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)

Drug resistance arises through either acquired resistance or transmission. As discussed in Chapter 3, there is evidence that transmission plays an important role in the development of MDR TB. Of the 511,000 individuals estimated by the World Health Organization (WHO) to have MDR TB in 2007, 57 percent had never previously been treated for TB and thus could not have acquired resistance from prior treatment. The remaining 43 percent were retreatment cases; however, it is known that reinfection through transmission also occurs in this population. This chapter reviews what is known about the transmissibility of TB from person to person, as well as experience with infection control programs to reduce or eliminate transmission. Overall, many presenters emphasized that a number of fundamental questions about transmission remain unanswered and that, if infection control is to be successful, much more research and evaluation of these issues will be necessary.

According to Mphahlele, experiments done for more than half a century have demonstrated great variability in the infectivity of patients with TB. Guinea pigs have often been used in these studies because they are highly susceptible to infection by human M.tb., more so than mice and rabbits and perhaps as much as AIDS patients; in addition, their immunological response is similar to that of humans.

Beginning in the 1950s, Riley and colleagues exposed guinea pigs to untreated TB patients to measure the infectiousness of M.tb. Of the 156 guinea pigs, 71 were found to be infected at the end of the 2-year study, demonstrating the ease of transmission in the absence of direct contact (Riley et al., 1959). Recent studies conducted at the Airborne Infection Research (AIR) facility in Witbank, Mpumalanga Province, South Africa, examined both the infectivity of MDR TB and the question of whether transmission can be interrupted. The infectivity experiment, which involved 360 guinea pigs exposed to 26 MDR TB patients, revealed that some infecting MDR and XDR TB strains may be immunogenic but not sufficiently virulent to cause progressive disease in guinea pigs and that guinea pigs may be clearing TB infection. In that study, 75 percent of the guinea pigs became infected, and of those strains recovered from the guinea pigs, only two matched the strains of patients. Despite large positive skin test results, however, many of the guinea pigs showed no histological evidence of residual TB infection, and the burden of TB infection was higher among recent large skin test converters than among more remote converters. Many of the guinea pigs

appeared to have had “transient” TB infection, which is consistent with the results of skin test and interferon-gamma release assay conversions reported in the older literature.²

Other studies have demonstrated some of the factors associated with the transmission of M.tb. among individuals, finding, for example, that transmission in a hospital setting can be widespread within a brief period of time (Haley et al., 1989). Another study that looked at TB transmission on airplanes revealed that duration of exposure and proximity to the index case have an effect on transmission (Kenyon et al., 2009).

The infection control experiment done at the AIR facility assessed the efficacy of an infection control intervention. In a two-phased study, ultraviolet (UV) radiation was switched on and off on alternate days in two animal rooms. When the UV radiation was switched on, air was transmitted to the intervention group of guinea pigs; when the UV radiation was off, air was transmitted to the control group of guinea pigs. In the first phase of the study, after 4 months of exposure, 9 guinea pigs in the control group were infected and none in the intervention group, showing 100 percent protection. In the second phase, after 3 months of exposure, 48 guinea pigs were infected in the control group and 15 in the intervention group, showing 85 percent protection.

Mphahlele observed that the findings of the AIR studies indicate a global need to better understand transmission of MDR and XDR TB through basic research. She noted that future studies planned for the AIR facility will address the efficacy of surgical masks for patients; novel interventions, such as inhaled antibiotics; transient TB infection, strain variation, and XDR TB transmission; and determination of when MDR and XDR TB patients become noninfectious.

The 2006 study of 53 patients with XDR TB in Tugela Ferry described in Chapter 2 provided evidence for nosocomial³ transmission (Gandhi et al., 2006). Evidence supporting the suggestion that infections had been acquired largely in the health care setting included four findings: (1) most patients had been admitted to the hospital during the preceding 2 years, (2) initial community contact tracing uncovered few additional cases,

(3) health care workers were among those who died from XDR TB, and (4) genotyping showed a similar strain in 83 percent of patients. According to Friedland, since that time, a number of other studies have confirmed transmission within the hospital setting from patients identified as having had XDR TB. TB infection is transmitted not only from patient to patient but also from patient to health care worker and from health care worker to patient. WHO and the Centers for Disease Control and Prevention (CDC) have formulated procedures for reducing nosocomial transmission through infection control measures. Infection control protocols are divided into three categories: administrative, environmental, and personal protection.

