The Emerging Threat of Drug-Resistant Tuberculosis in Southern Africa: Global and Local Challenges and Solutions: Summary of a Joint Workshop by the Institute of Medicine and the Academy of Science of South Africa (2011)

The workshop presentations and discussions highlighted a number of challenges and open questions for consideration in efforts to improve the treatment of drug-resistant TB. First, as a precursor to successful treatment, early diagnosis of drug-resistant TB is essential; otherwise, delays in treatment lead to higher mortality (see Chapter 5). Second, emerging information from a study of patient-specific resistance profiles and appropriate treatment regimens was presented (Keshavjee and Seung, 2008). Successful treatment outcomes for both MDR and XDR TB have been observed in low- and middle-income countries with a low prevalence of HIV infection, but this observation may not be applicable in a setting such as South Africa where HIV prevalence is high. With HIV coinfection, drug-resistant TB

appears to take a different and more aggressive course. Most patients with HIV-associated MDR and XDR TB die within 6 to 8 weeks, which is also the time typically required to make the diagnosis by conventional culture and drug sensitivity testing. Thus, the majority of patients die before their diagnosis can be documented and before second-line drug treatment can be initiated.¹ A third issue concerns the locus of treatment. On the one hand, it was noted that XDR TB treatment, especially in the earliest stages, should take place in a hospital to ensure management of side effects, treatment literacy in patients, nutritional support, and infection control. On the other hand, it was argued that a move toward a decentralized model of care is necessary for prevention of transmission. For example, as the number of patients on antiretroviral therapy increases, antiretroviral therapy facilities will become more crowded with patients, many of whom are highly susceptible to TB. It is important to recognize that combining HIV and TB programs could result in spreading TB among HIV-infected people. A final set of challenges relates to limitations of health care systems and cost issues.

Presenters offered general principles for the treatment of drug-resistant TB, including the need for a comprehensive approach that includes intensified case finding, preventive therapy, improved treatment literacy, and good infection control. In addition, presenters suggested that patients should have increased access to anti-TB drugs as well as other treatment modalities, including thoracic surgery and immune modulators.

Nesri Padayatchi, University of KwaZulu-Natal, elaborated on the objectives of MDR TB management in KwaZulu-Natal Province:

• to make the diagnosis of MDR TB in patients,

• to ensure that the best possible treatment is available,

• to provide support for each patient to ensure access to treatment,

• to facilitate the continuation of care for each patient,

• to factor the management of HIV into the treatment of TB, and

• to explore other options for treatment.

Gandhi described the program that has been instituted in Tugela Ferry, which has three main components:

• Prevention—Preventing the development of drug resistance by strengthening the DOTS program to curb the creation of drug-resistant strains and by implementing an infection control program to prevent the transmission of such strains.

• Diagnosis—Facilitating early diagnosis by testing and implementing rapid diagnostic assays, which may reduce the time required for diagnosis of drug resistance from 6–8 weeks to 1–10 days.

• Treatment—Decentralizing access to second-line TB therapy to reduce delays in referral, increase capacity, and improve treatment completion rates. When necessary, second-line drugs have been initiated for HIV-infected patients suspected of MDR or XDR TB, and antiretroviral therapy has been integrated into the drug-resistant TB treatment program.

Finally, Padayatchi described some of the common concerns about treatment default rates. The extent to which the current data may underestimate the default rate is unclear. Loss to follow-up is often due to the toxicity of drugs and the duration of treatment for drug-resistant TB, as well as the patient’s distance from the place of treatment. With respect to adverse drug effects, sudden deaths from hypokalemia have been found to be caused by capreomycin; however, renal failure may be related to the presence of a renal condition. More intensive monitoring may decrease deaths related to the drug.

Presenters summarized the findings from studies conducted to characterize TB patients in South Africa and to evaluate the effectiveness of treatment regimens that have been used in those with drug-resistant TB.

