Portal de Periódicos da CAPES

Challenges in delivering antiretroviral treatment in resource poor countries

2002; Lippincott Williams & Wilkins; Volume: 16; Linguagem: Inglês

10.1097/00002030-200216004-00024

1473-5571

Mina C. Hosseinipour, Peter N. Kazembe, Ian Sanne, Charles M. van der Horst,

Adolescent Sexual and Reproductive Health

Challenges in delivering antiretroviral treatment in resource poor settings While the AIDS epidemic has ravaged sub-Saharan Africa and other developing countries, industrialized nations have managed to reduce mortality and morbidity secondary to the disease with the introduction of triple drug therapy including protease inhibitors in 1996 [1]. As the deaths secondary to AIDS climb to nearly 3 million people per year, economically challenged countries have been devastated [2]. Worldwide twelve countries have a prevalence of HIV exceeding 10% [2]. The cumulative effects of AIDS compromise existing health programs, food security and economic welfare of these countries which has led to mounting pressure to provide universal antiretroviral treatment. Even with the provision of medications at a dramatically decreased cost, the lack of established clinical infrastructure, negative social stigma and paucity of social services in these impoverished countries remain significant obstacles to delivering effective antiretroviral therapy (ART). Citing these barriers and the considerable cost of treatment, many advocate prioritizing prevention activities over treatment [3–5]. Prevention should remain the cornerstone of AIDS efforts as education, condom promotion and lifestyle modification have successfully curbed epidemics in several countries including Uganda, Thailand and Senegal [6]. However, prevention efforts should not preclude antiretroviral therapy. Access to life-saving treatment can only serve to enhance prevention, improve clinical infrastructure and prevent economic collapse in countries with a high prevalence. Theoretically treatment could also decrease the sexual transmission of HIV by lowering the concentration of HIV in genital secretions [7–9]. Declining incidence rates in countries such as Brazil, where antiretroviral therapy is widely available, lends indirect evidence to this theory [10,11]. This concept will shortly be tested in a discordant couples trial to be conducted by the National Institutes of Health (MS Cohen, personal communication). The provision of ART to a select few capable of paying in urban centers of excellence in resource limited countries is already underway. Several middle income countries have extended services beyond prevention to the antiretroviral treatment arena. Brazil quickly implemented a government-sponsored program using locally produced off-patent and purchased on-patent ART [10,11]. By providing universal treatment to its infected population they have achieved country wide results comparable to Europe and the United States. Pilot programs have been initiated in several other countries [12–16]. In 1997, UNAIDS initiated an HIV Drug Access program introducing ART into Uganda, Cote d'Ivoire, Chile and Vietnam [17,18]. Médecins sans Frontiéres (MSF) has implemented a novel and very successful strategy at 9 pilot sites in 7 countries including Cameroon, Kenya, Malawi and South Africa [19,20]. With UNAIDS sponsored Accelerating Access Initiative pricing announced by western pharmaceutical companies as well as inexpensive generic manufacturers, several other countries are now developing antiretroviral programs. However, the majority of these programs are limited to urban referral centers and individuals capable of paying, excluding the vast majority of the infected population. The provision of ART under these extremes of poverty poses unique obstacles with complex solutions. The recognition of these barriers, potential solutions and future directions is the objective of this review. General infrastructure The magnitude of infrastructure deficiency is profound, beyond the realms of the imagination of most in the industrialized world. Statistics for Malawi, a small sub-Saharan African country consistently ranked among the world's 10 poorest countries, illuminate the shortcomings for basic human needs independent of health. Despite allocating nearly 15% of government spending to health expenditures, the health care needs are unmet by over 50 million dollars [2]. In Malawi, where the per Capita GNP of less than $200, a mere 5% of households have electricity and only 33% have access to a clean water source within 15 minutes of the home [21]. Access to most simple health care facilities remains limited or non-existent for many, particularly in the rural areas, which encompasses over 80% of the country. This bears consideration when contemplating storage of particular antiretrovirals requiring refrigeration (most protease inhibitors), barriers to communication methods requiring electricity or telephone service, and transportation difficulties from washed-out bridges or non-existent roads. Availability of