviors, the current study at the very least suggests that this phenomenon requires further attention. An understanding of how serum cholesterol relates to the suicidal behaviors may generate a biological marker or targets for novel therapeutic strategies.

References Anderson, I.M., Parry-Billings, M., Newsholme, E.A., Fairburn, C.G., Cowen, P.J., 1990. Dieting reduces plasma tryptophan and alters brain 5-HT function in women. Psychol. Med. 20, 785 – 791. Atmaca, M., Kuloglu, M., Tezcan, E., Gecici, O., Ustundag, B., 2002. Serum cholesterol and leptin levels in patients with borderline personality disorder. Neuropsychobiology 45, 167 – 171. Atmaca, M., Kuloglu, M., Tezcan, E., Ustundag, B., Gecici, O., Firidin, B., 2002. Serum leptin and cholesterol values in suicide attempters. Neuropsychobiology 45, 124 – 127. Atmaca, M., Kuloglu, M., Tezcan, E., Ustundag, B., 2003. Serum leptin and cholesterol levels in schizophrenic patients with and without suicide attempts. Acta Psychiatr. Scand. 108, 208 – 214. Buydens-Branchey, Branchey, M., Hudson, J., Fergeson, P., 2000. Low HDL cholesterol, aggression and altered central serotonergic activity. Psychiatry Res. 93, 93 – 102. Cockerill, G.W., Huehns, T.Y., Weerasinghe, A., Stocker, C., Lerch, P.G., Miller, N.E, et al., 2001. Elevation of plasma high-density lipoprotein concentration reduces interleukin-1-induced expression of E-selectin in an in vivo model of acute inflammation. Circulation 103, 108 – 112. Cowen, P.J., Anderson, I.M, Fairburn, C.G., 1992. In the biology of feast and famine. In: Anderson, G.H.K.S.H. (Ed.), Neurochemical Effects of Dieting: Relevance to Changes in Eating and Affective Disorders. Academic Press, San Diego, CA, pp. 269 – 284. Deisenhammer, E.A., Kramer-Reinstadler, K., Liensberger, D., Kemmler, G., Hinterhuber, H., Wolfgang, F.W., 2004. No evidence for an association between serum cholesterol and the course of depression and suicidality. Psychiatry Res. 121, 253 – 261. Druss, B., Pincus, H., 2000. Suicidal ideation and suicide attempts in general medical illnesses. Arch. Intern. Med. 160, 1522 – 1526. Ellison, L.F., Morrison, H.I., 2001. Low serum cholesterol concentration and risk of suicide. Epidemiology 12, 168 – 172. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III), 2001. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 285, 2486 – 2497. Ferketich, A.K., Schwartzbaum, J.A., Frid, D.J., Moeschberger, M.L., 2000. Depression as an antecedent to heart disease among women and men in the NHANES I study. National

Health and Nutrition Examination Survey. Arch. Intern. Med. 160, 1261 – 1268. Garland, M., Hickey, D., Corvin, A., Golden, J., Fitzpatrick, P., Cunningham, S., et al., 2000. Total serum cholesterol in relation to psychological correlates in parasuicide. Br. J. Psychiatry 177, 77 – 83. Goodwin, G.M., Fairburn, C.G., Cowen, P.J., 1987. Dieting changes serotonergic function in women, not men: implications for the aetiology of anorexia nervosa? Psychol. Med. 17, 839 – 842. Hibbeln, J.R., Salem Jr., N., 1995. Dietary polyunsaturated fatty acids and depression: when cholesterol does not satisfy. Am. J. Clin. Nutr. 62, 1 – 9. Jacobs, D., Blackburn, H., Higgins, M., Reed, D., Iso, H., McMillan, G., et al., 1992. Report of the conference on low blood cholesterol: mortality associations. Circulation 86, 1046 – 1060. Kaplan, J.R., Muldoon, M.F., Manuck, S.B., Mann, J.J., 1997. Assessing the observed relationship between low cholesterol and violence-related mortality. Implications for suicide risk. Ann. N. Y. Acad. Sci. 836, 57 – 80. Kim, Y.K., Lee, H.J., Kim, J.Y., Yoon, D.K., Choi, S.H., Lee, M.S., 2002. Low serum cholesterol is correlated to suicidality in a Korean sample. Acta Psychiatr. Scand. 105, 141 – 148. Kim, Y.K., Myint, A.M., 2004. Clinical application of low serum cholesterol as an indicator for suicide risk in major depression. J. Affect. Disord. 81, 161 – 166. Knox, S.S., Jacobs Jr., D.R., Chesney, M.A., Raczynski, J., McCreath, H., 1996. Psychosocial factors and plasma lipids in black and white young adults: the coronary artery risk development in young adults study data. Psychosom. Med. 58, 365 – 373. Lalovic, A., Merkens, L., Russell, L., Arsenault-Lapierre, G., Nowaczyk, M.J., Porter, F.D., et al., 2004. Cholesterol metabolism and suicidality in Smith–Lemli–Opitz syndrome carriers. Am. J. Psychiatry 161, 2123 – 2126. Lee, H.J., Kim, Y.K., 2003. Serum lipid levels and suicide attempts. Acta Psychiatr. Scand. 108, 215 – 221. Lindberg, G., Rastam, L., Gullberg, B., Eklund, G.A., 1992. Low serum cholesterol concentration and short term mortality from injuries in men and women. BMJ 305, 277 – 279. Maes, M., Smith, R., Christophe, A., Vandoolaeghe, E., Van Gastel, A., Neels, H., et al., 1997. Lower serum high-density lipoprotein cholesterol (HDL-C) in major depression and in depressed men with serious suicidal attempts: relationship with immune-inflammatory markers. Acta Psychiatr. Scand. 95, 212 – 221. Modai, I., Valevski, A., Dror, S., Weizman, A., 1994. Serum cholesterol levels and suicidal tendencies in psychiatric inpatients. J. Clin. Psychiatry 55, 252 – 254. Moscicki, E.K., 1994. Gender differences in completed and attempted suicides. Ann. Epidemiol. 4, 152 – 158. Muldoon, M.F., Manuck, S.B., Matthews, K.A., 1990. Lowering cholesterol concentrations and mortality: a quantitative review of primary prevention trials. BMJ 301, 309 – 314. National Center for Health Statistics, 1996. The Third National Health and Nutrition Examination Survey (NHANES III, 1988–94): reference manuals and reports. In Hyattsville, MD:

