Introduction
Several studies focused on the potential link between neurobiological factors and suicidal behaviors.Reference Kim and Lee 1 Of note, the possible role of lipids in modulating mental health was hypothesized according to the observations of an increased number of attempted suicides in subjects treated with lipid-lowering agents,Reference Muldoon, Manuck, Mendelsohn, Kaplan and Belle 2 or conversely of an amelioration of mood symptoms in patients affected by major depression and taking statins.Reference Köhler-Forsberg, Otte, Gold and Østergaard 3 In this regard, some studies suggested that statins may have some anti-inflammatory and antioxidant properties, possibly targeting the inflammatory and oxidative pathways associated with the pathophysiology of depression.Reference O'Neil, Sanna and Redlich 4 Furthermore, because patients using statins exhibit a limited risk of cardiovascular events, it cannot be ruled out that a better quality of life may be in turn associated with mood improvements.Reference Yang, Jick and Jick 5 Despite this, different neuropsychiatric adverse drug reactions such as violence, aggression, intentional injury, and mood changes were described in some observational studies as well as in pharmacovigilance databases regarding lipid-lowering drugs.Reference Tatley and Savage 6 However, potential confounders in the association between lipid-lowering drugs and behavior modifications were often not taken into account. For instance, lack of exercise is related to higher risk of cardiovascular accidents and was found to be associated with depression. Because hyperlipidemic individuals having this risk factor could be more likely to assume statins than hyper-lipidemic individuals who practice physical exercise, not controlling for physical exercise should turn out to an apparent relationship between statin use and depression.Reference Conklin and Stanford 7 In light of these gaps, in the last years, a growing number of studies has investigated the possible relationship between lower serum lipid levels and suicidal behavior on the grounds that suicidal behaviors are more likely to be not a result of lipid lowering therapies, but basically of lower lipid levels.Reference Leppien, Mulcahy, Demler, Trigoboff and Opler 8 In support of these clinical observations, some biological data show that low peripheral cholesterol and triglyceride levels might be associated with decreased lipids in synaptic membranes, leading to consistent changes in membrane viscosity and gene expression.Reference Freemantle, Mechawar and Turecki 9 These variations, in turn, could result in a poor serotonergic activity in lipid rafts, possibly related to a limited suppression of impulsive violent behaviors, such as suicidal conducts.Reference Vevera, Fisar and Kvasnicka 10 , Reference da Graça Cantarelli, Nardin and Buffon 11 An inverse association of suicidality with serum cholesterol and triglyceride levels was also remarked in a recent systematic review and meta-analysis.Reference Wu, Ding, Wu, Xie, Hou and Mao 12 Even though subgroup analyses revealed a potential association between a lower serum total cholesterol and violent methods of suicide attempts, some studies reported opposite results.Reference Wu, Ding, Wu, Xie, Hou and Mao 12 As a whole, the studies exploring a possible interaction between lipid serum profile and suicidal behavior in subjects with mental disorders produced mixed results, possibly due to methodological heterogeneity.Reference Atmaca, Kuloglu, Tezcan and Ustundag 13 , Reference Capuzzi, Bartoli, Crocamo, Malerba, Clerici and Carrà 14 The available data about this topic are also biased by confounding factors such as gender because males are more likely to use violent methods to commit suicide than women.Reference D'Ambrosio, Salvi, Bogetto and Maina 15 Furthermore, no robust evidence associates the severity of impulsivity with serum lipid profile,Reference Shaker, Sultan, Mohamed, Helal and Abd El Moneam 16 partly due to the fact that other biological factors such as the amount of circulating neurosteroidsReference Cantarelli Mda, Tramontina, Leite and Gonçalves 17 or severity of systemic inflammationReference Maes, Smith and Christophe 18 can both influence the cholesterol blood levels and the vulnerability to suicidal behaviors. In this regard, recent models included neuroinflammation as an important factor associated with the vulnerability to impulsive behaviors including suicide.Reference Capuzzi, Caldiroli, Capellazzi, Tagliabue, Buoli and Clerici 19 Recent studies identified neutrophil-to-lymphocyte ratio (NLR) as a measure of low-grade inflammation in different systemic diseasesReference Howard, Kanetsky and Egan 20 as well as a potential biomarker related to the severity of psychiatric disorders and suicide.Reference Ekinci and Ekinci 21 Specifically, a recent meta-analysis reported higher levels of NLR in subjects with bipolar disorder (BD) and major depressive disorder (MDD) than healthy controls.Reference Mazza, Lucchi, Tringali, Rossetti, Botti and Clerici 22 Further studies even suggest that NLR may be a trait marker of suicide vulnerability in patients with BDReference Ivković, Pantović-Stefanović and Dunjić-Kostić 23 or MDD,Reference Ekinci and Ekinci 21 , Reference Velasco, Rodríguez-Revuelta and Olié 24 although other findings contradicted these results.Reference Gundogdu Meydaneri and Meydaneri 25 Furthermore, two different articlesReference Orum, Kara and Egilmez 26 , Reference Kara, Orum and Egilmez 27 reported higher levels of NLR in subjects who attempted suicide with a violent method than the counterpart.
