The basis for the REMEMBER trial

REMEMBER stands for ‘Reducing Early Mortality and Early Morbidity by Empiric Tuberculosis (TB) Treatment Regimens’ and is a randomised controlled trial that is currently enrolling at the international ACTG sites in a number of resource-limited countries.

The study is evaluating the impact of providing any HIV-positive person who presents very late for care (with a CD4 cell count below 50) with empiric TB treatment first — on the presumption that there is a good chance that they have potentially life-threatening but hard-to-diagnose TB — followed a week or so later with efavirenz-based antiretroviral therapy (ART), compared to simply providing them with ART and TB treatment when indicated by local standard of care for TB diagnosis and treatment. The study will measure the effect of both strategies on AIDS progression; virologic and CD4 cell responses, HIV and TB drug resistance, as well as assess the safety and tolerability of and adherence to HIV and TB drugs and the cost-effectiveness of the two strategies. The primary endpoint will be survival status at week 24 (see http://clinicaltrials.gov/ct2/show/NCT01380080).

Background

Two years ago, there had been a debate about the merits of empiric TB treatment in advanced people living with HIV at the Union World Conference on Lung Health when it was held in Cancun. Nowadays because of advances in diagnostics, there are debates whether a trial of the approach remains ethical.

“HIV-positive patients starting ART in Southern Africa have a very high mortality risk, and such patients in the same part of Africa have a very high risk of active TB as well. Many patients die with undiagnosed active TB. Diagnostics in the patients are insensitive and slow, and the clinical trajectory in these patients is rapid, we don’t have time on our side,” Dr Steven Lawn of the London School of Hygiene and Tropical Medicine said. “But empiric TB treatment may reduce their mortality risk. So the conclusion I’ve come to is that the strategy should be tested in a randomised clinical controlled trial with mortality as the primary outcome.”

He had presented a very solid case in support of the study.

“There’s a huge amount of TB that we are missing,” he said. In one study on the burden of TB in people being referred for ART in a periurban community near Cape Town, 52% of the subjects had already had one or more episode of TB (and had completed treatment) while 26% had prevalent TB or were on TB treatment.¹ 11% developed incidence TB during the first year on ART.

“And that’s just the diagnosed TB,” he went on.

Mycobacterial load increases before TB becomes symptomatic, Dr Lawn believes, though it is not clear exactly at what point symptoms become apparent. When they do appear, symptoms are non-specific. However, several post-mortem studies in HIV-positive people have found high percentages with evidence of undiagnosed active TB.²

In another study in Guguletu (a periurban community near Cape Town), pre-ART screening of HIV-positive people, with a median CD4 cell count of 125, found that a quarter had active TB by culture, about a fifth were asymptomatic (and while culture might capture some cases that are primary infection, as described in the upcoming HATIP, this is not a population where one would want to take any chances). But diagnosis was difficult. In this study, fluorescence microscopy only had a sensitivity of 14% (this seems VERY low in comparison with other studies), about 30% of the culture positive TB cases had normal chest x-ray. Liquid culture took 23 days to become positive — the norm is usually around two weeks —suggesting that the specimens had very low mycobacterial loads. (Some of this could be early reactivation disease but some of it could be new primary TB or recent reinfection which appears to be quite common in such high burden settings). 

At the same time, there is a high risk of early mortality in the first year on ART (8-26% in one study) with TB being the leading cause of death.³

Looking for TB would clearly be justified in anyone with advanced HIV, but with old technology at least, this would take too long, Dr Lawn argued — plus, there are many cases where the lab tests go wrong or specimens become contaminated, which means the tests have to be run again. During that time, many of these often severely ill patients will get lost in the referral process and die.

Professor Helen Ayles of ZAMBART and the London School of Hygiene & Tropical Medicine also argued that “patients who enter into HIV care should be given a course of TB treatment. “

“An awful lot of people entering into care have active TB, around 25%. A study by Gavin Churchyard in the mines found that 30% had active undiagnosed TB, but to find that TB you have to do a lot of diagnostic tests [including culture], and how many of us in these high burden settings have access to that? And what is the cost of that?” she said.

“I’m not talking about the cost of the test per se which can be brought down, or the cost to the healthcare provider of setting up the systems [to diagnose TB]. What about the cost to the patient? The out of pocket cost to the patient of coming back again and again to the clinic, waiting for those results; the cost to the patient as they get sicker and sicker, and also the cost to patient and their healthcare providers, in terms of the confidence lost? We can’t even diagnose TB, so why should our patients believe us?” she added.

