TB management in refugees and prisoners: Botswana's experience

Two studies presented at the 1st National HIV and AIDS Conference (NHASORC) last month in Gaborone, Botswana suggest the TB epidemic there could be even worse among two "congregated" populations: refugees and prisoners. In fact, one of the studies reported a two-month incidence of TB in prisoners in Botswana that was 18 times higher and a rate in prison guards that was 7 times higher than that of the general population.

In the first study, Dr. C Bakgethisi and colleagues investigated TB in the Dukwe refugee settlement in Botswana. The rates of TB among refugees are often greater than the surrounding populations. Common problems for refugees such as malnutrition, crowding, poor access to care, and other underlying illness put them at risk of TB. Additionally, the mobility of refugees creates challenges for ensuring adequate treatment follow up. However, said Dr. Bakgethisi, "TB control programmes can succeed in refugee camps once basic needs such as water, food, sanitation, shelters are met and essential clinical services are functioning."

Dukwe was established in 1978. The camp accommodated about 4600 refugees in 2002 at the time of screening. Many refugees spent several years at Dukwe while awaiting voluntary repatriation or resettlement in other countries or Botswana citizenship. The camp is open, thus refugees frequently travel to surrounding communities for employment and leisure. Dukwe has a clinic staffed by a physician, six nurses and a number of family welfare educators. Refugees receive food, clothing, shelter, and can join schools and training programmes offered by the United Nations.

The study attempted to determine the prevalence of smear positive TB in refugees and the camp staff and to identify risk factors for TB in these populations. Anyone who reported a cough lasting two weeks or longer was asked to produce a three sputum samples: one at presentation, followed by an early morning specimen and another spot specimen the following day. TB was diagnosed using smear microscopy performed on the site. Mycobacterial cultures were requested on selected smear negative cases where there was a high index of suspicion and sent to the national TB reference laboratory. Persons unable to provide a specimen but with persistent symptoms underwent chest radiography.

Among 2365 resident refugees, 1615 refugees could be located and screened. 256 out of 1615 (16 %) refugees reported cough for two weeks or more. Of these, 205/256 (80 %) produced sputum. 15 refugees underwent chest radiography; seven (47 %) had abnormalities, though none was consistent with TB. 45 specimens were sent for mycobacterial culture but none was positive for MTB. Five refugees were on TB treatment at the time of screening. Four new smear positive cases were identified. The TB prevalence was 0.6 %. Risk factors for TB included contact with a TB case (Relative Risk [RR] 6.2 ; 95 % Confidence Interval [CI] 1.1, 33.7); and reporting any alcohol use [RR] 5.2; 95 % [CI] 1.4, 19.2).

Among 110 staff, 41 (37%) were screened. Three out of 41 (7 %) reported cough for two weeks or more. Three staff members were already on treatment but no new case was identified. The TB prevalence among the staff was 7.3%.

TB among refugees and camp staff was common, but consistent with the high burden of TB in Botswana. Nevertheless, Dr. Bakgethisi noted that the TB prevalence was a minimum estimate, chiefly because population screening was incomplete. "TB screening of the entire refugee population was difficult despite extensive efforts."

The doctor believes that targeted screening of high-risk groups (HIV-infected, contacts of smear-positive cases, and persons reporting alcohol use) might identify new cases sooner and reduce ongoing transmission of TB.

The crisis in the prison population seems even worse, according to the results of a project by M. Notha and colleagues to screen prisoners and guards for tuberculosis in Botswana.

A few studies have been performed in prisons in different African countries. These have found rates of TB in prisons that are five to ten times local/national rates. Dr Notha’s study was the first to determine the prevalence of TB among Batswana prisoners and prison guards, and to investigate the risk factors for TB in prisoners. Screening was offered at 4 prisons in the prison system in Gaborone, Botswana during 2002. All prisoners and guards at least 16 years of age were eligible for screening. Those who agreed to participate were administered a questionnaire in English or Setswana for demographic, treatment history and symptom information.

Three sputum samples were requested from any prisoners or guards reporting cough for smear microscopy and culturing. Chest radiographs were obtained for anyone with a cough who was unable to produce sputum samples.

1027 prisoners (or 88% of the prison population) were screened, including 20 who were on treatment at the time the screening began.

96% of the prisoners were male. The median age was 26 years with a range from 16 to 78 years of age. 83% of prisoners were incarcerated for the first time. The median duration of incarceration was 15 months with a wide range from 1 day to 22 years.

Out of 1027 prisoners, 509, or nearly half, reported a cough. Sputum was obtained on 371 (73%). 33 chest radiographs were obtained on those with a cough who were unable to produce sputum. None was counted as a clinical case.

A total of 41 cases, or 4% of prisoners, were identified with TB: 19 prisoners were identified during screening, including 8 who were smear-positive and 11 were smear-negative but culture-positive. 20 were already on TB treatment at the time the screening began. An additional four cases that were initially smear- and culture-negative but had persistent symptoms were found to be smear-positive on repeat testing after the initial screening period. These cases were included in the analysis of risk factors, but were not included in the calculation of point prevalence. Based on 39 prevalent cases identified or on treatment over a two-month period, the point prevalence of TB in prisoners was found to be nearly 3.8% (3797 cases/100,000 population). Using only the newly identified cases, the two-month incidence was 1850 cases/100,000 population.

After controlling for a range of variables, three factors remained significant in the multivariate model. Incarceration for more than six months was associated with a more than five-fold increased risk of TB when compared with those incarcerated for six months or less. Being a first time offender in residence in one particular prison was associated with nearly four times the increased risk of TB compared with residence in one of the other large facilities. Smoking was significantly associated with a decreased risk of TB in this analysis. Previous studies have found that smoking is a risk factor for TB. Dr. Notha said, "our finding may simply reflect the fact that prisoners who smoked were more likely to report cough and be screened with sputum, but less likely to have TB. When we restricted the multivariate model only to prisoners reporting cough, smoking was no longer significant."

Results for prison guards were similar to those of prisoners. 263 out of 288, (91%) of prison guards were screened. Only 45, or 17% reported any cough, and sputum was obtained on 25 (56%) of those with cough. A total of seven cases of TB were identified, five were on treatment at the time of screening. Two new cases, both smear-negative but culture-positive, were identified through screening. No new cases were identified through chest radiography.

"This project found critically high rates of TB among prisoners and guards in Botswana. The two-month incidence of 1850 among prisoners was nearly 18 times the two-month incidence in the general population. The two-month incidence among guards was more than 7 times the two-month incidence in the general population.

"Longer incarceration has been associated with a risk of TB in other prison studies and suggests that transmission within prison settings is causing high rates of TB.

"Several recommendations follow from these findings," concluded Dr. Notha. "First, a programme to screen for TB should be initiated. This should include prisoners at prison entry or transfer. However, since TB was associated with the duration of incarceration, screening at entry alone may be insufficient and screening on a periodic or ongoing basis should also be considered. Prison guards should also be screened on a periodic basis. Contacts such as cellmates of newly identified smear-positive cases should be evaluated for active disease. Finally, a formal assessment of administrative controls and environmental measures to reduce TB transmission is planned for February, 2004."

References

Bakgethisi C et al. Tuberculosis screening at a refugees settlement camp in Botswana, 2002 ?2003. First National HIV/AIDS/STI/Other Related Infectious Diseases Research Conference, Gaborone, Botswana, abstract MB11-7, 2003.

Notha, M. Rapid tuberculosis assessment in a prison system—Gaborone, 2002. First National HIV/AIDS/STI/Other Related Infectious Diseases Research Conference, Gaborone, Botswana, abstract MB11-8, 2003.