According to authors’ knowledge, no study has explored till now the concomitant association of lipid serum levels and NLR with violent/nonviolent method of suicide attempt. Thus, in order to overcome gaps of previous research, purpose of the present study was to explore the possible roles of lipid serum levels and of NLR on the occurrence of violent suicide attempts in a sample of patients affected by psychiatric conditions.
Methods
This study was drawn up following the Strengthening the Reporting of Observational studies in Epidemiology Statement items.Reference von Elm, Altman and Egger 28 The research project complied with the principles of the Declaration of Helsinki regarding medical research in humans, following local research ethical requirements.
Sample
A sample of 163 subjects aged ≥18 years was cross-sectionally recruited in a Psychiatric Inpatient Unit (Desio Hospital, ASST Monza, Italy), between January 2012 and December 2019. All patients were included if hospitalized for a suicide attempt occurred within the last 72 hours and if suffered from schizophrenia, BD type I or II (any acute phase), MDD and personality disorders according to ICD-10 (International Classification of Diseases) criteria. If patient had more than one psychiatric diagnosis, the main psychiatric condition was taken into account (ie, the disorder that conducted to hospitalization). Moreover, eligible subjects were medically stable, not needing treatment for any physical condition. We excluded subjects suffering from other mental disorders or mental retardation, and those with serious physical illnesses, including hepatitis B or C, and HIV infections (antiviral agents may modify lipid profile), or treated with thyroid hormone, antidiabetic, anticoagulant, antiplatelet, urate-, and lipid-lowering agents.
Assessments
All data were anonymously and retrospectively collected, not allowing subject identification. Standard demographic and clinical data (age, gender, nationality, occupational and marital status, characteristics of recent suicide attempts, diagnosis of mental disorders according ICD-10, ongoing psychopharmacological treatments, smoking status, co-occurring substance use disorders, and lipid profile) were retrieved from electronic health records. We considered as “suicide attempts,” only self-injuries of people willing to die, while suicidal ideation or plans without any action, as well as low-lethal self-harm behaviors, were excluded.Reference Velasco, Rodríguez-Revuelta and Olié 24 We used standard definitions to distinguish violent (firearm, hanging, cutting, jumping, car exhaust, and other violent methods) from nonviolent (drug overdose and poisoning) methods for attempting suicide.Reference Nock, Borges, Bromet, Cha, Kessler and Lee 29 Information on lipid serum levels, including total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), very low density lipoprotein (VLDL), triglycerides (TG), and NLR was retrieved from routine blood tests done in the first hours of hospitalization at 8.00 am, after an overnight fasting. We considered only serum blood tests carried out within 24 hours after the hospitalization. Serum lipid concentrations were examined using standard methods (enzymatic colorimetric ones) at the Department of Clinical Pathology of Desio Hospital, ASST Monza, Italy.
