By Theo Smart
A year and a half ago, investigators of the Starting Antiretroviral Therapy at Three Points in Tuberculosis (SAPIT) trial announced that starting antiretroviral therapy (ART) in people with both tuberculosis (TB) and HIV while they are still taking TB treatment dramatically reduces mortality (see http://www.aidsmap.com/en/news/81BF3B62-9197-42D6-9DB3-44A784E2F0B1.asp ). These results, recently published by the New England Journal of Medicine — support the integration of TB/HIV services.1 Other recent data from countries with a high burden of TB suggest that more coinfected people will receive integrated care in a timely fashion when HIV services, including the initiation of ART, are provided directly within the TB services.
To recap SAPIT’s findings: investigators in Durban, South Africa randomly assigned 642 patients with both smear-positive tuberculosis and HIV infection (and less than 500 CD4 cells) to one of three arms, receiving ART:
- within the first two months of starting TB treatment;
- within four weeks of completing the intensive phase of TB treatment; or
- only after the completion of TB treatment (sequential therapy).
Delaying ART was generally the practice at the time, although some observational studies had found improved outcomes with earlier initiation of ART. 2,3 Everyone in the study was taking the same once-daily regimen: efavirenz plus ddI and 3TC (today, ddI would not be part of the first-line regimen).
The study’s data and safety monitoring board recommended discontinuing the sequential therapy arm after a planned interim analysis found that after a median follow-up time of 12 months, mortality rates were 56% lower in the two integrated therapy arms.
Rates of adverse events were similar between the sequential and integrated arms; and although immune reconstitution inflammatory syndrome (IRIS) was three times more common in the two integrated arms, it did not appear to contribute to any of the deaths in the study.
Arms 1 and 2 are ongoing in order to determine the optimal time to initiate ART.
“Our findings provide compelling evidence of the benefit of initiating antiretroviral therapy during tuberculosis therapy in patients with HIV coinfection. The findings also support recommendations by the WHO and others for the integration of tuberculosis and HIV care,” Karim et al wrote.
Indeed, the results have already led to a change in WHO policy. The recent guidelines on ART for HIV infection in adults and adolescents recommend: “start TB treatment first, followed by ART as soon as possible after starting TB treatment.” This is categorised as a ‘strong recommendation’ based on evidence of a ‘moderate quality’ (the evidence would now have a stronger grading since its publication in a major peer-reviewed medical journal).
Several studies at the 40th Union World Conference on Lung Health held in Cancún, Mexico in December last year lend further support to TB/HIV integration, and the early initiation of ART in people on TB treatment.
Mortality
Globally, a third of the people with HIV are coinfected with TB, according to a poster from the KNCV Tuberculosis Foundation — and the leading cause of mortality in people with HIV in Africa. 4 The poster described a KNCV meta-analysis of 15 major cohort studies assessing the impact of TB on mortality in HIV-infected patients. Essentially, the study found that people with HIV who develop TB have twice the risk of death of those people with HIV who do not develop TB.
The effect was less pronounced in people with lower CD4 cell counts — probably because people with more advanced immune suppression are at greater risk of dying from other complications — and was absent in cohorts with access to ART, presumably because ART reduces the risk of TB-related mortality.
Conversely, a study from the International Centre for AIDS Care and Treatment Programmes (ICAP) and the Ministry of Health in Rwanda reported that even though an integrated TB/HIV care programme is scaling up, HIV is associated with a much increased risk of death among TB patients. 5
TB case detection rates remain low in Rwanda, though the recent efforts to strengthen the TB programme have improved treatment success rates, reaching the goal of 85% treatment success in 2006.
Mortality in people with smear-positive pulmonary TB peaked at 7% in 2002, and in 2006 and 2007 dropped to 5%. When all categories of TB (such as smear-negative and extrapulmonary TB) are included, the rate of mortality is higher: 7.3% in people who are HIV-negative. But in people with HIV it is 19.4%.
The researchers concluded that the strengthened TB programme activities are having a positive impact on mortality, but not enough yet in HIV-infected people. However, the mortality data are only available out to 2007 — and much of the progress getting coinfected people onto cotrimoxazole and ART has occurred in the last couple of years.
Problems coordinating care for TB/HIV coinfected patients between TB and HIV services
But the problem may also be that integration hasn’t gone quite far enough. Often, TB and HIV service integration means a system of referrals to a co-located clinic rather than full `one-stop shop` integration. According to another ICAP study, this can lead to disappointing results.