Voluntary Counseling and Testing (VCT) Prior to implementation of any antiretroviral treatment, knowledge of one's status is paramount. In a randomized controlled trial in Kenya, Tanzania and Trinidad, voluntary counseling and testing was found to reduce HIV risk behaviors both in individuals and couples [22]. Epidemic modeling based on this trial found VCT to be a cost-effective measure in preventing HIV infection. For every 10 VCT tests conducted, it is estimated that one HIV infection will be prevented [23]. While negative stigma, denial and lack of preventive care services for infected individuals remain significant deterrents to voluntary counseling and testing, access to HIV testing remains poor to non-existent to many of the most rural areas. HIV rapid tests, such as the Abbott Determine and the UniGold assays hold promise for reaching these remote rural areas and provide an example of an ideal assay for resource poor countries [24,25]. These use a simple finger stick and provide an answer within 15 minutes, with the additional benefits that non-clinicians or volunteers can perform the test, no venopuncture is required (with its attendant risk of needle stick injuries), and patients can see the results for themselves. These rapid tests have high acceptability by clients [26] and many countries such as Malawi have adopted rapid testing as the primary HIV diagnostic method due to its low cost and ease of use. Acceptance of VCT can be enhanced through the linkage to primary care, mother to child transmission prevention and tuberculosis programs and social support networks. The World Health Organization's ProTEST projects have demonstrated the benefits of integrating VCT and access to primary care, independent of ART [27]. Logic dictates that VCT efforts would be enhanced further when antiretroviral therapy is a component of a care package. Laboratory diagnosis and monitoring Low-cost monitoring of HIV disease and therapy continues to be an issue of importance necessary for the implementation of ART, and investigators across the globe are searching for cheaper ways to measure CD4+ cell counts and viral load [28]. In several publications, guidelines on laboratory assays required for ART have been suggested based on the experience of these medications in the developed world [29–31]. As the majority of this experience has occurred in relatively homogeneous populations, it is unclear whether the guidelines are applicable to a setting with a different environment as well as a different population with possible unique genetic, immunologic and virologic factors. The need for thorough evaluation of proposed guidelines and operations research remains. Problems with clean water, consistent, uninterrupted electricity, transport of specimens from remote areas, appropriately skilled staff and health expenditures limit the ability to perform diagnostic, safety and efficacy laboratory assays. In many referral hospitals throughout these countries the ability to consistently measure liver enzymes and to culture blood and body fluids for routine bacterial pathogens is often absent. The high cost of monitoring patients may prove to be a greater barrier to implementation than drug cost [32]. Possible clinical markers were described by Mekonnen and colleagues [33] who are conducting an ongoing cohort study in Ethiopia. The investigators found that simple markers such as anemia, lymphocyte count < 1500 cells/min3, low body mass index, and HIV symptoms were independent predictors of short-term survival. Using the presence of at least 1 of these markers as a criterion to initiate antiretroviral therapy, 126 patients would have begun treatment compared with 128 patients if international guidelines were used. Of the 116 who would have met both sets of criteria for starting treatment, 78% would have started therapy at the same time by either method. Similarly, in the Chiradzulu district in Malawi, WHO stage 3 or 4 disease or total lymphocyte count < 1500 cells/mm3, was found to have a 95% sensitivity to identify those with CD4 count below 200 cells/mm3, and may allow an appropriate surrogate to CD4 count for treatment initiation [34]. Several groups have described the use of total lymphocyte count (TLC) as a replacement for CD4+ cell count in untreated and treated patients [35–39] correlating a TLC < 1250 cells/mm3, with a CD4+ cell count < 200 cells/mm3 and a TLC < 1900 cells/mm3 with a CD4+ cell count < 350 cells/mm3. One group has shown that using a TLC cut off of 1400 cells or less would identify 73% of patients with CD4 < 200 cells/mm3 and using 1700 cells/mm3 or less would identify 70% of patients with CD4 < 350 cells/mm3 and they propose using TLC as a tool to initiate opportunistic infection prophylaxis [39]. Others suggest using change in TLC as a correlate of change in viral load or CD4 count to monitor ARV therapy [35–36]. In a study conducted in India, when TLC increased during the first 6 months on HAART the probability that the CD4 count had increased was 98% but with a