Padayatchi presented findings of a study conducted at King George V Hospital. She noted that historically, the outcomes of MDR TB in South Africa have been documented to a limited extent. One of the first outcome studies, by Simon Schaaf in the late 1980s, showed a 33 percent cure rate for MDR TB, with 48 percent of patients dying (Schaaf et al., 1996). In the late 1990s, a national study of outcomes showed that after 6 months of MDR TB treatment, more than 70 percent of patients discontinued the treatment. This remains a huge public health problem in South Africa and particularly in KwaZulu-Natal, where the cure rates are below 50 percent. Figures from 2005 to 2007 in Tugela Ferry show that by 30 days from sputum collection, the mortality rates in MDR and XDR TB patients were

already 40 percent and 53 percent, respectively (Gandhi et al., 2010). Most of the deaths occurred within 2 months, accentuating the need to diagnose MDR and XDR TB earlier.

At King George V Hospital, MDR TB cases increased exponentially from 1994 to 2008. Padayatchi’s team evaluated more than 1,200 patients at the hospital between 2000 and 2003 (Brust et al., 2010). The cure and completion rate was about 44 percent, 21 percent of patients defaulted, 18 percent died, and 17 percent failed treatment.

The 12-month outcomes for patients with XDR TB at King George V Hospital showed that antiretroviral therapy increases survival. Mortality was highest in the first 2 months after XDR TB treatment. Among those who survived 2 months on second-line drugs, survival rates were higher for those on antiretroviral therapy and those with CD4 cell counts above 200. Decentralized management will allow earlier access to care and initiation of treatment of MDR and XDR TB, as well as HIV coinfection, at decentralized sites (O’Donnell et al., 2009).

Padayatchi noted that several new and existing drugs are undergoing randomized controlled trials in South Africa, including TMC207, linezolid, and fluoroquinolones (e.g., moxifloxacin). The outcomes of the TMC207 study were published in the New England Journal of Medicine toward the end of 2009 (Diacon et al., 2009). Clearly this is a drug with potential, Padayatchi said. Linezolid has been known to have in vitro and in vivo activity against M.tb. and increasingly has been used in drug-resistant TB patients. Adverse side effects include anemia and peripheral and optic neuropathies related to mitochondrial toxicity. These neuropathies tend to occur after about 4 months of therapy and may not be resolved with the discontinuation of linezolid.

Several studies have examined the sensitivity of various strains of M.tb. to fluoroquinolones, and independent studies have confirmed that moxifloxacin is among the most active of the drugs tested. Use of this effective group of drugs needs to be continued, despite the levels of resistance that have been indicated, Padayatchi said. Except for linezolid and TMC207, other new classes of drugs have not yet been studied in randomized controlled trials for drug-resistant TB.

Padayatchi referenced a study in Europe evaluating 425 patients with fluoroquinolone-resistant versus fluoroquinolone-susceptible MDR TB. Study data reveal that a higher proportion of deaths and treatment failures occurred in the fluoroquinolone-resistant patient group (Migliori et al., 2008). According to Padayatchi, the judicious use of fluoroquinolones needs to be encouraged.

Charlotte Kvasnovsky of the South African Medical Research Council (MRC) reported findings from a 2008 MRC retrospective analysis of patients diagnosed between October 2006 and January 2008 in South Africa’s Eastern Cape Province, where MDR TB cases have been diagnosed and treated with a standardized regimen since 2000 (Kvasnovsky et al., 2010). The study looked at data collected during site visits to two TB hospitals in June 2008 and January 2009, with the data being censored (cessation of data collection) at 12 months of treatment or upon the death of a patient. The objectives of the study were

• to describe the characteristics of HIV-negative and HIV-positive patients with XDR TB in the Eastern Cape,

• to determine patient mortality after 12 months on XDR TB treatment,

• to elucidate risk factors for early mortality,

• to ascertain the role of HIV in mortality due to XDR TB, and

• to consider the effect of treatment regimens on patients with XDR TB.

Over the 15 months, 274 patients were diagnosed with XDR TB; 65 of those patients, or 23.7 percent, died before starting XDR TB treatment. Of the 206 patients in whom treatment was initiated, the HIV status of 195 was known: 108 patients, or 55 percent, were HIV-positive, and 87 patients, or 45 percent, were HIV-negative. Patients were treated with individualized regimens composed of the six first-line and second-line anti-TB drugs available in the province at the time, which were capreomycin, p-aminosalicylic acid, ethionamide, terizidone, ethambutol, and pyrazinamide. Capreomycin and p-aminosalicylic acid, which were available for treatment of XDR TB beginning in October 2006, formed the backbone of the regimen. The effectiveness of each drug was assessed, and drug resistance was tested. Most patients had undergone at least 7 months of previous MDR TB treatment and were dealing with a large burden of TB disease when the XDR TB treatment began.

