Globally and in India, tuberculosis (TB) is the most common opportunistic infection and the most common cause of death among HIV-infected individuals.1–4 Unless HIV-infected TB patients are initiated on cotrimoxazole preventive therapy and antiretroviral therapy (ART) in addition to TB treatment, mortality may remain high.5 The first step in this process is early identification of HIV-infected persons. Globally, less than half of those living with HIV know their HIV status.6 To facilitate early HIV case finding, World Health Organization and International Standards for TB Care recommend provider-initiated HIV testing and counseling (PITC) for people suspected and/or diagnosed to be having TB.7,8 India has adopted the policy of PITC for TB patients.9 The current policy of offering HIV testing only to TB cases may limit opportunities of early HIV diagnosis and treatment. Moving HIV testing upstream to include all persons suspected of having TB and referred for diagnostic evaluation (henceforth referred to as ‘TB suspects’), though recommended internationally, is not a policy in India and HIV prevalence in this population has never been reported. Previous studies on this topic are mostly from African countries where HIV epidemic is generalized; the findings from these studies cannot be generalized to Indian situation with a concentrated HIV epidemic.10,11 Indian policy makers seek evidence of the value of HIV testing of TB suspects for HIV case finding.12
We conducted this study in Mandya district of Karnataka, one of the highest HIV prevalent states in India, to assess the prevalence of HIV among persons referred to any microscopy center in the district for diagnostic evaluation. The specific objectives were (1) to determine the HIV prevalence among TB suspects examined for diagnostic smear microscopy, (2) to assess the number of HIV cases “newly” diagnosed by testing all TB suspects, (3) to assess the number needed to screen (NNS) to find an additional case of HIV-infected person among “TB suspects” in comparison to “TB patients”, and (4) to assess the potential contribution of this activity to HIV case finding.
This was a cross-sectional study done by offering HIV counseling and testing to TB suspects attending the microscopy centres for TB diagnosis.
The study was done in Mandya district (population 2 million) of Karnataka state, India. In the district, TB control program services are available through a decentralized network of primary health care facilities, which provide general health services including diagnosis and treatment for TB. Any patient with a cough of 2 weeks or more is considered as pulmonary TB suspect and referred to the microscopy centre for diagnostic evaluation. The sputum smear microscopy is done in 25 designated microscopy centres of the district, which are Revised National Tuberculosis Control Program-supported microscopy services embedded into the public health system clinics. Microscopy is supported by a district-wide system of external quality assurance. All TB patients diagnosed are treated with standardized fully intermittent thrice weekly short-course regimens (6–9 months) administered under direct observation and are registered at 1 of the 4 subdistrict level TB program management units according to Indian TB program guidelines.13
HIV counseling and testing services are offered through the National AIDS Control Program–supported network of integrated counseling and testing centres (ICTC). There are about 24 ICTCs throughout the district, and these are colocated in the same health facilities as designated microscopy centers. These service delivery sites under National AIDS Control Program follow the national guidelines for counseling, testing, care, and treatment of HIV-infected patients.14
The current national policy is to offer routinely HIV counseling and testing to TB patients; those identified as HIV infected are provided cotrimoxazole prophylaxis and referred to the nearest ART centre for assessment of ART eligibility and initiation on ART—if found to be eligible. Free ART is provided through the district ART centre, where HIV-infected patients (including TB patients) are screened for ART eligibility and offered treatment and care. In the year 2010, 20,165 TB suspects were examined for sputum microscopy and 2156 TB patients were registered in Mandya district. Of the TB patients registered, 74% were tested for HIV and 237 HIV-infected TB patients were identified.15
Sample Size and Sampling
Assuming 3% HIV seroprevalence (P) among TB suspects, an error margin (d) of 1%, 95% confidence limit (z = 1.96), and a possible attrition rate of 10%, the sample size was calculated using the formula
as 1280. Considering the TB suspect examination rate of 2010, it was decided to examine all the TB suspects attending the microscopy centres in 1 month.
The study population consisted of all the TB suspects attending the 25 microscopy centres for diagnostic sputum smear examination in the month of December 2010. TB suspects younger than 18 years were excluded from the study.