decrease in TLC, the probability of a CD4 decrease was only 40%–64% which persisted over the first 24 months (Anish P. Mahjan, personal communication). A recent report described 266 patients each evaluated at 5 time points [36]. There was a significant correlation between changes in TLC and changes in CD4+ cell count (R = 0.61; P < 0.01). Given the variability of TLC within subjects a single measurement while associated with CD4 count is too imprecise to serve as a reliable surrogate. Multiple measurements at the same time point would decrease that error. There is a need for careful validation of monitoring strategies relying on TLC to avoid emergence of drug resistant strains. Cheaper and simpler manual CD4 assays potentially useful for resource limited countries with a small volume of specimens are in development including Beckman Coulter's Coulter Manual CD4 Kit and Dynal's Dynabeads CD4 [40]. In addition, some of the traditional laboratory companies including Becton Dickinson have dramatically reduced their prices for resource poor countries (van der Horst, personal communication). The international group Afford CD4 have introduced affordable generic reagents for CD4, CD8 and CD45 [41–46]. The group has shown data comparing dual platform absolute CD4+ cell counts using a single-tube, 2-color CD45+/CD4+ protocol (PanLeucogate) with single platform gold standard, volumetric flow Cytometry (Ortho Cytoron), using a full triple tube, 4 color panel, demonstrating excellent correlation (R2 = 0.99). Dual platform CD4 PanLeucogating method, utilizing any standard haematology analyser plus flow cytometer, or the volumetric flow cytometry system that use generic monoclonal antibodies, cost less than $1–2 per assay for reagents alone [41–46]. Most clinicians in resource-poor countries feel that even with the recent dramatic reductions in cost announced by Roche for resource poor countries, viral load will never be cost-effective and that it is not essential for management. Recently released WHO Treatment Guidelines omit the use of viral load [29,47]. Withum and associates have evaluated an ultrasensitive p24 antigen assay for use in the diagnosis of pediatric HIV infection [48]. For Clade B the sensitivity is 97% with 100% specificity in diagnosing pediatric HIV infection making this a much cheaper method for diagnosis of perinatal infection in the United States ($8 per test, vs $50–$70 for PCR testing). However, it is not clear whether the same method will be useful for the HIV clades that are prevalent in Africa and elsewhere with sensitivity ranging from 76% for Clade A to 87% for Clade C (Susan Fiscus, personal communication). The results for subtype C are less consistent and technology is not yet robust enough for wide scale recommendation (Wendy Stephens, personal communication). HIV p24 antigen levels have been shown also to strongly correlate with HIVRNA levels, potentially providing a lower cost alternative for antiretroviral monitoring [49–51]. Careful pricing of assays needs to be conducted and should not simply look at the cost of reagents but also the staff requirements, equipment maintenance and shipping requirements. It is important that the lack of technology in resource poor settings not prevent the introduction of treatment. Most successful tuberculosis programs in resource-poor countries do not monitor liver enzymes. At the Barcelona Conference, Joep Lange, the president of the International AIDS Society said, ‘There is actually very little evidence that laboratory monitoring prevents mortality due to drug toxicity, but there is an awful lot of evidence that not treating symptomatic HIV infection is lethal, and usually not in a nice way; why are we always more concerned about doing harm than about not doing good?’ (Joep Lange, personal communication) Personnel Although provision of ART does not require physicians, many resource poor countries suffer a lack of health care givers at all levels of the hierarchy. Citing a baseline deficiency of health care providers even prior to the AIDS epidemic, the number of qualified providers has become more tenuous as personnel succumb to the disease themselves and the disease burden grows exponentially. HIV disease can be expected in health workers from sexual risk as well as potential nosocomial risk. Nursing and laboratory personnel represent part of the socio-economic group most affected by HIV/AIDS. In a country with 15% adult prevalence, as is now common in sub-Saharan Africa, one can expect between 1.6 and 3.3 percent of a country's trained health personnel will die from AIDS, while productivity losses from the illness are even greater [52]. The cost of training new personnel is significant and anticipated losses from AIDS deaths and illness would increase health care expenditures by nearly 4% [53]. Further, the inability of stretched government budgets to maintain competitive wages and benefit packages for their working staff has further compromised the pool of health care-givers through the exportation of