Shean reported on the preliminary results of an ongoing study of latent and active TB in 267 health care workers at five primary care sites, conducted by Dr. Shahieda Adams. Information had been collected on occupational and environmental determinants of TB infection, including infection control practices at work sites (Adams et al., 2010). In the studied population, the majority were female (78 percent), and all worked at community health centers or primary care clinics. The majority reported having had a BCG vaccination,⁵ 6 percent were diabetic, 15 percent were current smokers, and 27 percent reported using alcohol. A questionnaire on infection control practices was administered. HIV status could be disclosed to researchers, but participants were also encouraged to be tested; 72 percent of the respondents were willing to be tested for HIV at the time of the questionnaire. Those who had received previous TB treatment made up 18 percent of the group, and 31 percent had at least one current TB symptom. The majority of health care workers reported having daily contact with TB patients (89 percent), and 32 percent had some level of health qualification. Table 4-1 shows the percentage of the health care workers who reported that various infection control measures were in place in the facilities in which they worked.

Tests done on the health care workers showed that

• 31 percent were positive on a screen of symptoms,

• 19 percent were HIV-positive (either reported as or tested positive),

• 10 percent were positive on a chest x-ray,

• 84 percent were skin test positive,

• 66 percent were positive on the QuantiFERON-TB Gold In-Tube (QFT-GIT) test, and

• 57 percent were positive on the T-SPOT.TB test.⁶

Shean reported that active TB was found in five of the health care workers.

Friedland spoke about important work on transmission performed at the Church of Scotland Hospital (CoSH) in Tugela Ferry. CoSH is a rural district hospital with 350 beds, including male and female medical and TB wards of about 40 beds each. Both inpatient and outpatient facilities are overcrowded in a congregate area conducive to airborne transmission of TB. The local community has a 17.5 percent prevalence of HIV-positive adults and a TB–HIV coinfection rate of more than 90 percent. The TB case rate is 1,100 per 100,000 population per year.

CoSH has implemented infection control measures in each of the three WHO/CDC categories (administrative, environmental, and personal protection) as part of its infection control program:

• Administrative

  • A dedicated infection control officer was appointed.

  • Cough officers were placed in all ambulatory care areas to separate out suspected TB cases.

  • A TB infection control policy was ratified.

  • Reductions in admissions and lengths of stay were accomplished.

• Environmental

  • Isolation wards were created for male and female MDR TB patients.

  • Extractor fans were installed.

  • Natural ventilation was assessed, and an open-window policy was implemented.

  • The DOTS office was moved to the periphery of the hospital, and all patients were offered voluntary counseling and testing.

  • A well-ventilated antiretroviral therapy clinic was relocated to the periphery of the hospital, and all patients were screened by a cough officer.

  • Remediable environmental flaws were evaluated.

• Personal protection

  • Staff surveys were conducted regarding infection control knowledge, attitudes, and practices.

  • TB education sessions were held.

  • Regular staff screenings were performed by the occupational health officer.

  • The use of N95 respirators was introduced, along with fit testing and staff education regarding fit checks.

  • Voluntary counseling and testing was promoted for staff and inpatients, with provision for antiretroviral therapy and the option for health care workers to transfer discreetly to lower-risk areas.

In addition, a series of simple and inexpensive evaluations was performed, including

• assessment of lengths of stay and admissions by review of TB ward/hospital registers,

• cough officer program results,

• unannounced audits of ventilation and open windows,

• unannounced audits of N95 respirator use by health care workers,

• voluntary counseling and testing of staff and patients, and

• screening of health care workers for TB.

Outcome measures in inpatient TB wards showed dramatic results. In 2005, 88 percent of patients screened were culture positive, and 64 percent of the patients screened (16 of 25 patients) had MDR or XDR TB. In 2009 in the same wards, 59.3 percent of patients screened were culture positive, and only 7.4 percent of the patients screened (2 of 27 patients) had MDR or XDR TB. Thus, the effective implementation of administrative, environmental, and personal protection infection control measures created a hospital environment in which a decline in drug-resistant TB cases was possible and ultimately became reality.

A mathematical model was used to evaluate the infection control program by estimating the number of XDR TB cases that could be averted by the implementation of different combinations of infection control strategies (Basu et al., 2007). The model showed that using combinations of the available infection control measures over the period from 2007 to 2012 could result in preventing 48 percent of new cases. The model also showed that of the estimated 1,300 new cases of XDR TB, 3 percent would be among health care workers, and that 75 percent of those new cases among health care workers could be prevented through the use of natural ventilation; the use of N95 respirators; and voluntary counseling and testing for HIV, with redeployment of HIV-positive staff. The model showed further that the provision of antiretrovirals to eligible HIV-positive individuals would avert 24 percent of new XDR TB cases. Of note, involuntary confinement would increase transmission by 3 percent by bringing together individuals with strains of differing resistance.

