Prevalence of tuberculosis, HIV, and TB-HIV co-infection among pulmonary tuberculosis suspects in a predominantly pastoralist area, northeast Ethiopia

Background TB-HIV co-infection is one of the biggest public health challenges in sub-Saharan Africa. Although there is a wealth of information on TB-HIV co-infection among settled populations in Africa and elsewhere, to our knowledge, there are no published reports on TB-HIV co-infection from pastoral communities. In this study, we report the prevalence of TB, HIV and TB-HIV co-infection among pulmonary TB suspects in the Afar Regional State of Ethiopia. Design In a cross-sectional study design, 325 pulmonary TB suspects were included from five health facilities. Three sputum samples (spot-morning-spot) were collected from each participant. Sputum samples were examined for the presence of acid fast bacilli using Ziehl–Neelsen staining method, and culture was done on the remaining sputum samples. Participants were interviewed and HIV tested. Results Of the 325 pulmonary TB suspects, 44 (13.5%) were smear positive, and 105 (32.3%) were culture positive. Among smear-positive patients, five were culture negative and, therefore, a total of 110 (33.8%) suspects were bacteriologically confirmed pulmonary TB patients. Out of 287 pulmonary TB suspects who were tested for HIV infection, 82 (28.6%) were HIV positive. A significantly higher proportion of bacteriologically confirmed pulmonary TB patients [40 (40.4%)] were HIV co-infected compared with patients without bacteriological evidence for pulmonary TB [42 (22.3%)]. However, among ethnic Afar pastoralists, HIV infections in smear- and/or culture-negative pulmonary TB suspects [7 (7.6%)] and bacteriologically confirmed pulmonary TB patients [4 (11.8%)] were comparable. On multivariable logistic regression analysis, Afar ethnicity was independently associated with low HIV infection [OR=0.16 (95% CI: 0.07–0.37)], whereas literacy was independently associated with higher HIV infection [OR=2.21 (95% CI: 1.05–4.64)]. Conclusions Although the overall prevalence of TB-HIV co-infection in the current study is high, ethnic Afars had significantly lower HIV infection both in suspects as well as TB patients. The data suggest that the prevalence of HIV infection among Afar pastoralists is probably low. However, population-based prevalence studies are needed to substantiate our findings.

D espite the implementation of a widely adopted strategy to control tuberculosis (TB), the disease remains a major public health problem, particularly in developing countries (1). In 2013, an estimated 9 million individuals developed TB and 1.5 million died from the disease (2). Similarly, the HIV epidemic remains a huge global challenge, and in 2012, an estimated 35.3 million people were living with the virus, whereas 1.6 million died (3). The prevalence of TB-HIV co-infection widely varies; the prevalence of HIV among TB patients ranged from 3.8 to 72.3%, whereas the prevalence of TB among HIV-positive patients ranged from 2.9 to 64.5% (4). The African region accounted for 75% of the estimated number of HIV-positive incident TB cases (2). Although many countries have made considerable progress in addressing the TB-HIV co-morbidity, global targets for HIV testing among TB patients and provision of antiretroviral therapy to those who are HIV positive have not been reached (2). HIV markedly increases the risk of progression to active TB (5)   Likewise, TB is a leading killer among HIV-positive individuals (4); TB also hastens the progression of HIV disease, increasing the risk of developing other opportunistic infections (6). Thus, the influence of each infection on the other's natural history and pathogenesis has enhanced the magnitude of TB/HIV epidemic (7). The number of people dying from HIV-associated TB has been falling since 2003; however, there were still 360,000 deaths from HIV-associated TB across the globe in 2013 (2).
Although several studies reported TB-HIV co-infection in many African countries including Ethiopia (8Á10), to our knowledge, there are no reports on TB-HIV coinfection from pastoralist areas. Pastoralists constitute a substantial proportion of the population of Africa. Ethiopia, having the largest cattle population in Africa (11), is also home to an estimated 10 million pastoralists, about 12% of the total population (12). Pastoralists live in drought prone, arid areas and often face serious food shortages (13) and stress which are important risk factors for TB. Besides, their mobility could be a driving factor for HIV transmission as well as a challenge for intervention efforts. Here, we report for the first time, the prevalence of TB-HIV co-infection among pulmonary TB suspects as well as TB patients in the Afar Region, a predominantly pastoral region in the northeast Ethiopia.

Study area
The study area has been described elsewhere (14). Briefly, the study was conducted in two public (Awash Health Center and Dubti Hospital) health facilities in the Afar Region and three private (Selam hospital, Bati Hospital, and Amir Higher Clinic) health facilities in Dessie Town. The reason for including these private health facilities is because they provide diagnostic services for a substantial number of TB patients coming from the Afar Region. Laboratory work was performed at Aklilu Lemma Institute of Pathobiology TB laboratory, Addis Ababa University.

