1. Introduction

Stroke is the second leading cause of death [1] and the third leading cause of disability-adjusted life years worldwide [2]. In fact, 5.9 million deaths and 102.2 million disability-adjusted life years, lost throughout the world, were attributable to stroke in 2010 [3,4].

However, classification of causes of death and disability differs between high and low to middle-income countries (HICs and LMICs, respectively) [5], and changes dramatically over decades [1]. Studies have shown that stroke incidence in HICs has been always elevated compared with LMICs [6] until the past few decades, where incidence in LMICs has exceeded that of HICs by 20% [6,7]. These trends in incidence rates are accompanied by similar trends in mortality rates [6]. The expected effect of these changes, from 1990 to 2020, is an increase in stroke prevalence by 120–137% in LMICs compared to an increase of 30–60% in HICs [8].

Incidence and mortality rates of stroke are mainly affected by the prevalence of behavioral cardiovascular risk factors [9] and by the advances in its early diagnosis and management [8]. However, these factors are influenced by socioeconomic considerations (education, income, occupation) [10], which explains the difference in the epidemiology of stroke between countries of different incomes.

The Arab world is a region of 22 countries with wide socioeconomic disparities and unequal distribution of wealth, science, and technology [11,12], where trends of stroke are not precisely established [13]. However, many estimates about noncommunicable diseases (NCDs) exist mainly from studies conducted in HICs, such as Bahrain, Kuwait, and Saudi Arabia [13,14]. In 2008, 60% of all deaths in the region were attributable to NCDs, with large variations between countries: 27% in Somalia, an LIC, to 84% in Oman, an HIC [13]. Despite this, stroke is classified among the top five causes of death in all countries [12]. Furthermore, we find a high prevalence of physical inactivity (may reach 90% among women), smoking (38% in Lebanon), obesity (40% in the Arab Gulf countries), and diabetes (20% in the Arab Gulf countries) among Arab populations [15–18].

Lebanon is an Arab country of the Middle East region classified among LMICs, based on the 2008–2012 World Bank’s country classification and the World Health Organization (WHO) classification, where available studies on NCDs have shown a high prevalence of cardiovascular risk factors (obesity 26% [19], diabetes 14% [20], and hypertension 36% [21]). Moreover, stroke is assumed to be the second leading cause of death in the country.

However, studies focusing on overall stroke prevalence or incidence rates are scarce among all Arab countries [14]. Thus, we conducted a pilot study to assess the prevalence of stroke survivors in Lebanon and provide a reliable set of data on stroke epidemiology for subsequent studies and analyses.

2. Materials and methods

We performed a cross-sectional study between January and March 2012 to detect persons living with a history of stroke (stroke survivors) among Lebanese inhabitants. We adopted a multistage sampling. First, we calculated the sample size needed to obtain a 95% confidence level. For this purpose, community-based studies from Arab countries were examined to fix the minimum expected stroke prevalence (0.2%; CI width = 0.2%) [22,23]. A total of 7600 participants were needed. Then, we examined the 2005 distribution of the Lebanese population among governorates and districts to establish our sample in a more or less proportionate way (see Table 1). A list of landline phone numbers of all households, provided by the government, was obtained to randomly select the households’ sample among each governorate. Then, all members of the household were included in the study. We assumed that on average, four persons are found in each household. Thus, a total of 2000 calls were performed, allowing for a 10% nonresponse rate, which later resulted in 1779 answers more or less adequately distributed on all governorates.

After including all members of the household in the study, we obtained a total sample of 6963 participants. The interviewer explained the study objectives to the respondent, who was kept anonymous. After giving his oral consent, the respondent was asked a number of questions within 10 min. The questionnaire had two major sections; one related to sociodemographics, and one related to stroke identification. Each respondent was first asked about the number, sex, and age of inhabitants in his household and the number of rooms in the household (to later calculate the crowding index which is the number of persons per room, usually related to the socioeconomic status) [19].

Stroke cases were then identified with the question that raises the existence, among the house inhabitants, of a person having a history of “Jalta Dimaghiya” diagnosed by a physician (we asked ‘Have you ever been told by a physician that you suffered a stroke?’). “Jalta Dimaghiya” is the Arabic synonym of stroke and the most familiar and specific term for the disease in Lebanon. Then, we followed the WHO criteria for the definition of stroke to eliminate false positive responses: stroke is defined as “rapidly developing clinical signs of focal (or global) disturbance of cerebral functions, lasting more than 24 h or leading to death (not in our study), with no apparent cause other than that of vascular origin” [24]. For that, when a case was found, the interviewer asked for the acute symptoms that occurred and led to the ultimate diagnosis of stroke by a physician, especially face, arms, and speech signs. At the end, the list of medications administered for the stroke patient was requested to further confirm the case.