In March, 2008, ICAP supported Mozambique’s national TB programme to perform an evaluation of the implementation of TB/HIV collaborative services that had started in the country several years earlier.6 Data from a random sample of up to 25 coinfected patients were extracted from the registers in each of 22 TB facilities in 11 provinces, and then compared to the corresponding files for the same patients at the associated HIV services unit.
Matching files in the HIV services couldn’t even be located for 41% (181) of 446 TB/HIV patients selected. Of the remaining 265 files, the HIV registers could not confirm that many of the clients were receiving the services they were reported as receiving in the TB registers. The information on who was receiving cotrimoxazole could only be confirmed in 60%. According to the TB register, 154 of the subjects were on ART — but this could only be confirmed for 97 clients (63%).
The programme is redoubling training efforts, and a special form for referrals between the two departments has been introduced in an effort to improve communication.
Other studies have similarly reported that people with TB who test HIV-positive often fail to reach the HIV clinic even when it is right next door, indicating the importance of attempting to provide both HIV and TB services in the same clinic.
TB clinics should start providing all the basic HIV services on-site
A study from Durban reported a high rate of HIV among people who presented to the TB clinic — whether or not they were diagnosed with TB.7 The study was a cross-sectional survey of 741 clients (many of whom were self-referred) presenting to the clinic; 229 turned out to not have TB, 512 did (407 were cases of pulmonary TB).
Sputum specimens from the patients were screened for HIV using the OraQuick HIV1/2 Assay. The HIV prevalence among the suspects without a TB diagnosis was 66.8%; while it was 69.8% among those with pulmonary TB, and 81.6% among extrapulmonary TB cases.
“The high prevalence of HIV among TB suspects suggests that HIV services should be available to ALL individuals who present to the facility. The high prevalence of HIV among individuals who do not have TB, but who present to the facility, provides an opportunity to offer HIV services to a group of self-identified patients at risk… [and] provides further evidence to support the integration of TB and HIV services.”
Full TB/HIV integration at the TB clinic
Increasingly, TB clinics are becoming TB/HIV clinics. In a separate article in this issue, Lesley Odendal of Médecins Sans Frontières (MSF) summarises the experience of the Khayelitsha sub-district in the Cape Town Metro area, which has pioneered full TB/HIV integration in TB clinics.
Two studies from Kenya at the Union Conference this year described successful models where the integration of HIV services into TB clinics dramatically increased HIV screening and access of TB/HIV coinfected people to ART.
A poster presentation from Mugo et al in Kenya noted that the lack of integration between services results in missed opportunities for diagnosis and care, increases suffering, and increases the cost for patients of seeking out health care.8 So in 2008, ICAP working in collaboration with the Kenyan national AIDS and TB programmes established an integrated TB//HIV services site in the TB clinic at the Karatina District Hospital in Central Province, Kenya. Staffing for the site is minimal: a clinical officer and nurse in the TB clinic who have received intensive training in all aspects of TB/HIV management are running the project.
Between June and December 2008, every patient who was registered with TB at the clinic was offered and accepted an HIV test. Thirty-four (21%) tested HIV-positive and were then registered at the comprehensive care clinic (CCC) at the facility.
Their initial HIV care is provided by the TB clinic until they complete their TB treatment. Seventeen (half of the HIV-positive clients) qualified for ART (using the eligibility criteria at the time — TB and a CD4 cell count below 350) — and each one began ART while on TB treatment. The proportion of coinfected people enrolled onto ART increased by 45% after TB/HIV integration. Once TB treatment was concluded, participants were referred to the CCC for continuity of care.
Dr Helena Huerga from Médecins Sans Frontières (MSF) described the experience implementing a similar `one stop shop’ comprehensive TB/HIV care site in a rural district hospital in Homa Bay, Kenya.9
Kenya is a high-burden TB country, with an incidence of 353 new cases per 100,000 people per year. Although the HIV prevalence in Kenya was only 7.4% in 2007, the distribution varies markedly throughout the country — the Homa Bay District, situated on Lake Victoria in Nyanza Province, has the highest HIV prevalence in the country at around 24%.
So in July 2005, MSF in partnership with Kenya’s National TB and Leprosy Programme established a one-stop integrated TB/HIV care service in Homa Bay District Hospital.