Study design and study participants
A health institution based cross-sectional study was conducted in the five health facilities from September 2009 to March 2010. Pulmonary TB suspects (]18 years of age) who were residents of the Afar Region during the study period and had a cough lasting for 2 weeks or more were included consecutively. A sample size of 339 was estimated using a prevalence of 33% for bacteriologically confirmed pulmonary TB (15), margin of error of 5% and 95% CI.
Data collection and sputum culture Using a structured, pre-tested questionnaire, patients were interviewed by trained laboratory technicians on basic sociodemographic characteristics and their symptoms. Then, as previously described (16), ZiehlÁNeelsen staining was done on three sputum samples (spot-morning-spot) collected from each participant. The remaining sputum samples were stored at 48C and subsequently transported to Aklilu Lemma Institute of Pathobiology TB laboratory within 1 week. Subsequently, the three sputum samples from each participant were pooled, processed, and cultured according to WHO guidelines (17). Briefly, sputum samples were homogenized and decontaminated with equal volume of 4% NaOH and centrifuged at 3,000 rpm for 15 min. The supernatant was poured off while the sediment was neutralized with 0.1N HCL using phenol red as an indicator. Two to four loopfuls of the sediment were inoculated in four slopes of LowensteinÁJensen medium, and inspection of media was done every week to monitor the growth until 8 weeks. ZiehlÁNeelsen stain was done to detect acid fast bacilli.

HIV test
HIV testing was done using KHB (Shanghi Kehua Bioengineering, Ltd, Shanghai, China) as a screening test, HIV1/2 STAT-PAK † ASSAY (CHEMBIO Diagnostic systems, Inc., Medford, NY, USA) as a confirmatory test, and Uni-Gold TM (Trinity Biotech, Jamestown, NY, USA) as a tie-breaker according to the national algorithm (18). HIV testing was done after counseling, and patients with HIV infection were referred to HIV care and treatment clinics after post test counselling.

Data analysis
Datawere entered into EpiDataversion 3.1 and exported into SPSS for Windows version 20 for further analysis. Univariable analysis followed by multivariable analysis (logistic regression) was done to identify independent predictors of dependent variables. Variable selection was based on prior knowledge from literature and biological plausibility. The outcome variables were HIV infection and bacteriologically confirmed TB. A p-value for a two-tailed test less than 0.05 was considered as statistically significant.

Ethical consideration
The study has been ethically approved by the National Research Ethics Review Committee from Ethiopia and the Regional Committees for Medical Research Ethics Á South East Norway (Regionale Komiteer for medisinsk og helsefaglig forskningsettik, Sør-Øst) from Norway before commencing the study. Written informed consent was obtained from all study participants before the interview and sample collection.

Sociodemographic characteristics
A total of 339 pulmonary TB suspects were included in this study. However, 14 participants were excluded because of poor quality of sputum, and hence 325 (95.9%) were included for analysis. The sociodemographic characteristics of study participants have been summarized in Table 1. About 75% of the study participants were below 45 years of age with a median age of 30 years (interquartile range: 25Á45). Males constituted 58.8% of the study participants. Pastoralists represented 150 (46.2%) of the study participants.
Prevalence of smear-/culture-positive pulmonary tuberculosis and associated factors Among the 325 pulmonary TB suspects, a total of 105 (32.3%) were identified as pulmonary TB patients based on culture results. Among the 105 culture positives, 39 (37.1%) were diagnosed as smear-positive pulmonary TB patients at health facilities. Five more suspects who were reported as smear positive turned out to be culture negative. Therefore, the prevalence of smear-positive pulmonary TB patients among 325 pulmonary TB suspects was 13.5%. Based on bacteriology results (direct smear and culture), a total of 110 (33.8%) were diagnosed as pulmonary TB patients.
Results of analysis of the association between sociodemographic and other selected variables with pulmonary TB are summarized in Table 2. On logistic regression analysis, male sex, age group 31Á45 years, and HIV infection were independently associated with pulmonary TB. The odds of having pulmonary TB among female suspects was 52% [OR 00.48 (95% CI: 0.27Á0.86)] less compared with males, and the age group 31Á45 years had a 56% reduction in pulmonary TB compared with those less than 31 years of age. On the contrary, the odds of having pulmonary TB among HIV-positive pulmonary TB suspects was significantly higher [2.65 (95% CI: 1.41Á4.97)] compared with the corresponding odds among HIVnegative pulmonary TB suspects.

