We thank Dr. Chogle for his comments. In our study on lupus nephritis, there was a high prevalence of major infections (42.3%), including tuberculosis (11.5%). Importantly, apart from a direct effect on mortality (50% of all deaths), the occurrence of major infections was a major risk factor for poor outcome in lupus nephritis (chronic renal failure or death). Indeed, in light of these data, we share Dr. Chogle's concern about tuberculosis and in general about infections in lupus patients. However, the setting of our study needs to be made clear. India has the highest number of incident cases of tuberculosis in the world (2 million per year), and it is estimated that ∼40% of the Indian population is infected with tuberculosis bacillus (1). Thus, it is not surprising that our immunosuppressed lupus patients face the brunt of this onslaught too. Also, we have not been testing for latent tuberculosis (as a corollary, not giving tuberculosis prophylaxis) in our lupus patients.
Dr. Chogle has raised an important question: will prophylaxis help? Although there are no explicit guidelines in lupus nephritis, the American Thoracic Society does recommend targeted tuberculin testing and prophylaxis in patients on immunosuppression (2). Also, the Indian Rheumatology Association recommends prophylaxis in patients starting anti–tumor necrosis factor α therapy similar to Western guidelines, discounting the possibility of increasing resistance due to prophylaxis (3). Should we apply the same recommendations to our lupus patients? The answer may partly lie in the numbers. The estimated number of incident cases of tuberculosis in the general population of India is 168 per 100,000 people (1); however, the crude incidence rate in our patients was 1,923 per 100,000 patients per year (assuming the duration of the study to be 6 years, i.e., the median duration of followup). Compared with the risk of prophylactic isoniazid–induced hepatitis (estimated to be 278 per 100,000 people) (4), it seems clear that the medical benefit will outweigh the risk of prophylaxis.
However, what about other factors? An important consideration is the sheer magnitude of prophylaxis; we will have to treat 40–50% of our patients! There are concerns of an increased pill burden, cost (a major issue for our patients), and resistance. However, in view of the results of our study, a review of our practice is certainly warranted. A previous study from India where isoniazid prophylaxis was administered to all lupus patients requiring long-term steroids (without testing for latent tuberculosis) found a 1% incidence of tuberculosis in the first year, which was lower compared to a historical cohort (5). Nonetheless, a prospective study is warranted, and this remains an important area for research.
The second question raised by Dr. Chogle is with regard to testing for latent tuberculosis by IGRAs. IGRAs have been shown to have a higher sensitivity and specificity (6), and have been recommended over the tuberculin skin test by the Centers for Disease Control and Prevention in the case of BCG-vaccinated subjects (7). Despite the expected theoretical benefit in a highly BCG-vaccinated country like India, in a study of 726 health care workers in rural India, both tuberculin skin tests and IGRAs were positive in 30%, whereas each picked up an extra 10% that the other test did not. There was a high agreement among the tests (81.4%) (8). Indeed, in the absence of good evidence, the World Health Organization issued a policy statement in 2011 saying “There is insufficient data and low quality evidence on the performance of IGRAs in low- and middle-income countries, typically those with a high TB and/or HIV burden” (9). Based on a huge difference in the price of the two tests (the cost difference is 25 times), as of now, the tuberculin skin test seems to be the right choice until further studies are performed in our setting.