J. Zhang et al. / Journal of Affective Disorders 89 (2005) 25–33 US Department of Health and Human Service, Public Health Service, Center for Disease Control and Prevention. Neaton, J.D., Blackburn, H., Jacobs, D., Kuller, L., Lee, D.J., Sherwin, R., et al., Multiple Risk Factor Intervention Trial Research Group, 1992. Serum cholesterol level and mortality findings for men screened in the multiple risk factor intervention trial. Arch. Intern. Med. 152, 1490 – 1500. Ozer, O.A., Kutanis, R., Agargun, M.Y., Besiroglu, L., Bal, A.C., Selvi, Y., et al., 2004. Serum lipid levels, suicidality, and panic disorder. Compr. Psychiatry 45, 95 – 98. Partonen, T., Haukka, J., Virtamo, J., Taylor, P.R., Lonnqvist, J., 1999. Association of low serum total cholesterol with major depression and suicide. Br. J. Psychiatry 175, 259 – 262. Repo-Tiihonen, E., Halonen, P., Tiihonen, J., Virkkunen, M., 2002. Total serum cholesterol level, violent criminal offences, suicidal behavior, mortality and the appearance of conduct disorder in Finnish male criminal offenders with antisocial personality disorder. Eur. Arch. Psychiatry Clin. Neurosci. 252, 8 – 11. Shah, B.V., Barnwell, G.B., Bieler, G.S., 1997. SUDAAN, Software for the Statistical Analysis of Correlated Data, Release 7.5 ed. User’s Manual, vol 1. Research Triangle Institute, Research Triangle Park, NC.

33

Su, K.P., Tsai, S.Y., Huang, S.Y., 2000. Cholesterol, depression and suicide. Br. J. Psychiatry 176, 399 – 400. Sullivan, P.F., Joyce, P.R., Bulik, C.M., Mulder, R.T., OakleyBrowne, M., 1994. Total cholesterol and suicidality in depression. Biol. Psychiatry 36, 472 – 477. Tanskanen, A., Tuomilehto, J., Viinamaki, H., 2000. Cholesterol, depression and suicide. Br. J. Psychiatry 176, 398 – 399. Tanskanen, A., Vartiainen, E., Tuomilehto, J., Viinamaki, H., Lehtonen, J., Puska, P., 2000. High serum cholesterol and risk of suicide. Am. J. Psychiatry 157, 648 – 650. Tripodianakis, J., Markianos, M., Sarantidis, D., Agouridaki, M., 2002. Biogenic amine turnover and serum cholesterol in suicide attempt. Eur. Arch. Psychiatry Clin. Neurosci. 252, 38 – 43. Vevera, J., Zukov, I., Morcinek, T., Papezova, H., 2003. Cholesterol concentrations in violent and non-violent women suicide attempters. Eur. Psychiatry 18, 23 – 27. Zale, C., New, A.J., Trestman, R.L., 1996. Serum cholesterol and impulsive aggressive behavior in personality disorder patients (abstract). Biol. Psychiatry 39, 535. Zureik, M., Courbon, D., Ducimetiere, P., 1996. Serum cholesterol concentration and death from suicide in men: Paris prospective study I. BMJ 313, 649 – 651.