MATERIALS AND METHODS

This was part of a hospital based case-control study aimed at comparing the general health status between patients having a recent stroke and a group of eligible controls. A total of 150 cases along with 150 controls were studied. Study participants were selected from those patients hospitalized in Alavi, Bouali and Fatemi University Hospitals located in Ardabil, north-west of Iran. Cases were selected from neurology specialty ward of Alavi university hospital which is the provincial referral neurology center. Controls were chosen from other provincial referral wards located in any of the above mentioned university hospitals. This selection strategy was taken to make sure that the reference community becomes similar both for cases and controls. Cases were those with a definite diagnosis of stroke and controls were patients hospitalized without a known cerebro-vascular, TIA and hypercoagulopathic disease history. Computerized tomographic scanning of the brain was accomplished for all cases having stroke to make out a differential diagnosis. Controls were matched with cases for sex and also for age in a three years acceptable age difference range. Participants were interviewed and a questionnaire containing both demographic information, known stroke risk factors and GHQ 28 questions was filled out for each person either in case or control group. Data were entered into a computer and were analyzed by SPSS version 13 statistical package. Chi-square test was used for comparing the proportions between groups and t test was used to compare means. Odds ratio and its 95% confidence interval was calculated using Epi Info 2002 statistical software everywhere primary tests showed a significant distribution difference. A p-value less than 0.05 was considered as statistical significance level. The study was approved by high degree committee of research and ethics. Verbal consent was taken from all participants letting them know that there is no obligation in participating.

RESULTS

Demographic characteristics in case and control groups were not statistically different and the comparative data are given in Table 1.

Although an acceptable total GHQ 28 Farsi questionnaire reliability was measured but in social function and somatization domains we found a lower reliability.

Mean total GHQ score was 13.02±4.8 (Mean±SD). It was 13.85 in case group and 12.36 in control group and the difference was statistically significant (p<0.01). However those having a high GHQ score hadn=t different distribution between two groups.

Table 1:	Demographic characteristics compared between case and control groups

Fig. 1:	Frequency distribution of total GHQ scores

Fig. 2:	Mean score and confidence intervals among GHQ domains compared between stroke and non stroke patients

Figure 1 shows the frequency distribution of total GHQ scores and Fig. 2 compares the mean and confidence intervals among GHQ domains.

Anxiety, Social function and depression scores were higher in stroke patients compared to control group using t-test (p<0.05). Although the mean score in any of the Anxiety, Social function and depression domains were higher among stroke patients but in none of these domains the percentage of patients with abnormal score was not statistically higher among stroke patients compared to control group.

Out of our objectives detected risk factors for CVA in this study were found to be smoking (OR = 16.2 and 95% CI: 3.65<OR<100) and hypertension (OR = 3.48 and 95% CI: 2.09<OR<5.08).

DISCUSSION

Mean total GHQ score in our study was 13.02. This score was as low as 5.7 in a large study by Willmott et al. (2004). This lower GHQ mean score may be due to different study populations who were hospitalized patients in our study contrary to general population studied by Sasi. Similarly the GHQ 30 mean score in a study by May et al. (2002) was 3.1 in a large cohort in South Wales.

As it can be found in Fig. 1, we found a near normal GHQ score distribution while Willmott et al. (2004) had found the total GHQ score distribution to be skewed and had concluded that GHQ score median is more reliable than its mean.

A systematic review by Hackett showed a pooled estimate of post stroke depression equal to 33% (95% confidence interval, 29 to 36%). Differences in case mix and method of mood assessment could explain some of the variation in estimates across studies (Hackett et al., 2005).

Given the importance of mood, which along with cognition, motivation and social support is a key factor influencing recovery from stroke, it is surprising that there is much misconception over the epidemiology of stroke-associated depression, although the generally poor quality of studies has obviously contributed to this situation (Hackett et al., 2005).

Although in this study the mean score in any of the Anxiety, Social function and depression domains as well as total GHQ score were higher among stroke patients but in none of these domains even the depression domain the percentage of patients with abnormal score was not statistically higher among stroke patients compared to control group.

This means that if findings of our study are confirmed by other larger scale studies, It can be concluded that although stroke affects general health status and increases chance of depression but it may not be more than the effect of some other diseases or hospitalization itself.