In the workshop discussions of the implementation of infection control protocols, it was noted that available transmission control strategies and technologies, such as early diagnosis, triage and separation, natural ventilation, UV masks on patients, and treatment, need to be implemented. Once implemented, those interventions need to be evaluated to determine their efficacy. Monitoring and further direction by infection control practitioners will be required in areas where the interventions have not been successful to improve the measures and reinforce the importance of infection control. In addition, it will be necessary to develop novel effective transmission control interventions.

Another theme of the discussion was related to behavior change and best practices for health care workers. It was noted that infection control training for nurses is lacking, as is systemic training capacity. Large-scale, long-term training in infection control is required for the South African context. Other interventions that have shown some success include such simple measures as using checklists and having an infection control practi-

tioner. Finally, discussants suggested that there needs to be a general shift away from the concept of centralized infection control and toward a model in which each person in a health care unit is considered to be individually responsible for infection control.

In his presentation, Marra suggested that, although addressing the issue of transmission in the hospital is of critical importance, community-based interventions to combat the transmission of MDR and XDR TB are needed as well. According to Marra, contact tracing, which includes education in and awareness of infection control, should be part of the normal TB control program, and infection control should be part of every intervention. In Tugela Ferry, for example, health care workers have been visiting the households of MDR and XDR TB index cases to offer TB screening and education about the disease and its prevention, demonstrating the feasibility of household contact investigation.⁹ The household contact investigation and infection control strategy applied in Tugela Ferry, which started in 2006, involves

• visiting the homes of all index MDR and XDR TB patients;

• taking the global positioning system (GPS) coordinates of each home; and

• screening all household contacts by obtaining a TB symptom history, collecting sputum from all sputum producers, bringing all suspected TB cases to CoSH for evaluation by a medical doctor,

and having a chest x-ray taken (currently, 4,000 contacts are in a central database).

Most of the MDR and XDR TB cases were situated in Msinga, the subdistrict in which Tugela Ferry is located, in close proximity to the hospital, while fewer cases were found in the surrounding subdistricts. The greatest number of cases was found in 2007. The study evaluated but did not find associations with the terrain, access to health care, and the number of people sharing a home.

According to Marra, the infection control program in Tugela Ferry, the household contact investigation of MDR and XDR TB families, and the strengthening of the normal TB program have met with considerable success. The cure rate has risen from 70 to 82 percent, and a default rate that had been around 20 percent is now close to zero. No patients have been lost to follow-up. Marra stressed that the introduction of rapid diagnostic tests to reduce the infectious period is essential, and community spread of MDR and XDR TB needs to be further investigated.

In the discussion following Marra’s presentation, Jason Farley, Johns Hopkins University School of Nursing and Johns Hopkins AIDS Service, noted that infection control in the community requires a tremendous investment of resources and research. Moreover, it depends not only on the health care infrastructure but also on the infrastructure of individual households. To prepare patients for discharge from the hospital, this household infrastructure must be assessed and appropriate education ensured. Shean said that in the Western Cape, such cases are dealt with by a review board that makes decisions based on assessments of the home and the community by social workers, with the involvement of families to ensure treatment literacy. UV lights and mechanical ventilation are sometimes installed in homes. Friedland suggested that various practical approaches can be applied, such as keeping patients outdoors as much as possible, separating TB patient sleeping quarters, and implementing cough etiquette.

In her presentation, Shean observed that sub-Saharan Africa has 11 percent of the world’s population, 25 percent of the global burden of disease, and just 3 percent of the world’s health care workers. This limited number of health care workers presents one of the greatest challenges to health care systems in the region. Furthermore, the situation is becoming worse. The ratio of registered nurses to patients in South Africa decreased from 120 per 100,000 population in 2000 to 109 per 100,000 in 2005.

The HIV crisis is rapidly depleting current health care workforces, with an average of 15.7 percent of nurses in four provinces of South Africa being HIV-positive (Shisana et al., 2003). Moreover, drug-resistant TB has placed a significant additional strain on the overwhelmed TB control programs in high-burden areas. And according to Shean, the national policy of admitting all XDR TB patients to the hospital until they are smear negative has added to the challenge of providing access to affordable and appropriate infection control in health care facilities.

Very few data are available on the incidence and treatment-related outcomes of XDR TB in health care workers. A retrospective review of case records of passively detected XDR TB cases diagnosed between 1992 and 2008 from three provinces in South Africa found 36 health care workers who were sputum culture positive for XDR TB (Jarand et al., 2010). One case was diagnosed in 1996, with an incremental increase until 2008, when 14 cases were diagnosed, 80 percent of whom were on antiretrovirals. Ten of the health care workers had bilateral disease, and the mean number of drugs in the XDR TB treatment to which resistance was found was six. The majority of the health care workers worked in district hospitals. Thirty-four percent of the health care workers died, with a median of 3.7 months from diagnosis to death. These numbers reveal that XDR TB is an important risk for health care workers globally, particularly those who work in high-burden areas. Factors contributing to the risk of infection include

• the concentration of organisms in the source case (associated with severity of disease);

• the duration of exposure to air contaminated with tubercle bacilli;

• the type of health care facility; and

• comorbid diseases, such as HIV/AIDS, diabetes, or cancer.