Among the 206 patients, there were 95 deaths in the first year of treatment. Overall, 36 percent of HIV-negative patients died, as compared with 51 percent of HIV-positive patients. Most of the HIV-positive patients who died were not on antiretrovirals. HIV-positive patients not on anti retrovirals had 2.5 greater odds of death at 12 months compared with known HIV-negative patients. More than 90 percent of HIV-positive patients were diagnosed with HIV prior to admission for XDR TB treatment, and 55 percent of the HIV-positive patients were on antiretrovirals

at the time XDR TB treatment was initiated. As in the model including all patients, patients aged 25–42 or smear positive at the start of treatment had poorer 12-month survival. In assessing the effectiveness of the treatment regimens, a drug was considered effective if (1) it is recognized as an agent for the treatment of TB by the World Health Organization (WHO), (2) the patient had either never received it or received it for less than 3 months prior to XDR TB treatment, and (3) patient isolates were not found to be resistant to the drug by drug susceptibility testing.

Patients were treated with an average of 5.2 different drugs. Of those total drugs, patients received an average of 1.7 drugs that were considered effective. WHO recommends the use of at least four drugs to which the patient’s TB is considered susceptible. It was suggested that, in light of the overall poor outcomes for XDR TB, even the WHO recommendation may not be sufficient to cure patients with XDR TB, at least using those drugs available in the Eastern Cape at the time of the study.

Serious adverse effects were noted when a provider withheld a drug because of side effects or when a patient refused to continue with a drug. Overall, 36 patients (18 percent) experienced a severe adverse reaction to at least one anti-TB medication. HIV-negative patients experienced severe adverse effects at three times the rate of HIV-positive patients—31 percent compared with 9.3 percent. It was found that the use of antiretrovirals made no difference in the rate of adverse effects among HIV-positive individuals.

Of the 195 patients, 190 were culture positive at the start of treatment. Sixteen patients (8.4 percent) achieved sputum culture conversion in a median of 143 days, demonstrating the possibility that XDR TB can be cured. There was no difference in the frequency of culture conversion between HIV-positive (10/108, 9.2 percent) and HIV-negative (6/87, 6.9 percent) patients. In the year since these data were collected, four patients have been cured of XDR TB, and follow-up data continue to be gathered.

Kvasnovsky stated that this research confirmed the benefit of using antiretrovirals in HIV-positive XDR TB cases. Therefore, although the current protocol in South Africa is to start antiretroviral therapy when the CD4 count is under 200 and after the first month of XDR TB treatment, Kvasnovsky recommended the early initiation of antiretroviral therapy in all HIV-positive patients irrespective of CD4 count.²

For purposes of comparison, speakers described treatment of drug-resistant TB in two other countries—Ethiopia and South Korea.

Anne Goldfeld, Harvard University, described treatment of drug-resistant TB in Ethiopia. Ethiopia has the second-largest population in sub-Saharan Africa, with approximately 80 million people. WHO estimates that there are 129,000 new TB cases per year in the country and ranks Ethiopia seventh among the 22 countries with the highest TB burden worldwide. Based on a survey conducted in 2003–2006, 1.6 percent of these cases are MDR TB, and WHO estimates that there are at least 6,000 new cases of MDR TB each year. According to Goldfeld, the backlog of MDR TB cases needing treatment in Ethiopia is significant, and diagnosis remains a challenge.

In 2008, 221 MDR TB cases had been documented at the Ethiopian National Reference Laboratory in Addis Ababa. The Ethiopian Federal Ministry of Health submitted a funding application to the Green Light Committee (GLC) and received approval for treating a cohort of 45 patients in August 2008, with arrival of drugs anticipated in October 2008.