Statistical analyses
The data were analyzed using Statistical Package for Social Sciences—SPSS 26.0. Characteristics of the total sample were reported using mean and standard deviation for quantitative variables and frequencies and percentages for qualitative ones. Selected patients were divided into two groups according to the presence of violent or nonviolent suicide method. Quantitative and qualitative variables were respectively compared between the two groups by multivariate analyses of variance (MANOVAs) and Chi-square tests with eventual Bonferroni’s post hoc analyses. Finally, multivariate analyses of variance, inserting gender, and psychiatric diagnoses as covariates (MANCOVAs) were performed to compare the two groups in terms of lipid profile and NLR.
A p value of ≤.05 was considered as statistically significant.
Results
A total of 163 patients hospitalized for a recent suicide attempt were included in the study. Almost a third of the sample had experienced a violent suicide attempt in the 72 hours before admission. As a whole, 17 individuals had schizophrenia, 19 BD, 37 MDD, and 90 a personality disorder. More than 40% of the subjects was treated with antidepressants. Moreover, more than half of the sample was a smoker while about one in five patients had a comorbid substance use disorder. Descriptive analyses of the total sample and of the groups divided according to the type of suicide attempts are reported in Table 1.
Clinical Characteristics of the Total Sample and of the Two Groups Identified According to the Method of Suicide Attempt.
Values are numbers (column %), unless stated. P values in bold indicate statistically significant differences between groups.
Abbreviations: BD, bipolar disorder; FGA, first generation antipsychotics; MDD, major depressive disorder; SD, standard deviation; SGA, second generation.
* There are missing values for some variables: the greatest number of missing values is for smoking and comorbid substance use disorder with 79 (48.4%) and 132 (81%) missing values, respectively.
The groups of patients divided according to violent/nonviolent attempted suicide were significantly different in terms of gender (χ 2 = 9.46; df = 1; P < .01), antidepressant treatment (χ 2 = 15.85; df = 1; P < .001), and smoking status (χ 2 = 4.90; df = 1; P = .04). Indeed, male individuals were more likely to attempt suicide with a violent method than females. Furthermore, patients with a violent suicide attempt were more frequently smokers and were taking less antidepressant treatment than patients that attempted suicide in a nonviolent way. On the other hand, we did not find significant differences between the two groups in age (F = 1.44; P = .23), employment status (χ 2 = 0.28; df = 1; P = .60), marital status (χ 2 = 0.50; df = 1; P = .57), psychiatric diagnosis (χ 2 = 2.80; df = 3; P = .43), first-generation antipsychotic treatment (χ 2 = 1.14; df = 1; P = .32), second-generation antipsychotic treatment (χ 2 = 3.42; df = 1; P = .08), treatment with mood stabilizers (χ 2 = 0.48; df = 1; P = .63), and presence of a substance use disorder (χ 2 = 0.22; df = 1; P = .67).
Finally, serum levels of total cholesterol (F = 5.66; P = .02), LDL (F = 4.94; P = .03), VLDL (F = 5.66; P = .02), and NLR (F = 8.17; P < .01) were significantly lower among patients using a violent method compared to patients with a nonviolent method of suicide attempt (Figure 1).
Lipid profile and neutrophil/lymphocyte ratio in violent vs nonviolent suicide attempts.
Multivariate analyses of variance adjusted for gender and diagnosis. Abbreviations: HDL, High-Density Lipoprotein; LDL, Low-Density Lipoprotein; NLR, Neutrophil/Lymphocyte Ratio; VLDL, Very-Low-Density Lipoprotein.
Similarly, lower HDL (F = 0.006; P = .94) and TG (F = 0.05; P = .82) serum levels were found in violent vs nonviolent suicide attempters, but these differences did not achieve statistical significance.
Discussion
Main findings
In this study, based on real world sample of inpatients with different mental disorders, we found that subjects who attempted suicide with violent method showed lower serum levels of total cholesterol as well as of LDL and VLDL subfractions compared to their counterpart. In addition, NLR was significantly lower in violent vs nonviolent suicide attempters.