Data Collection, Study Variables, and Study Instrument
All the TB suspects aged older than 18 years were provided a patient information sheet by the laboratory technician at the microscopy centre and enrolled into the study. After obtaining a written informed consent, they were further referred to the ICTC for HIV testing located mostly in the same premises. From one microscopy center, which did not have a HIV testing centre in its premises, the patients were referred to the nearby ICTC. At the ICTC, the prevailing national guidelines for HIV testing were followed which includes a pretest counseling, posttest counseling and referral to ART centre in case the patient is found to be HIV positive. TB suspects with an already known HIV status were not tested again as per the national protocol. “Known” HIV status referred to those who were HIV-positive result or with a HIV-negative result within the past 6 months. Those TB suspects who had a negative sputum smear result were tracked for 1 month to assess if they were diagnosed as smear-negative pulmonary TB or extrapulmonary TB. The results of sputum smear microscopy and HIV testing were informed to the client and offered TB treatment and HIV care and support accordingly. The data were collected by the laboratory technician into a pretested structured questionnaire for all the eligible TB suspects; data collected included laboratory number, age, sex, sputum smear result, HIV status, and type of TB. The completed questionnaires were collected once a week by the trained laboratory supervisor of the national TB program and were checked for completeness and validated by the principal investigator. In addition, the data on total clients tested and diagnosed as HIV was collected from the local program division to assess the potential of this strategy to improve HIV case finding.
Data Entry and Analysis
Data were double entered into an EpiData software package16 by 2 data entry operators independently, databases were compared and discrepancies resolved through referral to the original questionnaire. Proportions of HIV-positive patients among the TB suspects and TB patients with 95% confidence intervals were calculated. The proportion of all HIV-infected persons who were “newly diagnosed” (after excluding previously known HIV cases) was calculated. To find 1 additional HIV case, the NNS among TB suspects and TB patients was determined. All variables have been described as proportions, and differences between groups were compared for statistical significance using the χ2 test or Fisher exact test, as applicable. The P value less than 0.05 was considered statistically significant.
All study subjects provided written informed consent for study participation and HIV testing. The study protocol was reviewed and approved by the Ethics committee of National Tuberculosis Institute, India, and the Ethics Advisory Group of the International union against TB and lung disease (The Union), France.
Of the 1723 TB suspects who underwent diagnostic sputum smear microscopy, 1668 TB suspects were eligible to participate in the study. Of those eligible, 1539 (92%) participated and had HIV status ascertained and recorded (Fig. 1).
The demographic and clinical characteristics of the 1539 who participated and had a HIV test result and the 129 who declined to ascertain HIV status are shown in Table 1. TB suspects who did not have a HIV test result were more likely to belong to older age groups. Among 1539 TB suspects with a known HIV status, 100 (6.5%) were eventually diagnosed as TB. Of the TB cases, 87 were sputum smear positive, 8 were sputum smear negative, and 5 were diagnosed as extrapulmonary TB.
HIV prevalence among TB suspects disaggregated by age, sex, and TB status is shown in Table 2. Overall, HIV prevalence among TB suspects was 7% with significantly higher prevalence noted among those aged 25–54 years, females, and the subset of TB suspects who were subsequently diagnosed as TB patients.
To highlight the incremental effect of the policy change from “HIV testing of TB patients” to “HIV testing of the larger group of TB suspects”, we calculated the difference in the number of instances of newly detected HIV infections (ie, excluding those persons with previously known HIV infection) (Table 3). Of the 108 HIV-positive patients, 65 (60%) knew their HIV status before the study. Although the current policy of HIV testing of TB patients would have yielded 7 instances of newly detected HIV infection, HIV testing of TB suspects added 36 instances of newly detected HIV infection. To detect a new case of HIV infection, the NNS among TB patients was 13 compared with an NNS of 37 among “TB suspects not diagnosed as TB” (Table 4). Among the subset of TB suspects aged 25 to 54 years who were not subsequently diagnosed as TB, the NNS to detect a new case of HIV infection was 21.
Application of the prevalence of new cases of HIV infection among TB suspects to the overall number of TB suspects examined in 2010 suggested that a policy change could yield markedly increased HIV case finding. Overall in 2010, in Mandya, 23,909 general clients (including TB patients) and 18,681 antenatal mothers were tested, and, respectively, 999 (4.2%) and 57 (0.3%) were HIV positive. If all 19,070 TB suspects were also tested, then approximately 534 cases of previously undiagnosed HIV infection could be expected. This would represent a 51% increase in district HIV case finding relative to current practice.