The Cambodian Health Committee conducted the first MDR TB training for doctors, nurses, health workers, and pharmacists in Addis Ababa in October 2008. As follow-up, the Ethiopian MDR TB team attended MDR TB training in Cambodia led by the Cambodian Health Committee. This program involved both didactic and hands-on training in the Cambodian Health Committee’s hospital-, community-, and home-based MDR TB care program. Along with a donation of capreomycin from Eli Lilly, other drugs were acquired with support from the Jolie-Pitt Foundation to begin treating the first cohort of Ethiopian MDR TB patients.

A cohort of 8 patients was started on treatment in February 2009, a second cohort of 12 patients was started in June 2009, a third cohort of 16 patients was started in August 2009, and a fourth cohort of 14 patients was started in October 2009. Treatment of these cohorts was initiated prior to the arrival of the GLC drugs later in 2009, a year after they were expected. When those 45 courses of drugs eventually arrived in Ethiopia, patients were rapidly initiated on them. The 92 patients started on therapy by March 2010 had received an average of three prior treatments for drug-sensitive TB, which had been given because no second-line drugs were available. Fifty patients completed the intensive phase of treatment and have been placed in the continuation phase at home using a community-based approach pioneered by the Cambodian Health Committee in collaboration

with the Ethiopian Federal Ministry of Health, with no defaulters. Approximately 25 percent of the 92 patients were HIV-positive. Three patients died in therapy, two of whom were HIV-positive.

Of the 221 patients who had been diagnosed with MDR TB as of August 2008, all eligible backlog patients who could be traced were started on therapy. Unfortunately, a house-to-house search for these patients revealed that 20 percent were confirmed dead and 50 percent could not be found, some of whom presumably had also died while awaiting therapy.

Goldfeld noted that many of the issues involving TB and MDR TB care in Ethiopia have previously been addressed in Cambodia. One of the first effective programs using short-course chemotherapy with directly observed treatment was begun on the Thai–Cambodian border in a refugee camp in an open war zone in 1980. Subsequently in Cambodia, beginning in 1994, the Cambodian Health Committee began to develop a community-based care approach, “community DOTS,” that has been scaled up to the entire country in recent years. A home-based care approach also has been developed and has been highly successful. In addition, the Cambodian Health Committee initiated integrated TB and HIV care in a provincial TB program and in the TB ward of the largest public hospital in Cambodia. Clinical research performed by the Cambodian Health Committee, supported by the National Institutes of Health (NIH) and the French National Agency for AIDS Research (ANRS), has addressed the question of the optimal timing for the introduction of antiretrovirals in immunosuppressed AIDS patients with TB. This research provided an opportunity for the country to obtain first-line AIDS drugs and then drugs for MDR TB cases. The Cambodian Health Committee subsequently obtained second-line drugs for MDR TB through the GLC process and initiated treatment for 136 patients throughout the country by March 2010. Twenty percent of these patients were initiated on MDR TB therapy at home in accordance with the community- and home-based models. An extensive community network of patient sup porters administers the drugs with the help of health workers and local health centers. Goldfeld suggested that these approaches are eminently suitable for application in Ethiopia and that this South-to-South transfer of knowledge can serve as a model for expansion.

In southern Africa, there is little doubt that HIV infection is accelerating the development of resistance to TB drugs, said Clifton Barry III, NIH. However, resistance is also a natural consequence of treatment with drugs, as is demonstrated by the fact that the same phenomenon is taking place in regions where HIV infection is not common.

Barry described the experience of South Korea, where HIV infection is rare. The National Masan Tuberculosis Hospital is the largest TB referral hospital in the country. It treats about 800 newly admitted patients every year and has about 350 outpatient visits per day. Patients are staying for an increasingly long time, with the mean average duration of admission currently being about 130 days.

Closely tied to the work of the National Masan Tuberculosis Hospital is that of the International Tuberculosis Research Center, which was established in 2005 in South Korea with the aim of improving the diagnosis and treatment of drug-resistant TB and the methodology used for clinical trials of new drugs, as well as understanding the virulence of drug-resistant TB. About 800 patients are currently on four open protocols, the first of which started in 2005. A natural history collection is used to benchmark every patient who comes to the hospital, and the treatment duration is 2 years. In South Korea, a diagnosis of TB requires a mandatory report to the ministry of health, which results in perfect recall for patients who relapse.