A personal challenge facing health care workers is the fear of being diagnosed with TB or HIV infection, which would lead to marginalization by colleagues and the possible loss of one’s job. Notwithstanding this challenge, drug-resistant TB should be considered in any health care worker suspected of TB, irrespective of HIV status, and implementation of infection control measures and rapid diagnostic testing for all health care workers suspected of TB are necessary. At the same time, health care institutions must support and protect health care workers by helping them cope with stressful situations and by acknowledging their sometimes dangerous and difficult work conditions, said Shean. Workable plans for emergency situations need to be in place. Employers are responsible for providing a safe working environment, which includes necessary and sufficient information, human resources, protective equipment, and supplies.

Lesotho is a land-locked mountainous country surrounded by South Africa. Its population is 1.8 million, with an HIV prevalence rate of 23.2 percent and a TB incidence rate of 544 per 100,000 population. The number of people living with HIV is 270,000; 22,000 are on antiretroviral therapy, and 81,000 still need it. The TB–HIV coinfection rate is about 64 percent, and the number of TB deaths is 107 per 100,000 population per year. The unemployment rate is 45 percent, and 56 percent of the population lives on less than $2 per day. In 2007 the textile and apparel industry employed about 46,000 workers in three districts of the country. In addition, Lesotho contributes more than 50,000 migrant workers to the South African mining industry, and thousands of migrant workers stay and work illegally in South Africa.

Thotolo stressed that in fighting TB, it is necessary to address the cross-border movement of migrant workers. At times, these workers visit their families in Lesotho without informing their clinics or doctors, thereby putting family members at risk. Illegal workers generally do not present at clinics upon realizing they are ill; instead, they treat themselves with traditional remedies and return to their homes when the illness is at an advanced stage. Some patients claim that they are refused TB treatment in South Africa because they are unable to present the necessary identity documents. Factors that contribute to the spread of drug-resistant TB in these workers include the following:

• Unauthorized home visits by patients sometimes interrupt and undermine treatment.

• Houses are overcrowded and have poor ventilation.

• Infection control is inadequate, particularly in relation to cough etiquette in public transport and crowded public places.

• Medical TB treatment is combined with remedies offered by traditional healers.

• Loss of income leads to poverty and malnutrition among the entire household.

According to Thotolo, transborder control of drug-resistant TB can be improved by

• establishing better systems for the prevention, diagnosis, and treatment of TB and HIV for mine workers, ex-miners, and their families in Lesotho and South Africa;

• establishing linkages between TB and HIV programs in and between Lesotho and South Africa, as well as between the public and private sectors, including traditional healers;

• building capacity among individuals and communities affected by the cross-border TB/HIV epidemic, including treatment literacy training; and

• addressing the legal, human rights, and socioeconomic issues affecting TB and HIV control among mine workers, ex-miners, and their families through an advocacy program.

The Department of Health’s Tuberculosis Strategic Plan for South Africa, 2007–2011 states that the South African gold mining industry probably has the highest incidence of TB in the world—between 3,000 and 7,000 per 100,000 population per year (DoH, 2007). The mining companies need to take the burden of TB in mine workers seriously and invest in the fight against the disease. In particular, the “three I’s”—intensified TB case finding, isoniazid preventive therapy, and infection control (see Chapter 3)—need to be advocated on behalf of migrant populations.

In May 2008, the AIDS and Rights Alliance for Southern Africa (ARASA) convened a meeting in Johannesburg together with mining companies, trade unions, the ministries of health of both countries, and activist groups to discuss the prevention and treatment of TB. An outcome of the meeting was a recommendation for collaboration between the ministry of health in Lesotho and the mining companies in South Africa so that when patients are referred to TB facilities in Lesotho, proper documentation concerning previous treatment will be made available and special procedures followed.

Warren observed that in South Africa, the mines have a far more efficient TB control policy than do communities. Among miners, all patients are diagnosed, hospitalized, and treated, and DOTS management is exceptionally good. Despite this infrastructure, and even though a high proportion of patients are cured, outbreaks of drug-resistant TB are still observed. These data have been interpreted to imply that the treatment of patients with standardized regimens for MDR TB inadvertently leads to the emergence of XDR TB. It could not, however, be discerned whether transmission occurred in the household, in the compounds where the miners lived, in the hospitals, or in the mines.