Our results are consistent with a previous meta-analysis by Wu et alReference Wu, Ding, Wu, Xie, Hou and Mao 12 and with a recent case–control studyReference Aguglia, Solano and Giacomini 30 which found lower serum total cholesterol and LDL levels in patients with high-lethality suicide attempts compared to low-lethality suicide attempters. Moreover, we found no differences in HDL and TG between nonviolent and violent suicide attempters similarly to previous studies.Reference Bartoli, Crocamo and Dakanalis 31 However, the literature exploring a possible interaction between low lipid levels and the method of suicide attempt in subjects with mental disorders produced mixed results.Reference da Graça Cantarelli, Nardin and Buffon 11 , Reference Papadopoulou, Markianos, Christodoulou and Lykouras 32 In particular, the high methodologically heterogeneity across studies due to the different assessment of suicide attempts,Reference Aguglia, Solano and Giacomini 30 the type of included mental disordersReference De Berardis, Marini and Piersanti 33 as well as the role of possible confounders (ie, economic factors, dietary patterns, body mass index [BMI], physical conditions, infectious disease, medical treatments, other psychological characteristics, and current psychopharmacological treatment) should be taken into account.Reference Capuzzi, Bartoli, Crocamo, Malerba, Clerici and Carrà 14 , Reference Aguglia, Solano and Giacomini 30 Nevertheless, the different subfractions of cholesterol (pro-atherogenic lipids, ie, LDL, VLDL, and TG vs anti-atherogenic lipids, ie, HDL) may have distinct roles in brain function and therefore in influencing the risk of suicide, although no definitive answers to this question exist.Reference Lutz, Mechawar and Turecki 34 Thus, different possible explanations were suggested to account for the association between cholesterol and suicidal behaviors.Reference Cantarelli Mda, Tramontina, Leite and Gonçalves 17 In particular, since violent suicide attempters usually have higher levels of aggressiveness and impulsivity, and more adult interpersonal violence and suicide risk than nonviolent attempters,Reference Stenbacka and Jokinen 35 serum cholesterol might be more linked to the violence of suicide attempts than to suicide itself.Reference Atmaca, Kuloglu, Tezcan and Ustundag 13 In support to this consideration, a correlation between infant exposure to violence and adult violent behaviors (ie, cycle of violence) was found in subjects with lower serum cholesterol levels.Reference Asellus, Nordström, Nordström and Jokinen 36 Early life adverse experiences (ELAs), such as sexual or physical abuse and parental neglect, are distal events related to suicide behaviors.Reference Zatti, Rosa and Barros 37 ELAs may produce chronic psychological stress and epigenetic modifications of multiple genes including those implicated in regulation of cholesterol levels and stress responses.Reference Asellus, Nordström, Nordström and Jokinen 36 In this regard, some studies reported that apolipoprotein E, that it is involved in the formation of synapses and prevention of neuronal death, might be associated with decreased lipid levels, which in turn might be related to the severity of suicidal behavior.Reference Asellus, Nordström, Nordström and Jokinen 38 , Reference Asellus, Nordström, Nordström and Jokinen 39 Nevertheless, some studies suggest that some cholesterol metabolites called oxysterols may cross the blood–brain barrier leading to modifications in central nervous system (CNS) cholesterol levels and lipid raft activity resulting in impairment of serotonin synaptic transmission.Reference Capuzzi, Caldiroli, Capellazzi, Tagliabue, Buoli and Clerici 19 As a consequence of altered phospholipid and cholesterol contents in CNS, a study reported that violent suicide completers may display a lower gray matter lipid content in both ventral prefrontal and orbitofrontal cortex than nonviolent suicidal people.Reference Freemantle, Mechawar and Turecki 9 However, the debate in the relationship between serum lipids and suicide is still ongoing. For example, an alternative hypothesis is that low serum cholesterol levels are explained by the over-production of neurosteroids promoting self-harm behaviors in suicide attempters such as dehydroepiandrosterone sulfate.Reference Chatzittofis, Nordström, Hellström, Arver, Åsberg and Jokinen 40 Furthermore, serum lipids may just represent an epiphenomenon of suicide attempts, interplaying with more complex processes, specifically associated with the dysregulation of the hypothalamic–pituitary–adrenal axis, and inflammation.Reference Walker and Spencer 41 In particular, interleukin-6 (IL-6), is hypothesized to be