Barry reported on a study begun in 2005 at the center, which involves collecting all the strains from patients infected with MDR TB, determining host genetic information, and characterizing different responses to therapy in a large cohort of infected patients. An extensive database of information has been collected by following patients essentially for life. Cohort A includes patients who have never before been treated for TB. Cohort B includes patients for whom retreatment is documented. The cure rate is approximately 58 percent. About 10 percent withdraw from the study for various reasons: 11 percent die; 18 percent default; and 3 percent are chronic cases, almost all of whom have TDR TB (where total drug resistance is defined operationally). Most of the strains are either MDR or XDR TB.

Genotyping and clustering were performed in the cohort enrolled from May 2005 to December 2006, which included 208 culture-positive patients. Even though the rate of clustering was not insignificant, it could be concluded that the XDR TB was not spreading effectively. However, the strains responsible for the individual branches of the epidemic were definitely expanding. In northern parts of South Korea, one subset of the Beijing strain was expanding; in the southern region, a different subset was found.

To date about 1,400 bacterial isolates have been collected. Patients who are on therapy can be followed individually, and drug susceptibility tests can be compared. The injectable drugs used to treat drug-resistant TB are associated with very different rates of adverse reactions, which constitute a significant and underappreciated problem for the management of these patients. Individualized regimens are used wherever possible and are adjusted for adverse reactions. The major predictor of adverse reactions is previous treatment history.

Researchers have compared the results of drug sensitivity testing with clinical outcomes and have found groups of organisms with discrete resistance patterns. They also have detected the major mutations associated with each strain. Four major allele changes are associated with injectable drug resistance, some of which confer cross-resistance. Streptomycin and capreomycin resistance involves discrete genotypic changes. Prior acquisition of kanamycin resistance confers capreomycin cross-resistance, but prior acquisition of capreomycin resistance does not confer kanamycin cross-resistance. Kanamycin and amikacin have particularly high levels of cross-resistance and therefore should be reserved for last use. This approach would require a change in policy guidelines, but it would bring some order to the way these drugs are used clinically and improve outcomes.

Barry stated that within a year or two, resistance will be determined using genotypic resistance determinations. Genotypic drug susceptibility testing has performed well in predicting clinical outcomes, as measured by 6-month sputum conversion.

Padayatchi discussed the recently published SAPiT (Starting Anti-retro viral therapy at three Points in Tuberculosis therapy) study, the first randomized controlled trial that corroborated previous evidence on the need to integrate TB care and antiretroviral therapy (Abdool Karim et al., 2010). The study showed that mortality was 56 percent lower in patients who began antiretroviral therapy as soon as possible compared with those who waited until TB treatment had been completed. There have been concerns about potential overlapping toxicities in drugs for MDR TB and HIV; some of the common concerns are peripheral neuropathy, psychiatric symptoms, and hepatitis. However, there are many ways in which these and other potential toxicities can be managed.

Of South Africa’s approximately 500,000 TB cases per year, 73 percent are associated with HIV infection, compared with about 12,000 such cases in the United States. It is impossible to “treat our way out of the epidemic,” said Stephen Lawn, University of Cape Town. Prevention strategies are essential.

Lawn cited recent data on nine cohorts around the world from both high- and low-prevalence countries showing that antiretroviral therapy substantially reduces TB rates in HIV-positive people, with a range of 54 to 92 percent protection (Lawn et al., 2009) and a pooled estimate of 67 percent (95 percent confidence interval [CI], 61–73 percent) (Lawn et al., 2010). The data demonstrate that there is a time-dependent reduction in TB rates, but the maximum benefit is probably acquired within the first 2 years of treatment. This represents a major advance in preventing TB across different settings, in patients with both moderate and advanced

immunodeficiency, and in both skin test–positive and skin test–negative patients. It is likely that this observation applies to both drug-sensitive and MDR TB, said Lawn.

