The first study group consisted of 43 male veterans with chronic combat-related PTSD, aged between 26 and 47 yr. The duration of their combat activity ranged from 1 to 4 yr. The number of their combat traumas ranged from 2-5. Time elapsed since they had experienced combat traumas ranged from 6 to 10 yr. The second study group consisted of 37 male veterans with chronic combat-related PTSD comorbid with MDD, aged between 25 and 45 yr. The duration of their combat activity ranged from 1 to 4 yr. The number of their combat traumas ranged from 1-5. Time elapsed since they had experienced combat traumas ranged 6 to 10 yr. The third study group consisted of 38 male veterans with MDD, aged between 25 and 46 yr. The duration of their combat activity ranged from 1 to 4 yr. The number of their combat traumas ranged from 2-4. Time elapsed since they had experienced combat traumas ranged from 6 to 10 yr. Subjects from all study groups, when they had been drafted for active military service, they went through rigorous entrance tests, which ensured that they had not had any previous major psychiatric or somatic disorders. All investigated patients did not have any other psychiatric comorbid disorders or medical problems, and took different medications in medical history prior to this study, mostly antidepressants (fluoxetine 8.9%, fluvoxamine 5.2%, paroxetine 25.4%, sertraline 27%, clomipramine 18.7%, trazadone 1.4%, and maprotiline 13.4%) and anxiolytics (alprazolam 18.7%, oxazepam 8.6%, diazepam 57.6%, clonazepam 15.1%). 75% of the patients sometimes took some hypnotics during therapy (nitrazepam, funitrazepam or zolpidem). None of the above mentioned drugs influence serum lipids concentrations (23). The control group consisted of 39 healthy men, mostly our hospital workers, aged 28-47 yr. They had negative anamnesis data of alcohol or other drugs of dependence abuse. Informed consent was obtained from patients and controls after complete and extensive description of the study profile.

Because lipid concentration are sensitive to even relatively subtle psychosocial influences, our experience has indicated that it is important to describe these factors in some details, which characterize PTSD patients as well as the control group. Descriptive parameters of all included subjects are presented in Table 1. Almost half of the subjects were married (65.1% in PTSD, 48.6% in PTSD comorbid with MDD, 60.5% in MDD and 58.3% in the control group). Most of the subjects had secondary schooling (74.4% in PTSD, 62.2% in PTSD comorbid with MDD, 68.4% in MDD, and 71.8% in control group). Most of them lived in urban areas (60.5% in PTSD, 56.8% in PTSD comorbid with MDD, 60.5% in MDD, and 69.4% in control group). Subjects with PTSD, PTSD comorbid with MDD, or MDD were in inpatient hospital treatment. During hospitalization, they undertook an intensive schedule of 30 hr of individual and group therapy per week, as well as other socio-therapeutic activities. Treatment program was based on the therapeutic community (24).

From the pool of all male patients (n=891) who were receiving treatment in the University Department of Psychiatry in "Sestre milosrdnice" University Hospital in 2001 and 2002, we elected three study groups; patients with combat-related PTSD only (n=43) or comorbid with MDD (n=37), and patients with combat-experiences with MDD only (n=38). Patients with other psychiatric disorders (n=673), and comorbid psychiatric disorders in PTSD group (n=63), (except MDD in PTSD group) or somatic conditions (n=28) were excluded from the study in all study groups. Also because the patients did not give informed consent we excluded them from the study (2 in PTSD group, 3 in PTSD comorbid with MDD group, and 4 in MDD group).

A psychiatrist performed the structured clinical interview and made diagnosis based on DSM-IV criteria for PTSD or/and MDD (25). Clinical psychologist applied Clinician Administered PTSD Scale (CAPS) PTSD interview based on DSM-IV criteria to measure post-traumatic stress reaction (26). Final diagnosis of PTSD was reached only in cases where both sets of criteria were fulfilled. Diagnostic agreement between psychiatric and psychological criteria was high κ=0.96. The alpha coefficient for the CAPS questionnaire in our study was 0.91. We accepted the diagnosis of MDD if the patients fulfilled two criteria: indicative result on the Montgomery-Asberg Depression Rating Scale (MADRAS) (27) and sufficient number of symptoms shown in the structural clinical interview based on DSM-IV criteria for MDD (25). Diagnostic balance between the MADRAS and clinical interview based on DSM-IV criteria for MDD was κ=0.98. MADRAS questionnaire was used because of its simplicity and reliability, in our study alpha coefficient was 0.93. Two psychiatrists performed this part of the evaluation, each of them examining independently all the subjects. The agreement between two psychiatrists was high, κ=0.97.

We also compared PTSD group and MDD group according to DSM IV criteria for nicotine dependence (25) because smoking has a strong effect on elevation of serum lipids (28). And last, body mass index (BMI) was calculated for PTSP and MDD group because BMI correlated well with serum lipids level (28). BMI is ratio representing each patients weight in kilograms divided by square of height in meters (BMI=kg/m²). Since diet and activity can influence the results of examinations, laboratory measures and psychiatric examinations were carried out between days 14 and 16 of inpatient stay. Blood lipids measurements at this time may be preferable to values on admission to ensure that all patients had equivalent diets and activity levels.

Blood samples were collected from forearm vein, in glass red-topped, vacuum tubes without any anticoagulant, in the morning between 8-9 a.m., after overnight fast of 12 hr and 30 min resting immediately prior to blood collection. Serum concentrations for cholesterol, HDL, and triglyceride were measured by enzyme assay (EA), immediately after taking the samples with commercial kits (Olympus Diagnostic, GmbH, Hamburg, Germany) on Olympus AU 600 automatic analyzator. The interassay coefficient of variation in our laboratory was 3.2% for cholesterol, 2.5% for triglycerides, and 3.0% for HDL-C. Serum LDL-C concentrations were calculated by following formula: LDL-C=cholesterol-(HDL-C -tryglicerides/5)] (29). Our laboratory's referent intervals for lipids are parameters measured as follow: cholesterol 147-220 mg/dL, LDL lower than 150.0 mg/dL, HDL higher than 42.0 mg/dL, LDL-C/HDL-C ratio lower than 130.0, and triglyceride 53-177 mg/dL.

We use chi-quadrate test to analyze difference in nicotine dependence or sociodemographic variables between four groups. The normal distribution was assessed for all measures and for each group by Kolmogorov-Smirnov test. Subjects with PTSD, subjects with PTSD comorbid with MDD, subjects with MDD, and control group (except in war variables for control group) were compared in age, blood glucose concentration, systolic or diastolic blood pressure, duration of combat activity, duration of PTSD symptoms, and number of combat traumas by analysis of variance (ANOVA) and Scheffe's post-hoc analysis for pair-wise comparisons. Analysis of covariance (ANCOVA) with age, BMI, blood glucose concentrations, diastolic or systolic blood pressure as a covariates, was used to analyze differences in serum cholesterol, triglycerides, LDL-C, HDL-C, and LDL-C/HDL-C ratio in subjects with PTSD, subjects with PTSD comorbid with MDD, subjects with MDD, and control group. Age, BMI, blood glucose concentrations, diastolic or systolic blood pressure was used as covariates because that variable can influence serum lipid concentrations (28).

A p value of <0.01 was considered to denote the presence of a statistically significant difference. Statistics was done with SPSS software (SPSS for Windows 8.0, SPSS, Chicago, IL, U.S.A.).