Prevalence of HIV infection and associated factors among study participants
A total of 287 (88.3%) participants agreed to be tested for HIV infection, and 82 (28.6%) were positive. The prevalence was comparable in males (26.4%) and females (31.5%) ( Table 3). However, the peak HIV prevalence in females was at a younger age (18Á24 years) compared with males (35Á44 years) (data not shown). A very low proportion (8.7%) of HIV infection was found among ethnic Afars compared with the proportion of infection among other ethnic groups ( Table 3).
The association of HIV infection with sociodemographic variables was assessed using logistic regression analysis, and results are summarized in Table 3. On multivariable logistic regression, the proportion of HIV infection in TB suspects was significantly lower among ethnic Afars [OR 00. 16

Discussion
The HIV epidemic has influenced the epidemiology of TB significantly. HIV is a well-known risk factor for progression to active TB among those infected with Mycobacterium tuberculosis (7). Studies reported the prevalence of HIV infection among TB patients as well as TB suspects in settled populations as reviewed in Gao et al. (4). However, to our knowledge, this is the first report on TB-HIV co-infection and the prevalence of TB and HIV among pulmonary TB suspects from a predominantly pastoral community.
In this study, among pulmonary TB suspects, 13.5% were diagnosed as smear-positive pulmonary TB patients in agreement with a previous report from a rural hospital (19), but a higher prevalence of smear-positive pulmonary TB was observed in Addis Ababa (20). The prevalence of pulmonary TB among pulmonary TB suspects based on smear microscopy and culture was found to be 33.8% in agreement with a previous finding in a tertiary hospital in Addis Ababa (15). Such a high TB prevalence among suspects might be related to delay in the diagnosis of TB as previously reported in the study area (21). Gender disparities with regard to TB have been documented, although the reason for such differences is not well known. In agreement with a previous report from Ethiopia (15), a significantly higher proportion of male pulmonary TB suspects were diagnosed with TB compared with females suspected of pulmonary TB. As reviewed by Holmes et al. (22), gender disparity in TB notification has been reported mainly in high TB burden settings.
The proportion of HIV infection among pulmonary TB suspects was 28.6% which is in agreement with a previous report from Addis Ababa (20). Interestingly, ethnic Afar pulmonary TB suspects had the lowest proportion of HIV infection compared with other ethnic groups, and such low prevalence has been previously documented among pregnant Afar women in the study area (23). Generally, the prevalence of HIV infection among TB suspects is believed to reflect the HIV prevalence in the general population, and our data may suggest a low HIV infection among ethnic Afar population. The relatively low HIV prevalence in ethnic Afars might be related to their geographical and cultural inaccessibility to the rest of the country. However, changes in the life style of ethnic Afars with a tendency of settlement as well as urbanization could facilitate the spread of HIV infection among this population. In this study, the proportion of HIV positivity among pulmonary TB patients (40.4%) was comparable with a report from Ethiopia (9). However, HIV co-infection among TB patients in our study was higher compared with the co-infection rate reported in urban (20.2%) (24) as well as rural (18% and 19%) Ethiopia (25,26). The average TB-HIV co-infection for Africa (32%) (7) is also less compared with the co-infection rate in the current study.
Although low compared with non-Afars, the prevalence of TB-HIV co-infection among indigenous Afar population with traditional culture is still high and worrisome for many reasons. First, TB-HIV co-infection is usually associated with urban poor rather than inaccessible remote regions with traditional ways of life such as cattle herding. Second, pastoralists have less access to education, health services, communication, and transportation and, therefore, poor knowledge and practices regarding prevention of HIV infection (27). Third, pastoralists are migratory people who move with their animals from season to season in search of water and pasture, which implies that the current TB control strategy (directly observed treatment, short course at fixed health facilities), antiretroviral therapy, and counseling and testing for controlling HIV infection, which are primarily designed for settled population, are not easily available to them as documented in other mobile populations (28). When such health services are available, they are easily interrupted because of seasonal migration with their animals, suggesting the need for designing a suitable TB/HIV control strategy pertinent to pastoral way of life to contain the spread of HIV infection (29).
The World Health Organization recommends integrated TB and HIV activities on the interventions needed to prevent, diagnose, and treat TB in people living with HIV. These include establishing and strengthening coordination mechanisms for delivering integrated TB and HIV services (1). However, pastoral communities lack basic health services and depend on traditional healers rather than health services provided by health authorities, which is usually meager and ineffective. Therefore, policy makers need to consider how new interventions can be adopted by TB and HIV control programs, and implemented on a sustainable basis in the pastoral communities to limit the spread of HIV infection. The low prevalence of HIV among ethnic Afar pulmonary TB suspects and pulmonary TB patients may not necessarily reflect HIV prevalence among ethnic Afars, and therefore, a community-based study is needed to determine the population prevalence. Besides, not all health facilities were included in this study, and, hence, our findings may not be generalizable to all ethnic Afar pulmonary TB suspects and pulmonary TB patients in the region.

Conclusions
Although the overall prevalence of TB-HIV co-infection in the current study is high, ethnic Afars had significantly lower HIV infection both in suspects as well as TB patients. These data suggest that the prevalence of HIV among Afar pastoralists is probably low. However, population-based prevalence studies are needed in designing intervention strategies to limit the spread of HIV infection among Afar pastoralists.