Patients referred into the antiretroviral therapy program in Gugulethu in the Western Cape, where the burden of TB is high, undergo a 1-month period of screening and investigation for TB before starting antiretroviral therapy. This period is one of great risk for nosocomial TB transmission. More than half of patients referred to the antiretroviral therapy program have already undergone one or more courses of completed TB treatment in the past, a quarter have prevalent disease at baseline, and 11 percent develop TB during the first year of antiretroviral therapy (Lawn et al., 2006). Forty percent of patients are newly diagnosed with TB and commencing treatment for TB at the time they begin antiretroviral therapy, showing that presenting for antiretroviral therapy is unmasking cases of TB that would otherwise go undiagnosed. If these patients were rigorously screened for TB using active microbiological screening prior to starting antiretroviral therapy, the risk of nosocomial TB and MDR TB transmission could be decreased.

Observations made in Tugela Ferry from 2005 to 2007 provided valuable data concerning drug-resistant TB. Virtually all the cases of MDR and XDR TB were associated with HIV and occurred in patients with very low CD4 cell counts (see Box 6-1 for a discussion of CD4 count as a predictor of mortality among MDR and XDR TB patients).³ The proportion of patients who were on antiretroviral therapy upon diagnosis of MDR and XDR TB was 15 percent and 22 percent, respectively (Gandhi et al., 2010). Almost two-thirds of patients had received a previous diagnosis, most within the previous year and most of drug-susceptible TB. Thus, many of these patients had been under the care of the health care system for some time prior to acquiring MDR TB, and an opportunity to prevent these infections using antiretroviral therapy was missed.

Because many MDR TB outbreaks are driven by HIV-positive patients with very low CD4 cell counts, antiretroviral therapy has a potent preventive effect. The development of new drugs to treat MDR and XDR TB patients effectively will be extremely costly and will take a very long time. By contrast, more than 20 antiretroviral drugs are currently available, and the potential to use them more effectively to prevent HIV-associated TB and MDR TB is currently being squandered. An aggressive scale-up of antiretroviral therapy at the earliest opportunity is warranted, said Lawn.

It is important that antiretroviral therapy be initiated as soon as possible, with patients receiving treatment immediately upon presenting to the health care system. Lawn noted that current WHO guidelines recommend that antiretroviral therapy be initiated in all HIV-infected individuals with active TB as soon as practicable after the start of TB treatment, irrespective of CD4 cell count (WHO, 2010c).⁴

Several workshop participants noted that physicians can inadvertently increase the non-TB patient population’s exposure to important anti-TB drugs. For example, common infections, such as urinary tract infections, often are treated with fluoroquinolones (which are also used to treat MDR TB), a practice that can increase an individual’s resistance to this important category of antibiotics. Doctors may need an intensified education program concerning the management of TB and the importance of accurate diagnoses, suggested one participant.

In addition to the cotoxicities of antiretrovirals and second-line TB drugs, the risk of immune reconstitution inflammatory syndrome (IRIS) is greater when the interval between the start of TB treatment and the start of antiretroviral therapy is shorter. Patients thus need to be monitored for cotoxicities and IRIS, said Lawn. An early start on antiretroviral therapy can lead to toxicity problems, but given the positive effect of antiretroviral therapy on lowering mortality and morbidity rates, an option to be treated early is necessary. At the same time, Lawn suggested, research is needed to assess the relationship between the early initiation and the toxicity of antiretroviral therapy, and ways must be found to make early anti retroviral therapy sustainable. Moreover, Lawn noted, the recent revision of the WHO guidelines has promoted a phasing in of less toxic antiretroviral regimens (WHO, 2010c). According to the WHO guidelines, initial anti-retroviral therapy should include a non-nucleoside reverse transcriptase inhibitor plus two nucleoside reverse transcriptase inhibitors, one of which should be zidovudine or tenofovir. WHO also recommends that countries begin to phase out the use of stavudine (d4T) in first-line antiretroviral regimens because of its toxicities.

Chaisson suggested that much more clinical research on the treatment of MDR TB is needed. For example, more research is needed on existing drugs and their effectiveness against particular TB strains.