Some of the first actions that were taken were to improve infection control precautions in the clinic “because these HIV-positive patients are currently seen in the TB clinic (and this is including TB suspects),” said Dr Huerga. These measures included the construction of an outside waiting area, and the addition of a few more windows and doors to the facility which are kept open to maximise natural ventilation, while nurses are supplied with N95 masks to wear in the unit.
“In terms of human resources, this was a mixed team of MSF and MoH staff. MSF is also working in the HIV clinic of the hospital, so we brought someone from the HIV clinic who was already trained on HIV to the TB clinic to start,” said Dr Huerga. “Now there are people who are rotating between the HIV and TB clinic, both clinic officers and nurses were trained and can do six months in the TB clinic and then six months in the HIV clinic, and then some other staff will come in. The MoH has a different system, but they also rotate everyone in the hospital — but they rotate them to more places. But this helps foster understanding between one programme and the other.”
The patient file system, reporting and recording system were also improved for better programme monitoring.
Dr Huerga described a retrospective analysis that was performed to determine the impact of the one stop care site on HIV testing for TB patients, management of HIV-positive TB patients, and TB programme outcomes, using data collected from the TB register book and patient’s files. To measure the impact, the investigators looked at three periods of time: 1) before implementation from January to June 2005, 2) shortly after implementation, from January to June in 2006, and 3) a ‘’medium term’ after implementation, January to December 2007.
Prior to integration, around 30% of the TB clients were tested for HIV. Shortly after implementation, close to 80% were tested, while a year later, roughly 90% were tested. After implementation, cotrimoxazole and ART initiation also increased significantly from ~45% and 10% respectively to 85-95% and 40-50% afterwards. (The differences between the short-term and middle-term were non-significant).
As for TB diagnoses, “we expected to see an increase in smear-negative TB diagnosed (~40%) but this proportion remained stable, but on the other hand, we did see an increase in extrapulmonary TB from 22% before integration, to 29-30% afterwards (p<0.05).
Of note, by the medium term follow-up period, TB programme outcomes improved — with a significant increase in success rates (p<0.01), and a drop in mortality (p<0.01), though the rate of default was fairly constant.
“We were surprised by this. We would have expected an impact on the defaulter rate since there was more counselling and more patient education, so we need further investigation to understand why,” she said.
The investigators also looked at TB outcomes based upon the HIV status of the TB subject (HIV-negative, HIV-positive on ART, HIV-positive not yet on ART and HIV status undetermined). Although the HIV-negative patients had the best TB outcomes (with an 82% treatment success rate), the difference between their outcomes and the outcomes of HIV-positive clients who started ART was insignificant. “This probably means that these patients are properly managed for both their TB and their HIV status,” said Dr Huerga.
However, HIV-positive patients who were not on ART had a significantly lower treatment success rate (p<0.01) and higher rate of death (p<0.05). People with an unknown HIV status had the worst outcomes, and were significantly more likely to default treatment as well.
“The implementation of one-stop comprehensive TB/HIV care in the Homa Bay District Hospital was very successful and provides a model which should be implemented in all of our TB clinics,” said Dr Huerga. “It is important to note that this improved not only the management of HIV, but outcomes for the TB programme in general.”
One audience member asked whether there were any plans to scale the model up.
“Yes, I think the Kenyan authorities are trying to implement `one-stop` services wherever it is possible to do it,” she said. “The problem is that in Kenya and many other African countries, the TB programme is much larger and far more decentralised than the HIV programme, so there are many more TB clinics than HIV clinics. To introduce HIV care and treatment to all the TB clinics is going to take some time and effort, but I think it is on the mind of the national programme.”
“Nowadays this is the only way of correctly managing TB and HIV — managing them together — there is no other way. 80% of the TB patients in this hospital are coinfected with HIV so it doesn’t make any sense to separate the patients and have 80% of the patients going to the TB clinic and going to the HIV clinic.
Some common themes
There are indeed more TB clinics and units than HIV services, and given the burden of coinfection in many settings — and the need to scale-up to universal access to ART — there will be an increasing emphasis on how to deliver HIV care in these settings
There are some common factors to the successful scale up of TB/HIV integration in these facilities.
One is that TB/HIV services can be successfully provided by trained nurses and clinical officers — who can start and follow-up ART patients.
Another is the importance of TB/HIV training — and the transfer of experienced staff, who can help launch the service and provide ongoing mentoring to less experienced healthcare workers.
It is also essential to implement good infection control — both to protect TB suspects, staff and HIV-infected clients from people with infectious and possibly drug resistant TB.