888    

Background:  The HIV epidemic in sub Saharan Africa has resulted in a 4-fold increase in tuberculosis (TB) incidence in some countries. With the emergence of high rates of TB/HIV co-infection, sputum smear for acid fast bacilli (AFB), the sole widely available diagnostic tool for pulmonary TB, may be less reliable than in HIV^– patients. Newer methods such as interferon-γ release assays using TB-specific antigens may be more sensitive than sputum smear for AFB in diagnosing pulmonary TB disease.

Methods:  A cross-sectional study of 38 adult patients from the NY1 TB clinic in Gugulethu, South Africa, with culture positive pulmonary TB and self-reported HIV status. Sputa were analyzed at the National Health Laboratory Services for AFB smear and mycobacterial culture. Blood for the in-tube whole blood IGRA, which uses the antigens ESAT-6 and CFP-10, was collected at the start of TB therapy. Means were compared using Student’s t-test and categorical variables using Fisher’s exact methods. Any culture positive for MTB was considered a true positive. Confidence intervals for sensitivity were determined using Wilson’s procedure. Overlap of 95% CI was interpreted as lack of statistically significant difference in sensitivity between groups.

Results:  Median age was 38 years (range 18 to 62), 55% female and 25 were new TB cases. Of 38 patients 24 were HIV⁺ of whom 16 were female (p = ns). HIV⁺ patients had significantly lower mean Karnofsky scores compared to HIV^– patients (82 vs 88, respectively, p = 0.004) but did not differ by age or duration of prior TB symptoms. There was no significant difference in BCG scar status or type of TB (new vs retreatment) by HIV status. The sensitivity of the IGRA in HIV⁺ and HIV^– patients was 78.9% (95% CI 56.7 to 91.5%) and 76.9% (95% CI 41.7 to 91.8%), respectively. The sensitivity of a single sputum smear for AFB in HIV⁺ and HIV^– patients was 12.5% (95% CI 4.3 to 31%) and 15.4% (95% CI 4.3 to 42.2%), respectively. Use of 2 sputum smears slightly increased (p = ns) the sensitivity in both groups to 33.3% (95% CI 17.9 to 53.3%) and 46.2% (95% CI 23.2 to 70.9%), respectively.

Conclusions:  The sensitivity of a single or 2 AFB sputum smears for TB diagnosis is very poor, irrespective of HIV status. The IGRA had high sensitivity in both HIV⁺ and HIV^– patients. The IGRA may be a useful tool to aid in the diagnosis of TB/HIV. Studies examining the specificity of this new in-tube whole blood IGRA are needed.

Keywords: TB/HIV Co-infection; Tuberculosis Diagnostics; Interferon-gamma release assay