this talent to developed countries or private facilities [54]. Budgetary demands to maintain a workforce can force temporary closure of clinical and technical training facilities or prevent the development of facilities to meet the medical needs of the country. The maintenance of the health care force must be prioritized as ART is contemplated. Analysis of impact of ART in certain sectors of the population reveals ART to be highly cost-effective. While some debate exists concerning prioritizing certain groups to receive ART over others, the argument to treat health care workers is particularly strong. Many companies and institutions are recognizing the importance of treating their employees [55,56, Brian Brink, personal communication]. The maintenance of health for this care giving pool can increase their productivity, preserve funding for the training of additional personnel, and provide an educated population by which messages of ART can be transferred to those seeking medical care. Medical schools and health sciences training facilities must be strengthened to advance the general health systems of developing countries. The addition of qualified midwives, clinical officers and medical officers can extend care to the most remote areas to improve the overall health standard of these countries. For South Africa, the International Monetary Fund has estimated that the introduction of ART will require a three fold increase in medical practitioners. ART will require an even larger pool of clinicians including laboratory and pharmacy support. The implementation of such training can in part be accomplished by partnerships between western universities and developing country universities and hospitals to provide exchange of education and financial support as proposed in the United States Senate Authorization Bill S2649. Training programs, including those sponsored by the Fogarty International Center of the NIH and the World Health Organization already exist to foster this laudable goal and expansion of such programs should continue. Beyond training is the need for there to be salary support for full time physicians that is continuous as well as adequate. Currently many physicians either leave the resource poor countries as there is no salary support or they spend a large amount of time in private practice or working for NGOs to supplement their income. The management of ART is complex and even with simplified algorithms there will be a need for highly trained physicians to lead the expansion of care. Given the shortage of trained doctors and nurses in sub-Saharan Africa, any country-wide scale-up of medical care including antiretroviral therapy would require novel approaches including community-based caregivers, traditional birth attendants, and traditional healers. Already in some countries such as Lesotho and Malawi, Primary Health Care Sisters serve as the prin-ciple health care provider in rural areas. In Africa, the ratio of doctors to population is 1:40 000 but the ratio of traditional healers to population is 1:500. These healers play a key role in delivery of care as many individuals seek care initially among them. Their role must be enhanced to ensure the healer recognizes the role of ART in the treatment of the HIV infected and refers such individuals to medical care. This importance was underscored in a recent study in Malawi, in which a large percentage of tuberculosis patients sought care from the traditional healer, and in some instances were referred to the TB registry by the traditional healer [57]. Traditional healers have been utilized in a controlled clinical trial in a directly observed tuberculosis (TB) treatment program in Hlabisa, South Africa [58]. After two days of training, 25 traditional healers volunteered for the project, in which 53 patients were supervised by healers and 364 by clinics. Overall, 89% of those supervised by healers completed treatment, compared with 67% of those supervised by others (P = .002). The mortality rate among those supervised by healers was 6% compared with 18% for the others (P = .04). Interviews with patients completing treatment said they liked the convenience and short waiting time for the healers, as well as their more empathic caring approach. Some healers conducted regular home visits, particularly in the early days of treatment when the patients were more ill. The role of traditional birth attendants must also be emphasized as most women in rural areas are attended by such individuals. The training of such women in perinatal HIV transmission prevention could promote VCT and acceptance of interventions such as antepartum nevirapine [59]. The organization, Partners in Health, has successfully integrated community health care educators in combating multidrug resistant tuberculosis in Peru and Haiti [60] and likewise has promoted this approach toward HIV [61]. Integration and training of traditional healers, traditional birth attendants and community health educators may serve as a route of identification of individuals