involved in the pathophysiology of suicidal behavior since it may promote inflammation with consequent astrocyte and microglial activation.Reference Song, Li and Zhang 42 IL-6 may be overproduced by microglia or may enter in the brain from the periphery via a compromised blood–brain barrier.Reference Pandey, Rizavi and Ren 43 In this regard, NLR is currently considered a reliable marker of low-grade systemic inflammation.Reference Mazza, Lucchi, Tringali, Rossetti, Botti and Clerici 22 Our study found that individuals who committed a suicide attempt with violent method had a lower NLR compared to nonviolent ones. This result seems to conflict with two consecutive studies,Reference Orum, Kara and Egilmez 26 , Reference Kara, Orum and Egilmez 27 which found higher NLR values in violent suicide attempters in comparison to nonviolent suicide attempters and healthy controls. However, it appears to be difficult to compare our results with those reported by OrumReference Orum, Kara and Egilmez 26 and KaraReference Kara, Orum and Egilmez 27 since they excluded patients with a history of psychiatric treatments and with current psychiatric disorders, apart from those with major depression or anxiety, while we included a sample of patients affected by psychiatric disorders requiring hospitalization. As a whole, to our knowledge, few articles investigated the potential role of NLR in suicidal behavior and these studies present different methods and samples than our research.Reference Gundogdu Meydaneri and Meydaneri 25 , Reference Ucuz and Kayhan Tetik 44 Nevertheless, even though most studies reported higher levels of NLR in suicide attempters than individuals without suicidal behavior, these results can be discordant respect to our findings for the following reasons. First of all, the severity of depression and suicide may be related to distinct immune-biological profiles in peripheral blood.Reference Keaton, Madaj and Heilman 45 This aspect is also supported by our findings as subjects with nonviolent suicide attempts were more frequently in treatment with antidepressants. In addition, a very recent study confirmed that the severity of depressive symptoms is linearly associated with LDL serum levels, but no changes in NLR.Reference Buoli, Caldiroli and Guenzani 46 Furthermore, suicidal behavior may be accompanied by an increased number of lymphocytes whereas depressed patients may be less likely to show elevated lymphocytes.Reference Barzilay, Lobel, Krivoy, Shlosberg, Weizman and Katz 47 Similarly, distinct patterns of inflammation might be associated with a different impairment in self-regulation and consequently with a variable risk of impulsivity and violent suicide attempts.Reference Shields, Moons and Slavich 48 , Reference Gassen, Prokosch and Eimerbrink 49 Second, most studies about the association between inflammation and suicide considered the total count of white blood cells while just few studies examined lymphocyte subtypes. For instance, patients affected by post-traumatic stress disorder, a condition associated with high risk of suicide, can show a decreased number or a decreased functional capacity of natural killer cells as a result of the high glucocorticoid receptor expression on lymphocytes.Reference Gotovac, Vidović and Vukusić 50 In addition, cortisol may promote the differentiation of regulatory T lymphocytes as well as repress the production of T-helper 1 and T-helper 17 lymphocytes.Reference Walker and Spencer 41 , Reference Bellavance and Rivest 51 Third, many possible confounders should be considered in the potential association between NLR and suicide attempts. In particular, our study found significant differences between nonviolent and violent suicide attempts in terms of gender, frequency of antidepressant treatment and smoking status. One study reported statistically significant differences in NLR between men and womenReference Meng, Chang and Liu 52 and another article found that the NLR is normalized by the antidepressant treatment in subjects affected by MDD.Reference Demir, Atli and Bulut 53 Finally, smoking seems to affect NLR valuesReference Tulgar, Cakar, Tulgar, Dalkilic, Cakiroglu and Uyanik 54 and a recent study, conducted on patients undergoing coronary intervention, reported that changes in the lipid profile influence the variability of NLR in the light of the inflammatory properties of atherogenic lipids.Reference Zhao, Xu and Li 55 Taken as a whole, in the light of the above mentioned considerations, we cannot fully exclude that lower serum levels of total cholesterol, LDL, and VLDL are concomitant with lower levels of NLR among patients using a violent method compared to patients with a nonviolent suicide attempt.