Routinely testing TB suspects for HIV infection in Mandya detected large numbers of new cases of HIV infection and may represent a major opportunity for increased HIV case finding. Surveillance during our 1-month study period found overall 15 cases of HIV infection among TB patients, including 8 (50%) with newly diagnosed HIV infection; these would be detected by the existing policy. HIV testing among TB suspects, limited to those not subsequently diagnosed with TB, detected 93 cases of HIV infection, including 36 (39%) with newly diagnosed HIV infection. In the overall district context, effective application of a policy of HIV testing of TB suspects in 2010 could have increased HIV case finding by 51%. TB suspects are an easily accessible group of population, already at health facilities with HIV testing. These finding suggests that HIV testing of TB suspects should be carefully explored in India, particularly in high HIV prevalence settings where HIV testing services are widely available and colocated with TB diagnostic services.
This is among the first studies to report HIV prevalence among TB suspects from India or any concentrated HIV epidemic setting globally. A similar study done during the same period in Andhra Pradesh (India) showed HIV prevalence of 10.3% among TB suspects compared with 8.1% in TB patients (Shanta Achanta, MPH, personal communication, September 2011). Limited data from generalized HIV epidemics has also shown HIV prevalence in TB suspects to be similar to HIV prevalence among TB patients.10,17–21 The World Health Organization has recommended that in generalized HIV epidemics, HIV testing be offered to all persons attending health care facilities, including TB suspects. Although the magnitude of HIV prevalence among TB suspects and TB patients was substantially higher in those generalized HIV epidemic settings than found in our study, the principle of a comparable HIV case-finding yield remains.
With more than 90% of TB suspects accepting HIV testing, this study showed that it was feasible to routinely implement PITC among TB suspects in the general health system with the existing staff. Availability of decentralized HIV testing services with colocation in the same facility as the TB microscopy centers might have facilitated such high levels of acceptance.
While ”PITC among TB patients” was clearly a more efficient strategy of HIV case finding than “PITC among TB suspects”, HIV testing of TB suspects remained reasonably efficient and yielded major improvements in HIV case finding. Although the NNS to detect a case of HIV in the general ICTC client pool was 23, the NNS for TB suspects not diagnosed with TB was 37 and was 20 among those aged 25–54 years. Given the feasibility of HIV testing clients already in health care facilities for TB diagnosis and the high yield in terms of absolute numbers of HIV infections diagnosed, we believe that the strategy of “routine offer of HIV testing for TB suspects” in high HIV settings may be recommended by the National program. The effective implementation of this strategy is highly dependent on the presence of a colocated HIV testing service in the same facility as TB diagnostic center; hence, it should not be recommended in areas where HIV testing services are not sufficiently decentralized. Another important challenge would be to maintain functionality of HIV testing centers by ensuring uninterrupted supply of HIV testing kits and availability of trained and motivated staffs—this requires continuous collaboration between the TB and HIV programs. Additionally, this strategy is likely to be beneficial by resulting in earlier diagnosis of HIV among TB patients, earlier initiation of cotrimoxazole preventive therapy and ART, and consequent reduction in mortality.
The study had a few limitations. First, about 8% of the study population could not be assessed for HIV due to various reasons; however, if we conservatively assume all of the untested patients were HIV negative, then the minimum HIV prevalence among TB suspects would be 6.5%. Hence our interpretation of the findings would not change. Second, the HIV prevalence reported in our study was mainly among pulmonary TB suspects who attended the government-run sputum microscopy centers. We were not able to assess HIV prevalence among isolate extrapulmonary TB suspects who were not eligible for sputum examination or TB suspects evaluated at private laboratories outside national program. Though this will remain a limitation in assessment of true HIV prevalence among all TB suspects, we feel that information is not required to inform the policy question facing national TB and AIDS programs. Third, our study did not assess whether the strategy of moving HIV testing upstream to test all TB suspects instead of TB patients actually resulted in increased linkage to HIV care and support and reduced mortality among HIV-infected TB patients. This needs further research. Fourth, the higher HIV prevalence among females compared with males may be due to sociodemographic structure of this setting or may be indicative of the maturity of HIV epidemic in this area. This needs further exploration in future studies. Finally, though indicative, results from a single district is not sufficient to inform national policy decision, as the HIV epidemic is heterogenic with varied HIV prevalence across states and districts of India.22 Hence, broader surveillance from across the country is required before national policy decision.