Dr Lawn also said that there is “no time for complicated smear-negative TB diagnostic algorithms.” He noted that WHO had developed a more simplified algorithm that ends up providing empiric treatment for symptomatic patients if there is no response to a short course of antibiotics, “we’re just saying perhaps you should do that a little earlier on,” he said.⁴

A paper subsequently published reporting on the debate in the International Journal of Tuberculosis and Lung Disease provided some elucidation on this point.⁵

“This strategy could be viewed as an extension of the existing WHO algorithm for the management of seriously ill patients with confirmed HIV infection living in HIV-prevalent settings.⁶ This algorithm recommends empirical TB treatment for those with cough for 2–3 weeks and danger signs, but who have negative sputum smears and fail to respond to parenteral antibiotics for bacterial infection or other specific treatments, such as for Pneumocystic jirovecii pneumonia, within 3–5 days. The potential advantages of routine empirical TB treatment are the elimination of the 2–3 week cough screen (which has very limited sensitivity for TB in this patient group), expediting the start of TB treatment by excluding the need for a prior trial of 3–5 days antibiotic treatment and including many other patients who may also benefit,” Lawn et al wrote.

Among the other patients the approach might potentially benefit, Professor Ayles included patients without active TB. The empiric treatment approach might have direct and indirect benefits.

In our part of Africa, most (~70%) of the general population are latently infected with the mycobacterium Tuberculosis,” she said.

“If they don’t have active TB, chances are they have latent TB and [empiric treatment] lowers the risk of them developing TB for sometime to come,” said Dr Lawn.

Treatment of active cases could also be expected to benefit uninfected people by reducing the risk of exposure in health care settings.

Professor Ayles pointed out that the empiric treatment approach might be a more acceptable alternative for programmes that are afraid to provide isoniazid preventive therapy (IPT) to people living with AIDS, for fear that they might undertreat cases of active TB that their diagnostic process misses. But Professor Ayles added that if specimens taken when the patient enters care show no evidence of TB, it may be possible to discontinue TB treatment after the induction phase of therapy.

The case against empiric treatment

Dr Saidi Egwaga of the National TB and Leprosy Programme in Tanzania spoke against providing empiric TB treatment for all people about to start ART, primarily because of the health systems, logistical and resource challenges he felt the approach would present.  Treating all HIV patients about to go onto ART for TB would dramatically increase the number of people on TB treatment, and the clinical workload. He noted there are drug supply management challenges in many settings, poor record keeping, inadequate staff, and that TB control is highly donor dependent in some countries.

“How would such cases be registered? " he asked (Clearly a TB control programme concern). He also pointed to the weak referral systems between HIV and TB clinics.

“In terms of programmes, I think we have to stop being so rigid regarding the roles of the HIV programmes, and the TB programmes,” replied Professor Ayles. “I think TB treatment offered in ARV clinics would be given through the HIV programme; and of course, recording and reporting needs to be refined. But I don’t see a problem with that.“

At the time, I was concerned about providing empiric treatment to ALL HIV patients about to start ART, regardless of whether they had symptoms, because I didn’t believe that people with high CD4 cell counts who feel well would be willing to go on a full course of TB treatment, particularly with TB being such a stigmatised disease.

However, Dr Haileyesus Getahun from the WHO injected some common sense into the discussion relating to this point.

“The critical point is really to identify which HIV patients would actually benefit from empirical TB treatment. Steve showed the WHO policy – which actually suggested empiric treatment for a certain category of HIV-infected TB patients, those who are defined as seriously ill. Which means they are on the verge of death. The whole objective of empiric TB treatment should be really to increase survival – to prevent death. So in that context I think we have to be very clear, this is not a blanket recommendation for every HIV-infected patient,” he said.

Dr Lawn clarified that he believed the approach should be tested first in those with less than 50 CD4 cells and then perhaps less than 100 CD4 cells, who are at greatest risk of rapid clinical progression and death.

Other concerns centred around the toxicity of the TB drugs, particularly in patients who may be needing treatment for other concurrent conditions; and perhaps most importantly, that the approach would lead to the under-investigation of other alternative diagnoses.

“There is an enormous amount of under-diagnoses of TB in patients with low CD4 cell counts, but there’s also a huge amount of over-diagnosis of TB,” said Dr Sandy Wells from Namibia. “There are also post-mortem studies that show that people die of quite a lot of other diseases in their respiratory systems – other than TB – and a lot of them while they are on TB therapy.”