requiring antiretroviral therapy, appropriate referrals to clinicians for interventions and ongoing monitoring of adherence in the community. Endemic diseases, nutrition and primary care Various factors unique to resource constrained countries must be taken into consideration as ART is introduced. Nutritional status of many individuals in developing countries is marginal with 35% of children in sub-Saharan Africa exhibiting stunting on growth charts [62]. Food shortages are common and Africans can demonstrate marked fluctuations in body weight based on the season of the year [63,64]. Dependence on staple foods, lack of diversity in diet, and food shortages commonly result in deficiencies of both macronutrients and micronutrients with women and children bearing the most significant ramification [65]. For example, among pregnant women in Malawi, 81 percent demonstrated some form of nutritional deficiency [66]. The impact of ART in this setting is unknown. Medications requiring administration with food or large quantities of water may be problematic. The spectrum of infections in developing countries does not mirror that of the United States. Endemic infections including tuberculosis, bacterial infections such as bacteremia due to non-typhoidal salmonella and Streptococcus pneumoniae, and malaria represent a great percentage of the AIDS mortality in sub-Saharan Africa [67–80]. The success of many government-sponsored tuberculosis programs has been severely jeopardized by the AIDS epidemic as the continued transmission and reactivation of disease propels both epidemics and negatively impacts AIDS mortality [69,70]. Given the poor sanitary conditions of many developing countries, fecal-oral transmission of bacterial and protozoal pathogens promotes increased morbidity and mortality. Non-typhoidal Salmonella bacteremia remains among the chief causes of mortality in developing countries; its association with HIV has clearly been established [71–73]. The lack of adequate quantity or active antibiotics hampers treatment, promoting relapse [71]. Pneumococcal bacteremia is also one of the leading causes of death among HIV infected patients [74]. Some of the endemic infections exacerbate HIV infection. Helminth infections, associated with poor living standards, have been proposed as promoting the epidemic [75]. While the interaction of malaria and HIV is not yet entirely understood, the impact on viral replication has been demonstrated [76]. Incident malaria results in a steep 100-fold increase in viral replication which gradually reduces after treatment [76–78]. HIV impacts on malaria as well with an increased prevalence of malaria in HIV infected women and a decreased response to prophylaxis [79, Peter Kazembe, personal communication]. The use of cotrimoxizole for bacterial prophylaxis and its potential induction of cross-resistance of malaria to pyrimethamine is very concerning [80–82]. HIV care must encompass prophylaxis for tuberculosis and bacterial infections including Salmonella, pneumococcus and malaria [83]. Nonetheless, any guidelines must be first evaluated in the different settings [84–85]. In South Africa the efficacy of cotrimoxizole was limited to patients with a CD4 count less than 200 cells/mm3 unlike Cote d'Ivoire. Africa is not one population or geographic area. It consists of vastly different environments and peoples, each with different characteristics, endemic diseases and resistance patterns. All endemic diseases and their treatments have the potential to impact drug interactions among the antiretroviral medication and the clinical course of HIV. These potential interactions can influence treatment guidelines including the choice of medications, when to initiate therapy and potentially the ultimate success of therapy. Clinical management — unique issues for developing countries Among the first questions which must be answered in the development of ART programs in resource poor countries is when to initiate therapy. While this question remains unanswered in the developed world, the decision may bear even more significance globally as long term costs of medication and monitoring for a large percentage of the population can quickly consume health care budgets. In the United States, guidelines have steadily been modified and now suggest the initiation of ART at a CD4 count of 200 to balance the required life-long use of therapy and potential toxicity of ART [86–88]. The concern, however, is that a delay in initiation of therapy will lead to irrecoverable damage to the immune system [89]. There is no difference in overall mortality of untreated HIV-1. infected patients in historical US and European cohorts compared to African cohorts [90–92]. There is no evidence that HIV patients from resource poor countries will respond less to ART than their wealthy counterparts [93]. Suggestions for initiation of therapy in resource poor countries have considered both available laboratory monitoring capabilities and clinical status and generally suggest later