Limitations
There are some important methodological limitations in the present study and results should be interpreted cautiously. First, as this study was a cross-sectional one, we cannot definitively define whether the selected variables preceded or followed the suicide attempt. Consequently, we could not evaluate the direct causal relationship between suicidal behavior and lipid profile as well as NLR. Second, the sample size was relatively small and almost all subjects were Italian inpatients from one hospital, which limit the generalization of our findings. For the same reason, we were unable to stratify the sample according to the different type of mental disorders. Third, since data were retrieved from electronic medical charts, without information from face-to-face interviews, some variables including previous suicide attempts, poor adherence to medications, dietary habits, nutritional status, physical activity, and BMI were not available.Reference Vogelzangs, Beekman and van Reedt Dortland 56 , Reference Furuncuoğlu, Tulgar, Dogan, Cakar, Tulgar and Cakiroglu 57 Fourth, we should consider that most subjects were treated with psychopharmacological agents, leading to a possible alteration of lipid serum levels and NLR. Indeed, it is widely known that both first- and second-generation antipsychotics as well as antidepressant treatments may cause lipid dysregulation, thereby contributing to the onset of the metabolic syndrome. 58-60 Moreover most mood stabilizers, including valproic acid, carbamazepine, lamotrigine and lithium could be typically associated with alterations of Leukocyte Count, Reference Post, Ketter, Uhde and Ballenger 61 leading to possible change of NLR.Reference Spuch, López-García, Rivera-Baltanás, Rodrígues-Amorím and Olivares 62 However, we excluded subjects with physical conditions or medical treatments, in particular lipid lowering agents, that may interfere with lipid metabolism and NLR values. Furthermore, we considered only certain mental disorders, excluding those conditions (ie, eating disorders, organic mental disorders, and mental retardation) that might confound the association between lipids, NLR, and suicidal behaviors.Reference Diaz-Sastre, Baca-Garcia and Perez-Rodriguez 63 Nevertheless, we cannot fully ruled out that part of our sample may be affected by subclinical conditions (eg, hypothyroidism, cardiovascular and autoimmune diseases, and infections) which in turn may translate into alterations of both lipid metabolism and NLR. 64-66
Conclusions
Despite some limitations, the current study may support previous evidence of lower serum lipids in subjects with mental disorders and who attempted suicide with violent method as compared with nonviolent attempters. Moreover, low NLR levels might increase the risk of violent suicide attempts. To the best of our knowledge, no previous studies investigated lipid profiles concomitantly with NLR and other confounding factors in relation to the violence of suicide attempts. NLR could be a biomarker of clinical utility to stratify psychiatric subjects according to diagnoses or severity of symptoms.Reference Brinn and Stone 67 Further studies should therefore focus on this parameter, including other important markers of inflammation such as glucocorticoids, cytokines, and lymphocyte subpopulations. Reference Keaton, Madaj and Heilman 45 Longitudinal studies on larger samples of subjects, stratified by type of mental disorder, could better clarify the role of some biological pathways in the onset of suicidal behavior.
Disclosures
Enrico Capuzzi, Alice Caldiroli, Martina Capellazzi, Ilaria Tagliabue, Annamaria Auxilia, Giulia Ghilardi, Massimiliano Buoli, and Massimo Clerici have no conflict of interest with any commercial or other association in connection with the submitted article.