Jacobs emphasized the potential of research on preventing the emergence of drug resistance with simple compounds available in the diet. He explored the possibility that particular compounds can block the mutations that lead to resistance. When excess cysteine was added to a culture undergoing selection for isoniazid resistance, it sterilized the culture. In other words, mutants were unable to arise from the persistent cell population when cysteine was present. The effect is dose- and time-dependent, and the thiol group of cysteine is essential for activity. Jacobs suggested that this idea should be examined in the search for compounds that can prevent the emergence of drug resistance.

Sturm noted that the variation among strains in southern Africa is

much greater than is suggested by the broad categories of MDR and XDR TB. Standardized treatment regimens for these categories are therefore inadequate and will lead to greater resistance; however, the region is not yet at the point at which individualized treatments can be instituted, according to Sturm. Much more needs to be known about resistance patterns for TB strains throughout southern Africa to support the development of individualized treatment regimens, he noted. Barry pointed out that XDR TB is not a biological descriptor of the strain; the question that should be asked is how the structure has changed and what is seen in the patient.

Wallengren explained that the experience in KwaZulu-Natal demonstrates some of the challenges and opportunities of community-based care. As MDR TB cases have increased, there has been an increase in admission of MDR TB patients to King George V Hospital in Durban, South Africa, which has only 320 beds. It takes approximately 16 weeks to initiate treatment once a sputum culture has been taken and the patient has been admitted to the hospital. To accommodate the dramatic increase in admissions, it has become necessary to discharge patients, many of whom are not yet healthy and are still infectious. The average length of stay in the hospital has decreased from 6 months to 2 months. In 2006, more than 30 percent of patients were culture positive, and more than 15 percent were smear positive when they left the hospital. These patients went home to households that were unprepared to manage side effects or practice infection control. Patients had to return to King George V Hospital once a month for follow-up visits, sometimes traveling distances of up to 500 km. A lack of infrastructure support for patients has resulted in poor treatment outcomes and a cure rate of just 40 percent over the last few years at the hospital.

Today hospitalization is not feasible because of the high numbers of MDR TB patients. In KwaZulu-Natal Province, alternative treatment is already taking place, with many patients returning to their households and communities, where treatment is continued. The challenge is to provide these patients with the support they need.

To this end, decentralized MDR TB treatment is being piloted at four sites situated in four districts of KwaZulu-Natal; patients began treatment in the decentralized system in 2008. King George V Hospital has a supervisory role and provides technical support to the decentralized sites. Complicated cases are admitted to the hospital. Eligible patients are started on the intensive phase of treatment. They receive appropriate education

about their disease, their homes are assessed with respect to decentralized care, and a plan for continued care is issued. When patients are discharged from the hospital, they are assigned to teams within the decentralized unit. The teams administer injections, monitor side effects, support treatment, refer complications to the unit, and continue patient and family education in TB and infection control. The teams are assisted by DOTS supporters or family members who are trained as treatment supporters. Family members are involved in both the intensive phase and the continuation phase of treatment. Patients return to the decentralized clinics for a monthly visit, where treatment is assessed and side effects are addressed. The teams transport drugs in cooler boxes and use a GPS unit to locate patients. When roads are impassable, team members sometimes have to walk considerable distances to get to patients. The selection criteria for decentralized care are as follows:

• Patients must have noncomplicated MDR TB.

• The decision to place a patient under community care is made by a multidisciplinary group consisting of a medical doctor, a nurse, an MDR TB coordinator, and a social worker.

• The patient should be smear negative and have stable accommodations and nutritional support.

• The patient and family must have received adequate education and be able to follow the plan for the administration of drugs.

Since July 2008, 420 patients have been placed in decentralized care, and their progress has been compared with that of patients in centralized care at King George V Hospital. Some early results are as follows:

• Most patients are tested for HIV, and the HIV-positive rate is 74 to 78 percent.

• Decentralized sites have been effective in ensuring that HIV-positive patients receive antiretroviral therapy.

• Integration of antiretroviral therapy into the MDR TB treatment program has been problematic. A lack of communication among vertical programs is a notable challenge. Moreover, it seems that the TB/HIV integration itself has become a vertical program. Wallengren noted that the challenge for both TB and HIV treatment lies in the clinics delivering the care, rather than in the patients not complying with treatment.

• Mean TB smear positivity at diagnosis is 73 percent in the decentralized sites and 57 percent in the centralized hospital.