initiation of therapy [29,47]. These guidelines recommend initiation in patients with WHO stage IV disease (clinical AIDS) regardless of CD4+ cell count and in anyone else who has a CD4+ cell count < 200 Cells/mm3 or a total lymphocyte count < 1200 cells/mm3. For children younger than 18 months, they recommend therapy for those with WHO Pediatric stage III disease (clinical AIDS) regardless of CD4+ cell count, and other stages with a CD4+ percentage < 20. The integration of tuberculosis and antiretroviral programs must be considered in terms of both treatment, prophylaxis and ‘when to start’. Waiting for a CD4 count to drop to 200 cells/mm3 before initiation of ART may continue to fan the flames of the tuberculosis epidemic as tuberculosis occurs at the entire range of CD4 counts. In South Africa, the initiation of ART in individuals averted 7.3 cases of TB per 100 patient-years overall including 18.8 averted cases for patients with WHO AIDS stage 3 or 4 disease, 14.2 cases for patients with a CD4 count less than 200 cells/mm3, 10.6 averted cases for patients with CD4 count 200–350 cells/mm3 and 2.3 averted cases for patients with greater than 350 cells/mm3 [94]. In addition to the public health ramifications of preventing further tuberculosis infections with initiation of ART, simultaneous treatment of HIV and tuberculosis must be studied [95]. In Malawi, where Cipla's Triomune (a combination tablet of 200 mg nevirapine, 40 mg stavudine and 150 mg lamivudine) is the primary agent for the government sponsored program, 50% of the 500 patients presenting for ART to date have had a history of active tuberculosis, 36 patients deferred antiretroviral therapy due to active TB treatment and an additional 8 interrupted ART for TB treatment due to the drug interaction between rifampin and nevirapine [unpublished data]. In addition to drug interactions, factors to consider include pill burden, overlapping drug toxicities, paradoxical or immune reconstitution syndromes, and the high morbidity and mortality of patients with active tuberculosis and low CD4 counts. Patients with tuberculosis and CD4 counts less than 100 cells have a high risk of developing opportunistic infections [96]. Simultaneous treatment can result in interruption of therapy in up to 34% of patients due to severe adverse events [96]. As the majority of protease inhibitors and NNRTIs interact with rifampin-containing tuberculosis medications, implementation must proceed cautiously to maximize the treatment of both conditions. Rifampicin is a potent inducer of P450 isoenzyme CYP 3A4 which affects drug levels of NNRTIs and protease inhibitors which are also metabolized by this enzyme. Rifamycins also induce the multi-drug resistance-associated protein (MRP) associated with protease resistance [97]. Likewise, immune reconstitution syndromes to tuberculosis are common in resource poor countries and require antiretroviral modifications [98–101]. Several groups including WHO have made recommendations concerning simultaneous versus staggered treatment of both HIV and tuberculosis. The WHO recommends simultaneous therapy for patients with pulmonary TB and a CD4 count less than 50 cells/mm3 and any patient with extrapulmonary tuberculosis [29]. For other patients, a staggered start is recommended [29,102]. Duration of therapy must be considered as well [102,103]. Clinical management must weigh the options of including expensive anti-tuberculosis agents such as rifabutin, modifying tuberculosis treatment to omit rifampin, or choosing highly potent antiretroviral agents without drug interactions. While triple NRTIs regimens are devoid of drug interactions with TB medications, the question of their potency remains among those with high concentration of plasma HIVRNA that may occur more commonly in developing countries [104,105]. While clinical markers and absolute lymphocyte counts may be sufficient for decision-making with respect to initiation of therapy, viral load monitoring will be more important for deciding when to switch or discontinue in treatment failure. Virologic failure precedes measurable clinical and immunologic decline and intercurrent illnesses complicate clinical measures. In a prospective cohort study following HIV negative and positive individuals in rural Uganda, clinical symptoms suggestive of disease progression such as weight loss, diarrhea, and bacterial infections were extremely common even in HIV negative patients [106]. While tuberculosis, shingles and oral candidiasis were better predictors of HIV disease progression, reliance on purely clinical measures could result in either terminating effective antiretroviral therapy or delaying antiretroviral adjustment allowing the accumulation of resistance mutations. The new WHO Guidelines are purposely vague as to when to switch medications as the panelists could not agree on an approach [29]. It is obvious that any severe toxicity would require switching the offending agent. Poor adherence to a particular complicated regimen might also prompt a switc