One long-standing challenge for HIV-infected people with smear-negative TB has been the difficulty in obtaining timely diagnosis and treatment — but one study presented at the conference provided dramatic evidence that clinicians who follow the 2006 WHO smear-negative TB diagnostic algorithm can “drastically reduce mortality in seriously-ill, HIV-infected, sputum smear-negative TB suspects in South Africa,” said Dr Thuli Mthiyane of the South African Medical Research Council (MRC).
In a coinfected person, HIV doesn’t just change the clinical course and nature of TB, it can make it extremely difficult to detect TB by the usual methods, such as smear microscopy and chest x-rays. Smear microscopy involves using a microscope to look for TB bacilli in a sample of sputum (or other biological specimen) that has been stained with a special dye — but specimens from people with HIV and TB are frequently ‘smear-negative.’
According to Dr Mthiyane, in South Africa, the incidence of smear-negative TB has increased 35-fold from 1990-2010 (since the HIV epidemic). Unfortunately, under the previous guidelines for the diagnosis of smear-negative TB — which required performing microscopy on at least three sputum specimens, most of these cases were diagnosed late if at all, resulting in delayed or no treatment, and increased mortality.
In 2006, WHO released revised guidelines for the diagnosis of smear negative pulmonary TB (SNPTB) in countries with a significant burden of HIV. There were a few key changes.
First, in people with or strongly suspected of having HIV, a diagnosis of TB could be made even if only one smear was positive after performing microscopy on only two specimens. A diagnosis of SNPTB could be made in anyone with a cough for two weeks or more if both smears were negative, and there was evidence consistent with TB on chest x-ray and the clinician had decided to treat and monitor. A specimen should be sent for cultural confirmation but treatment should not wait for results.
Likewise, although a short trial with antibiotics (to check for effectiveness against other possible causes of the illness) could be useful during the three-five day period of the initial TB investigations, clinicians should not wait for a week or two for antibiotics to fail before initiating TB treatment. The guidelines also suggest that a diagnosis of SNPTB can be made in, and empiric TB treatment offered to, people with HIV who are dangerously ill with suspected TB even if chest x-ray does not suggest TB, if they do not respond to the antibiotics (including treatment for Pneumocystis jirovecii pneumonia ) after three to five days .
WHO acknowledged that the guidelines were, in some areas, based on expert opinion rather than a strong body of clinical evidence because of the urgency of responding effectively to TB in HIV prevalent settings. In light of this, WHO encouraged operational research to strengthen the evidence base of the guidelines.
So between 2008 and 2009, researchers in South Africa conducted a prospective observational study at three hospitals in KwaZulu Natal (McCord, St. Mary’s and Stanger Hospital) comparing whether people suspected of having TB who were managed under the WHO algorithm for smear-negative TB fared better than those who were managed under current standard practice.1
The study included seriously ill people with HIV over the age of 15 with cough for more than two weeks, a chest x-ray suggestive of TB and at least two negative smears. Researchers assessed whether there was a significant difference in the proportion of TB suspects alive at 8 weeks after admission (about the time it takes to get culture results back). A secondary outcome was whether there was a significant difference in the proportion of patients discharged from the hospital, due to clinical improvement (according to the clinician’s reports in the medical records) at one week after admission.
The study enrolled 351 subjects into the standard practice cohort, and 187 subjects in the WHO algorithm cohort. Demographic characteristics were well matched in both groups, with the exception that subjects in the WHO algorithm group were more severely ill (with more WHO stage IV disease). At admission, most participants had a CD4 cell count below 200, although some had CD4 cell counts that were much higher, and a few patients were already on ART.
Seven days after admission, 37% of patients were still in hospital in the standard practice group, compared to the WHO algorithm group [Odds Ratio (OR) 0.61, 95% confidence interval (CI) 0.41-0.90, p=0.01.] There wasn’t a significant difference in the number who died in hospital; but after 8 weeks, a significantly greater proportion of subjects were alive in the WHO algorithm group (84%) compared to the standard practice cohort (68%) [OR 2.5, 95% CI 1.6-3.8, p=0.001].
“There was a huge difference,” said Dr Mthiyane. This was driven by the success of the WHO algorithm at getting people onto appropriate treatment sooner — less than half (47%) of SNPTB suspects in the standard practice cohort were started onto TB treatment either before or after discharge.
“There was high mortality among those in whom appropriate TB treatment was delayed or not given at all due to a trial of antibiotics,” said Dr Mthiyane. “Eight week survival was highest in those in whom appropriate TB treatment was started within three days in patient’s meeting WHO criteria for SNPTB.”
Some in the audience were left wondering whether such a prospective study was truly necessary.
“How many lives would have been saved if the WHO Guidelines should simply been followed when they came out? We’ve known for years that there is little time to waste in getting a severely ill TB suspect with HIV onto TB treatment,” one doctor told this reporter. “In a high burden setting like this, it is extremely likely that the person has TB. In fact, as long as the clinician keeps an eye out for other possible causes, we should probably put all advanced people with HIV and symptoms of TB onto TB treatment.”