Interviews were done in the presence of two interviewers (a pharmacist and a nurse) to better ensure the quality of the interview. Data entry was performed by two persons unaware of the study objectives. A total of 50 questionnaires were checked for errors by a third person and the error rate was <1%; thus, the entry was considered adequate.

Statistical analysis was performed using SPSS software, V. 18.0. A weighting was performed according to the number of inhabitants by age group and sex in each governorate, as described by the official 2005 estimates (local age-standardization), to obtain representative results of the Lebanese population. We also performed an age standardization to the WHO standard world population estimates for 2000–2025 [25] to allow comparisons with other countries.

Stroke prevalence was calculated representing the number of stroke survivors at the time of data collection over the total number of the sample. Estimates of stroke prevalence were also calculated for each governorate for comparative purposes. Proportions were compared using Pearson’s Chi-square test, with further standardized residuals’ two-by-two comparisons. Means were compared using the analysis of variance (ANOVA) test with Bonferroni correction on post hoc tests. A p value <0.05 was deemed statistically significant.

The study was conducted for observational purposes with no participants’ traceability. Thus, no ethical approval was found to be necessary by the Lebanese University ethics committee.

3. Results

Among the 6963 inhabitants included in the study (mean age 36.8 ± 21.2 years; 48.6% males), 56 persons had suffered a stroke, which provided a crude prevalence of 0.80% (95% CI = 0.60–1.00%). Adjusted prevalence decreased to 0.50% (95% CI = 0.33–0.66%) and 0.60% (95% CI = 0.42–0.78%) after standardization to the local and world populations, respectively. Stroke patients had a mean age of 63.2 ± 15.8 years and 21 (60%) were males. Among those aged >40 years, adjusted stroke prevalence was 1.83% (95% CI = 1.20–2.46%).

Most commonly reported stroke symptoms were sudden headache (40%), facial weakness (40%), and arm weakness (37%).

No significant difference was found for stroke prevalence between males and females (p = 0.21). Significant differences were found among age groups and governorates, stroke prevalence being significantly lower in “North Lebanon” [0.13% (95% CI = 0.05–0.31%) vs. 0.61% (95% CI = 0.40–0.82%) in all other governorates; p = 0.021]. Moreover, stroke prevalence was significantly higher among less crowded households (crowding index ⩽1 vs. >1, 0.64% vs. 0.28%; p = 0.025). When examining “North Lebanon” characteristics, population mean age was significantly lower compared to all other governorates (26.2 ± 19.7 vs. 29.5 ± 20.3 years; p = 0.000). Furthermore, the mean crowding index was significantly higher in “North Lebanon” vs. other governorates (1.7 ± 0.9 vs. 1.2 ± 0.6; p = 0.000).

More details among subgroups are provided in Table 2 .

The age- and sex-specific prevalence estimates of stroke in the 10-year age groups are shown in Table 3 . In general, the prevalence increased with age for both males and females with a prevalence reaching 9.38% (95% CI = 2.24–16.52%) in the group aged 80 years and older. No stroke cases were found among males aged <50 years whereas six female cases were found (p < 0.05). There was also a significant sex difference in the age group (60–69) years, where stroke prevalence was higher among males than females (3.53 vs. 0.56%; p = 0.039).

4. Discussion

According to our findings, age-adjusted prevalence of stroke survivors among a sample of Lebanese inhabitants on March 31, 2012 was estimated to be 0.50% (95% CI = 0.33–0.66%) and 0.60% (95% CI = 0.42–0.78%) for the local population and the WHO world population standardizations, respectively. However, epidemiological inference to the Lebanese population may be difficult to obtain in our study settings, although results were age- and sex-standardized, due to our data collection method and the relatively small sample size. In fact, our sample was chosen among Lebanese inhabitants that possess a landline phone number. This group may be of a relatively higher socioeconomic status or include more aged participants. Moreover, telephone interviews may have slightly underestimated the real prevalence in the country. However, many studies have shown that telephone interviews have a high validity and reliability when compared to face to face interviews [26,27] and that there is a moderate to substantial agreement between self-reported questionnaire data and clinical examinations for chronic diseases [28]. Hence, the interview survey in our study may be considered more or less valid and may have some advantages over other collection methods such as higher response rates and lower cost of study materials.

Thus, our findings may be valuable to draw hypotheses on stroke burden in the country for future analyses. They may also add valuable epidemiological data on stroke burden in the Arab and Middle East countries where studies are found to be scarce [18].

