High prevalence of drug-resistant TB amongst HIV patients in South Africa

There is a high prevalence of multidrug-resistant tuberculosis amongst patients with HIV in South Africa, investigators report in the April 1^st edition of Clinical Infectious Diseases.

Overall, almost 20% of HIV-positive patients with TB in a cohort receiving care through the South African armed forces had culture-positive tuberculosis (TB), with additional testing revealing that a fifth of these individuals had multidrug-resistant TB (MDR-TB), with a further 3% of patients also having resistance to second-line drugs, (extensively drug-resistant TB, or XDR-TB).

Smear screening alone was revealed as a poor method of identifying those with drug-resistant TB.

TB is the leading cause of illness and death amongst patients with HIV. The burden of TB is especially high in resource-limited settings. Access to laboratory facilities in such settings is often poor, meaning that healthcare workers diagnose the infection empirically on the basis of symptoms alone. The diagnosis of multi-drug resistant strains of TB is especially difficult in these settings, due to the limited availability of drug resistance tests.

There is concern in South Africa (and many other countries) about the emergence of strains of TB which are resistant to first-line (MDR-TB) and second-line (XDR) TB therapy. Successful treatment for these forms of TB is reliant upon an accurate understanding of resistance patterns.

However, South African TB treatment guidelines do not recommend routine drug susceptibility testing for patients newly diagnosed with TB. Furthermore, there is limited information about the prevalence of MDR-TB and XDR-TB in the country.

Therefore, investigators from the Phidisa group (an observational and randomised HIV treatment study involving members of the South African military and their families), designed research with three specific aims.

• The reliability of empirical TB diagnoses.
• Rates of MDR-TB and XDR-TB.
• The ability of sputum cultures, microscopy and PCR analysis to identify TB cases.

Study staff identified patients with symptoms suggestive of TB: persistent cough, weight loss, and unexplained fever.

Sputum samples were obtained from these patients. In a laboratory these were evaluated using culture (the gold standard test), two microscopy smear techniques (Ziehl-Neilson and auramine O) and PCR reaction.

The sputum samples were then used for drug susceptibility testing. Patients were diagnosed as having MDR-TB if resistance to isoniazid and rifampicin was present, and XDR-TB if resistance to the key second-line drugs ethionamide, kanamycin, or ofloxacin was detected.

Results

A total of 785 sputum samples from 584 patients were collected. These patients had a mean age of 36 years, and median CD4 cell count was low at 144 cells/mm³. Consistent with this degree of immune suppression, there was a high prevalence of TB. A total of 116 patients (20%) had at least one culture result that was positive for TB.

Drug susceptibility testing was performed on samples from 107 of the culture-positive patients.

These tests showed that 21% had MDR-TB and a further 4% had XDR-TB. A significantly higher rate of MDR-TB was seen in patients with a history of TB therapy than those without (27% vs. 12%, p = 0.05).

Smears were an insensitive technique for confirming a TB diagnosis, with only 33% of culture-positive patients having their infection confirmed using this method.

Half of the culture-positive samples also tested positive using PCR, but the investigators believe that this was largely because most of the culture results were negative.

“This study demonstrated that clinicians could identify patients with a high likelihood of TB using general clinical parameters that focused on self-reported or clinically observed chronic cough, apparent weight loss, or fever”, comment the investigators. “In these HIV clinics, 19.9% of patients who were asked to provide sputum samples did in fact have culture positive TB.”

They emphasise that only 33% of culture-positive samples were determined to be so by smear and that “PCR alone was not sensitive or specific enough…reliably and definitively diagnose TB.”

However, “PCR testing of isolates for isoniazid and rifampicin resistance showed a high [86%] correlation with culture-based results, providing additional confirmation of the accuracy of these results.”