• The treatment delay is shorter at the decentralized sites, with an average of 90 days from diagnosis to treatment, compared with 108 days at King George V Hospital.

• TB culture conversion is 66 percent in the decentralized sites.

• With respect to early treatment outcomes, 14 percent died, 3 percent failed, and 2 percent defaulted.

According to Wallengren, the challenges faced by the decentralized program include high staff turnover, inadequate staff training, inadequate support for nurses (clinically, monetarily, and administratively), drug shortages (due to either reduced orders at the district level or problems with procurement and distribution at the provincial level), and integration of antiretroviral therapy.

The current TB epidemic is reminiscent in some ways of the HIV epidemic in the late 1980s and early 1990s, said Robin Wood, University of Cape Town. The numbers of patients were so great that the capacity of existing health care facilities was inadequate. There was great fear among health professionals, some of whom shunned infected patients. A major difference, however, is that the HIV epidemic was more linked with human rights issues, and patient and activist groups served as advocates for change.

Because sufficient beds are not available, drug-resistant TB patients diagnosed in the general ward of a hospital generally remain there while they wait for a bed to become available at a referral hospital. These patients are being cared for by people who are insufficiently trained and unaware of both the needs of the patients and infection control. Building more capacity requires money and time. A short-term solution, said Wood, is to modify existing health care facilities, improve infection control, and incorporate community-based management of TB. Wood emphasized that concern for the human rights of patients should underpin any strategy used to manage TB. Of equal concern is the health of health care workers.

Padayatchi called for simultaneous advances on several fronts. She cited the need for (1) rapid diagnostics and new TB agents; (2) health systems research to analyze such options as decentralized treatment and integrated management of HIV, MDR TB, and XDR TB; (3) better referral and outreach systems to minimize the default rate; (4) advocacy; (5) improved surveillance; (6) an appropriate prophylactic for contacts (for both adults and children); and (7) further exploration of empiric treatment, which is already being started in Tugela Ferry.

Lerole David Mametja, Department of Health, South Africa, stressed the need for financial support. When WHO reviewed South Africa’s TB control program in July 2009, it estimated that 1 percent of the population or approximately 490,000 South Africans had TB in 2008 and that 2 percent of the TB cases were MDR. Data from the National Health Laboratory

Service showed that about 6,200 cases of MDR TB were diagnosed in 2008, but only 4,600 became part of the treatment program. Thus a large number of MDR TB cases are disappearing from the system.

WHO has estimated that successful TB control and management in South Africa would require R5.2 billion. Mametja stated that not enough funding is available to produce the required response to TB in the country. Challenges involving the interactions between national and provincial governments have resulted in the inability of the national TB program to access all the funds assigned to it. These challenges have impeded the reversal of the spread of TB and drug-resistant TB.

According to government policy, all drug-resistant TB cases must be hospitalized, but only 2,000 beds are available for this purpose. Thus the policy is unworkable, said Mametja. Although South African governments have tried to expand the bed capacity, funding this expansion has been difficult. Moreover, the solution will have to go beyond creating more beds; it may require returning to the basics of TB control and management so as to preempt the onset of MDR and XDR TB. One workshop participant added that the cost of drugs is an issue as well, with a course of moxifloxacin and capreomycin, for example, costing thousands of U.S. dollars. Barry said that the choice is either to add to the problem by using suboptimal drugs or to design a regimen that contains the epidemic. Pressure must be applied to bring the price of drugs down, said a participant.

Cassell noted that a good DOTS program will not necessarily prevent the spread of MDR TB, which has continued to increase even in areas with such programs. However, a major requirement for accessing global funds is to have a good DOTS program in place. In fact, patients are spreading resistant strains for almost 2 years until drugs arrive.

Gandhi observed that since 2007, Tugela Ferry has seen a major influx of resources, primarily financial, that have allowed staffing for the TB treatment program to be increased. From two nurses in 2004, the TF CARES facility now has more than 30 staff members working in the program. As a result, treatment for drug-susceptible TB has improved from a little more than 60 percent to more than 83 percent at present, and the default rate for 2 consecutive years has been zero.

This page intentionally left blank.