Studies on stroke prevalence are relatively limited compared to studies on stroke incidence and mortality rates. However, available data show that high prevalence estimates are found mainly in developed countries where a history of stroke is present among >1% of the general population [29–32]. The lowest prevalence is found in Africa, South Asia, and Arab countries, although two studies from Egypt showed prevalence estimates as high as the one in developed countries (0.98% and 1.22%; age-adjusted to WHO world populations) [33,34]. Our study prevalence is two to five times lower than that of developed countries [29–32], but higher than many other studies from developing countries [35–37]. However, it could be compared to South American and South Asian countries [38,39]. For example, one study from India (Kolkata) has given an age-standardized prevalence to world standard population of 0.54% [35], very similar to our result. Nevertheless, comparisons between countries should be interpreted with caution due to differences in study design, stroke definition, and time of data collection.

A review by Benamer and Grosset [17] on stroke epidemiology in Arab countries shows a prevalence of 0.04–0.07% in 10 Arab countries, a percentage 10 times lower than that of our study. However, most studies were hospital-based and long-established, which makes comparison slightly difficult. Nevertheless, there were two community-based studies that also showed lower prevalences: in Tunisia (Kelibia), age-adjusted prevalence according to the WHO world population was 0.07% [23], whereas stroke crude prevalence was 0.2% in Saudi Arabia (Thugbah community, Eastern Province) [22]. However, most studies from Arab countries were restricted to certain regions or governorates and may not be representative of the whole country.

Stroke incidence rates are reported to be increasing in Arab countries [4] and with the high prevalence of cardiovascular risk factors, we expect a higher prevalence of stroke at the present time. However, population aging in Lebanon is higher than in any other Arab country and life expectancy has reached an average of 80 years and more [40,41]. Moreover, Lebanon has the highest smoking rates [13] and current smokers constitute almost one third of its adult population [13,42]. Other risk factors, especially alcohol consumption (one fifth of the Lebanese population [42]), are also highly prevalent. In these settings, we expect that stroke prevalence would be still higher in Lebanon even if some other risk factors are higher in other Arab countries, especially the Arabian Gulf oil-rich countries (a review by Rahim et al. [13] shows that obesity and sedentary lifestyle are more or less higher in Arab countries compared to Lebanon). On another side, Lebanon has the highest number of physicians per inhabitants among all Arab countries, which with its advances in the medical and technological fields may have reduced stroke mortality in recent years [11]. However, these hypotheses should be tested in future analyses.

Two studies have reported stroke prevalence in Lebanon: the study by Jurjus et al. [43] and the study by Farah et al. [44]. However, the first one was not a community-based study, although projections to the general population were made. It was based on physicians’ reported data on stroke (3.9%) and prevalence of transient ischemic attacks (1.3%) among patients aged 30 years and older. The second study reported an adjusted prevalence of 3.6% (95% CI = 2.3–4.9%) of a self-reported stroke history among people aged 40 years and older (note that the confidence intervals of Farah et al. [44] and our study (95% CI = 1.20–2.46%) for the age group >40 years slightly overlap).

For age influence on stroke, our results were similar to many other studies where stroke prevalence clearly increases with age [33,45]. Moreover, our study resulted in a higher stroke prevalence among men than women (ratio 1.5:1) with a significant difference for those aged 60–70 years. Similar results for males’ predominance were found in many countries, such as Egypt [33] and Australia [46]. Studies from Kuwait [47] and South Asia [48] have, however, led to contrary results. Moreover, all cases found in the sample group aged <50 years were females, a finding that may be correlated to the higher incidence of subarachnoid hemorrhage among relatively younger females [49].

Another major finding is discrepancies between governorates. In fact, a higher stroke prevalence in our sample is found in Beirut and other governorates compared to North Lebanon, a governorate more or less distant from the capital. This may reflect a lower incidence of stroke or a higher mortality rate in North Lebanon. When examining stroke prevalence among different crowding index groups, we found a lower prevalence in more crowded households which are more frequent in North Lebanon. Thus, we might associate the socioeconomic status correlate “crowding index” to the lower prevalence found in North Lebanon in our sample. Moreover, a significantly younger population was found in North Lebanon compared to other governorates, which might indicate lower incidence rates.

Our findings match with the real socioeconomic situation of the country as available national data show that poverty is not equally distributed in the country, the North being the most deprived [50].

5. Conclusions

Although the study results could not be epidemiologically inferenced to the Lebanese population, they shed lights on some important findings. Stroke prevalence might be high in Lebanon, especially among older groups. However, the relatively high prevalence might be associated to a high incidence of stroke in the country much more than a reduced mortality rate. Moreover, it is clear from existing data that some governorates are less privileged than others, and stroke prevalence differences found in the present study, among governorates, may also reflect this situation and trigger further attention on stroke mortality in less privileged areas.

Thus, it is urgent to implement further studies to understand the disease burden in Lebanon, and establish strong control procedures on cardiovascular risk factors (smoking), especially with further demographic transition and urbanization of the population.

Conflicts of interest

All authors have no conflicts of interest to declare.