“In my setting this will be a typical patient coming in with a CD4 cell count of 15 and some pathology. They don’t respond to antibiotics; they are sputum-negative; they are put on TB treatment; they don’t get better. And then you think, “Oh they’ve got drug-resistant TB.” And while you’re thinking of that, they’re dying of PCP, they died of Kaposi’s or common pneumonia or something else. In Namibia, at the moment, we tend to suspect – pretty much anyone when they walk through the door – as having TB. And our patients are very underworked [diagnostically], the staff very rarely think to suspect opportunistic infections. I worry that if you bring in empirical TB therapy for everyone, they’ll be even less likely to think of opportunistic infections,’ and we’ll start to lose more patients from PCP, from pneumonia.”

Dr Lawn agreed this is a valid concern.

“In our patients with CD4 cell counts below 100, 40% have TB,” he said. “With this approach, we are taking over- or under-diagnosis out of the equation and saying: ‘This group has got a very high risk, let’s treat them.’ But I think [ it’s important] that healthcare providers don’t turn their brains off, and keep in mind that these patients could have alternate or multiple pathologies.”

Indeed a recent paper on causes of early death (within two months of presentation) showed that among 353 East African HIV-positive patients who came into Mulago Hospital in Uganda, with complaining of a cough for two weeks, a confirmed diagnosis could be confirmed in only 74 (66%). These included cryptococcal pneumonia in 1%, Pneumocystis pneumonia in 3%, pulmonary Kaposi sarcoma in 4%, and pneumonia caused by two or more disease processes (3%). Evidence of active TB, however, was found in 56%.⁷ Of those who remained undiagnosed, most died before the diagnostic workup could be completed, and ten died after starting a course of broad-spectrum antibiotics.

“Newer and more rapid diagnostic tests are needed to facilitate diagnosis and treatment of respiratory disease, but empiric treatment of smear-negative TB suspects with anti-TB drugs should also be considered in settings with high TB prevalence,” the authors wrote.

“Of course we do need to think about the other opportunistic infections,” Prof Ayles had said back in Cancun. She stressed that the routine diagnostic process should be performed for all patients, though she added that even in the case of TB, already clinicians aren’t waiting for those test results.  “We just all pretend that we’re diagnosing on lab results. But lets face it, we’re not, most people are empirically started on TB treatment. That’s what happens in most cases.”

“Of course, if we get a point of care diagnostic, this whole thing is out of the window,” she added.

An upcoming HATIP will address whether we have reached this point yet.

References

[1] Lawn SD et al. Burden of tuberculosis in an antiretroviral treatment programme in sub-Saharan Africa: impact on treatment outcomes and implications fro tuberculosis control. AIDS: 20: 1605-12.

[2] Lucas et al.  The mortality and pathology of AIDS in a West African city. AIDS 7 (12): 1569-79, 1993; Rana et al. Autopsy study of HIV-1-positive and HIV-1-negative adult medical patients in Nairobi, Kenya. J Acquir Immune Defic Syndr 24 (1): 23-9, 2000. Ansari NA et al. Pathology and causes of death in a group of 128 predominantly HIV-positive patients in Botswana in 1997-1998. Int J Tuberc Lung Dis 6 (1): 55-63, 2002.

[3] Lawn SD et al. Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS: 22:1897-1908, 2008.

[4] Saranchuk P et al. Evaluation of a diagnostic algorithm for smear-negative pulmonary tuberculosis in HIV-infected adults. S Afr Med J 97 (7): 517-23, 2007.

[5] Lawn SD et al. Potential utility of empirical tuberculosis treatment for HIV-infected patients with advanced immunodeficiency in high TB-HIV burden settings. Int J Tuberc Lung Dis 15(3):287–295, 2011.

[6] World Health Organization. Improving the diagnosis and treatment of smear-negative pulmonary and extra-pulmonary tuberculosis among adults and adolescents. Recommendations for HIV-prevalent and resource-constrained settings. WHO/HTM/ 2007.379 WHO/HIV/2007.1. Geneva, Switzerland: WHO, 2007. http://whqlibdoc.who.int/hq/2007/WHO_HTM_TB_2007.379_eng.pdf

[7] Kyeyune R et al. Causes of early mortality in HIV-infected TB suspects in an East African referral hospital. J Acquir Immune Defic Syndr 55 (4): 446-50, 2010.

Visit the HATIP Archive.