This study demonstrated that routine HIV testing of TB suspects was feasible in settings with decentralized availability of HIV testing services and yielded a large number of HIV cases in absolute terms. HIV testing of TB suspects was somewhat less efficient than current efforts for HIV case finding among TB patients or then general ICTC population, but remained reasonably efficient in the context of a concentrated HIV epidemic. More importantly, in the context of overall district HIV case-finding efforts, a policy of HIV testing of TB suspects could substantially increase HIV case detection relative to existing practice. Though these findings are promising, given the heterogeneity of HIV in India broader surveillance of HIV among TB suspects is indicated to inform national policy decisions.
We gratefully thank the staff of State TB Cell and Karnataka State AIDS prevention and control society, District TB Officer of Mandya and his staff, and the staff of the microscopy centres and ICTCs of Mandya for helping in completing this study. Finally, the authors are grateful to the TB suspects who have indirectly contributed to this learning.
1. Ghate M, Deshpande S, Tripathy S, et al.. Incidence of common opportunistic infections in HIV-infected individuals in Pune, India: analysis by stages of immunosuppression represented by CD4 counts. Int J Infect Dis. 2009;13:E1–E8.
2. Rajasekaran S, Raja K, Jeyaseelan L, et al.. Post-HAART tuberculosis in adults and adolescents with HIV in India: incidence, clinical and immunological profile. Indian J Tuberc. 2009;56:69–76.
3. Sharma SK, Kadhiravan T, Banga A, et al.. Spectrum of clinical disease in a series of 135 hospitalised HIV-infected patients from north India. BMC Infect Dis. 2004;4:52.
5. Harries AD, Zachariah R, Corbett EL, et al.. The HIV-associated tuberculosis epidemic—when will we act? Lancet. 2010;375:1906–1919.
6. UNAIDS. Report on the Global AIDS Epidemic. Geneva, Switzerland: UNAIDS; 2008.
7. World Health Organization and UNAIDS. Guidance on provider-initiated HIV testing and counselling in health facilities. Geneva, Switzerland: World Health Organization; 2007.
8. Tuberculosis Coalition for Technical Assistance. International Standards for Tuberculosis Care (ISTC). 2nd ed. The Hague; 2009. Available at: www.istcweb.org
. Accessed June 13, 2011.
10. Srikantiah P, Lin R, Walusimbi M, et al.. Elevated HIV seroprevalence and risk behavior among Ugandan TB suspects: implications for HIV testing and prevention. Int J Tuberc Lung Dis. 2007;11:168–174.
11. Porskrog A, Bjerregaard-Andersen M, Oliveira I, et al.. Enhanced tuberculosis identification through 1-month follow-up of smear-negative tuberculosis suspects. Int J Tuberc Lung Dis. 2011;15:459–464.
16. EpiData Association. EpiData entry version 3.1. Available at: http://www.epidata.dk/
. Accessed May 9, 2011.
17. Munthali L, Mwaungulu JN, Munthali K, et al.. Using tuberculosis suspects to identify patients eligible for antiretroviral treatment. Int J Tuberc Lung Dis. 2006;10:199–202.
18. Odhiambo J, Kizito W, Njoroge A, et al.. Provider-initiated HIV testing and counselling for TB patients and suspects in Nairobi, Kenya. Int J Tuberc Lung Dis. 2008;12(3 suppl 1):63–68.
19. Deribew A, Negussu N, Melaku Z, et al.. Investigation outcomes of tuberculosis suspects in the health centers of addis ababa, ethiopia. PLoS One. 2011;6:E18614.
20. Adjei AA, Adiku TK, Ayeh-Kumi PF, et al.. Prevalence of human immunodeficiency virus infection among tuberculosis suspect patients in Accra, Ghana. West Afr J Med. 2006;25:38–41.
21. Dimairo M, MacPherson P, Bandason T, et al.. The risk and timing of tuberculosis diagnosed in smear-negative TB suspects: a 12 month cohort study in Harare, Zimbabwe. PLoS One. 2010;5:e11849.
Keywords:© 2012 Lippincott Williams & Wilkins, Inc.
HIV testing; India; TB suspects; tuberculosis