The present study revealed that 12.2% (95% CI 10.8% - 13.9%) of students smoked cigarettes (4.4% among females, 20.6% among males) and the current smoking practice is 4.2%. This seems to be similar to findings in Ethiopia and elsewhere. In Ethiopia, a study conducted in Addis Ababa for the Global Youth Tobacco Survey the prevalence reported that 10.1% of students had ever smoked cigarettes (boys, 15.2%; girls, 5.7%) where as 2.9% currently smoke cigarettes (boys 4.4%; girls, 1.0%). About 13.6% of never smokers were likely to initiate smoking in the next year . In the same study, the overall prevalence of ‘ever users’ of tobacco products was 13.9%, 20.5% among boys and 2.9% among girls. It was estimated in a previous survey among high school students in northwest Ethiopia that the lifetime prevalence of smoking was 13.1% and the current prevalence rate of smoking was 8.1% . A study among junior collegiate students in Nepal in 2007, reported that the prevalence of ‘current use’ of tobacco products was 10.2% . A study of young people aged 15 to 24 years in Addis Ababa showed a current smoking prevalence of 11.8% in males and 1.1% in females . In a study of rural populations in southern Ethiopia in 2005, the percentage of ever-smokers among respondents aged 15 years or more was 5.8% for ever smoking and 4.4% for current smoking .
In this study, one of the predictors of cigarette smoking was sex whereby more percentage of males smoked cigarettes than females. According to a study in Jakarta (Indonesia), Guangdong (China) and Nepal, male predominance was reported in the habit of smoking, while in Zambia , in the Indian cities of Delhi and Goa, and in the Czech Republic, no gender differences have been observed . Smoking is found to be strongly associated with male sex in almost all populations in studies conducted in Africa [17, 20–24]. This may be because females are more socially restricted than male counter parts and mostly young people imitate and exercise what they observe from their elders, parents and friends. Furthermore, familial relationships including care and family related activities may protect females from involving in tobacco use .
In this study, we detected a significant positive association between smoking and age (OR 1.20). Similarly, previous studies elsewhere have reported an association between age and cigarette smoking, where as age increases the odds of smoking also tend to increase [17, 24, 26–28]. This might be because among adolescents, an increase in age might lead to an increased trial of risky practices and experiment with substances such as tobacco. The subjective reason given for ever cigarette smoking in this study are also in agreement with this conjecture where 42.9% reported to smoke for “trial” and 52.6% for “enjoyment” among others.
We found that having friends who smoked cigarettes had an eight fold increase in the odds of smoking compared to non-smoking friends (OR 8.14). Several studies have previously reported the association between adolescent smoking and having a friend who smoke. This is consistent with studies conducted in parts of Ethiopia like Gondar , Addis Ababa and Butajira  as well as elsewhere . For instance the GYTS study among adolescents in Addis Ababa revealed that among participants most or all of whose friends are smokers, there was a more than 30-fold increase in the odds of smoking compared to those who had no smoking friends, while those who had some smoking friends had a 9-fold increase in the odds of smoking . The report that current smoking is associated with having friends that smoke could indicate a role for peer pressure or that risk taking students who would like to try smoking might be befriending smokers. For instance a study by Bricker and colleagues  reported that smoking among adolescents’ friends could influence both the initiation of smoking and its maintenance. This is affirmed by the report that most students obtained their first cigarette from a friend, indicating that initiation of smoking is linked to social relationships . Furthermore diverse psychosocial factors have been associated with the use of cigarettes by adolescents which include having smoking parents , teachers  and peer pressure [27, 33]. Also leisure activities, especially the ones that involve the company of friends [26, 34] were associated with higher odds of smoking.
Strengths and limitations of the study
Our survey is not representative of all adolescents or age groups, and as it is conducted among students it might not necessarily be generalizable to adolescents not enrolled in secondary- or high schools. Secondly, the survey applies only to adolescences who attended classes on the day of the survey and who completed the GYTS questionnaires. Hence delinquent students might be different from the population studied as they are expected to have higher levels of substance use. However, as the response rate was high (more than 90%) among the students present at the time of the survey, bias attributable to non-response among the population studied are expected to be limited. Furthermore, since smoking status was not validated by using biomarkers such as nicotine levels in saliva or exhaled carbon monoxide, it is difficult to estimate the actual extent of under or over reporting of smoking. Furthermore, there may be recall biases as students may not recall whether or not they smoked within the past 30 days prior to the day of the study. Despite the above limitations, the study has strengthens. We employed a standardized questionnaire that enabled us to compare to other studies conducted in similar settings. The prevalence estimates obtained are likely to closely represent the smoking prevalence among school going adolescents as we employed probability methods for selecting the sample.