Declining trends in early warning indicators for HIV drug resistance in Cameroon from 2008–2010: lessons and challenges for low-resource settings

Low- and middle-income countries had just over 8 million people receiving highly active antiretroviral therapy (HAART) by end of 2011, representing 54% [50–60%] coverage of eligible patients based on World Health Organisation (WHO) guidelines (CD4≤350 cells/μl) [1]. As compared to 2010 (<6 million) and 2003 (400,000), this coverage has been greatly favored by the rapid scale-up of antiretroviral therapy (ART), with sub-Saharan Africa being the main beneficiary (6.2 million people were receiving antiretroviral therapy in 2011, up from just 100,000 in 2003) [1, 2]. In Cameroon, these efforts led to reductions in AIDS-associated morbidity and mortality, and a relative decrease in HIV prevalence (from 5.5% in 2004 to 4.3% in 2011), with close to 44.5% eligible patients on ART [3, 4]. Furthermore, the number of ART clinics and of treated patients has increased overtime: 5 clinics for 116 (1.3%) eligible patients in 2002, to 145 clinics for 89,455 (36%) eligible patients in 2010 [4–6]. Since scale-up of ART is known to be associated with a high risk of HIV drug resistance (HIVDR), strategies to combat HIVDR are of public health priority in Cameroon [6–8]. Most importantly, with limited access to reference laboratory equipment (CD4 count, HIV viral load, HIVDR testing) and the use of drugs with low genetic barrier for resistance, HIVDR population-based surveillance and prevention are recommended to ensure long term efficacy of treatment guidelines [8–11]. As response to this crucial need, a national HIVDR working group (HIVDRWG) was created based on the WHO global HIVDR prevention and assessment strategy which includes the laboratory-based surveys of transmitted and acquired HIVDR, and monitoring of HIVDR early warning indicators (EWIs); with the guidelines being recently revised and updated in 2012 [12–17]. Surveys in Namibia, Malawi, South-Africa, and other African settings [18–24]; in Central American and Caribbean countries, as well as in Asia and Oceania (Papua New Guinea) [23–31], also supported the use of EWIs to evaluate the risk of ART failure and HIVDR emergence. Such evaluations use existing clinic- and pharmacy-based data on ART prescribing at the ART clinic [12, 15, 16]. Six strongly recommended and two optional EWIs are proposed by the WHO, among which ≥4 feasible EWIs should be chosen for an effective ART programme evaluation [15]. In this prospect, we aimed to evaluate the levels and trends of five WHO-strongly recommended EWIs, in order to identify potential strengths and weaknesses (i.e. gaps in service delivery that might inform policy changes to improve performance) of the national ART program and to target appropriate interventions that can optimize care and potentially minimize the risk of emerging resistant patterns that could compromise the commonly used antiretrovirals in the country. Of note, results presented in this study are based on WHO HIV drug resistance early warning indicator guidance [15], and not upon the updated 2012 guidance as presented in the latest WHO HIV drug resistance early warning indicator meeting report [16].

The increasing need to control HIVDR in low-resource settings (LRS) requires the implementation of efficient and rapid approaches to limit the emergence of preventable HIVDR, among which a population-based survey, such as EWI evaluations, could yield relevant evidences based-recommendations for best clinical practices. Overall, this is a paper on a topic of high importance in HIV medicine, transmitted drug resistance, as well as the methods for detecting it in a LRS where genotyping is not available, and remains challenging and clinically important issues to all the stakeholders who are involved in the global scale up of ART, particularly in Africa.

Physicians’ prescribing practices (EWI₁) remained nationwide in conformity with the national guidelines. Interestingly, even sites with the 98-99% performances for EWI₁ had no cases of dual- or mono-ARV prescriptions, which are more likely to select for drug resistant HIV rather than an inappropriate prescribing of drugs such as boosted protease inhibitors, abacavir, or tenofovir (which may not be appearing on several guidelines) which likely have less significance related to HIVDR. This good practice was mainly due to the use of standardized ART regimens and the institution of a consortium agreement (therapeutic committee) for every treatment initiation. This is a weekly consortium body made up of health staff all trained in HIV/AIDS management (medical doctors, nurses, medical biologists/laboratory technicians, medical counselors, pharmacists/pharmacy clerks, data manager and/or community relay agents; a practice which merits to be encouraged. Globally, 75% of clinics monitored worldwide met the WHO-recommended target on prescribing practices [16]. Specifically, other African countries (Malawi, Namibia, and others) also reported good practices [18–24], while poor performances were observed out of Africa (Central American, Caribbean, Asian, Oceanic and Western Pacific: from 38 to 75%) [23–31]. Lack of evaluation overtime in the above mentioned studies could not enable in-depth comparison to ours. Still, the use of a therapeutic committee as well as continuous training may be vital practices to maintain and/or improve good ARV prescriptions.

The high and increasing rates of lost-to-follow-up are indicative of a growing rate of patients harboring potential drug resistant viruses within the national context, which in turns supports a growing risk of HIVDR emergence in Cameroon. These would have been favored by the observed heavy workload, and also probably due to termination of working contracts for community relay agents (CRA). Thus, task shifting (from medical doctors to nurses), and probably community empowerment (by allocating resources to CRAs), as well as free consultations and reduced laboratory costs in such poor settings, may help in reducing the rate of missing patients. Indeed, the numbers of required health care workers to provide ART in LRS (1–2 physicians/1000 patients, 2–7 nurses/1000 patients, <1-3 pharmacy staff/1000 patients, and wider ranges for other health-allied staffs) estimated by Hirschhorn et al. [32], was based on the WHO-3by5 target (i.e. treatment of 5 million patients by 2005), and currently need to be revised in order to match the growing ART coverage (presently >8 millions) [1]. With restriction to the WHO-protocol, a direct evidence-based evaluation has not yet been conducted to investigate on the real impact of CRA’s disengagement on lost to follow-up. Globally, 69% of clinics monitored worldwide met the WHO-recommended target [16]. Particularly, in other African countries, lost-to-follow-up is also challenging (40 to 75%) [18–24], whereas out of Africa higher performances were recorded (54 to 100%) [23–31]. Furthermore, the impact of distance to the clinic, waiting time prior to medical consultation, stigma and patient educational level, are factors that may help in reducing lost patients [33]. Indeed, ART uptake has been negatively associated with distance from the nearest primary healthcare [34, 35], thus indicating a possible need for creating new ART clinics to foster adherence.

The gradual poor retention of patients on first line ART after 12 months ART may be indicating a rapid switch to second line regimens that could be explained by numerous other factors (wider availability of HIV Viral Load testing, greater clinician experience/awareness on identifying treatment failure, a preexisting drug resistant mutation, etc.). Further studies are therefore needed to determine the time-to-treatment, to implement measures toward long term efficacy of first line ART and to limit events of inappropriate switch to second line ART in these settings. Furthermore, failure to be on first line therapy one year after initiation might in fact not necessarily be a negative indicator, since it might indicate a correct and necessary treatment switch to second line. This observation makes questionable the overall utility of EWI₃, which would likely need to be revised for better utility in future. Globally, 67% of clinics monitored worldwide met the WHO-recommended target [16]. More interestingly, our performances in 2008 (70%) were similar to those found in other African settings (among which Malawi: 53% and Namibia: 67%) [18–24], while higher performances (62-90%) were recorded out of Africa [23–31]. Since first line ARV drugs (≈$100/patient/year) are about 4 times less costly than second line, an effective retention on first line is economical [36] and may regulate the current fast switch to second line ARV (2% of total patients on ART in 2006, to 13% of total patients on ART in 2009) in Cameroon [4]. These analyses are consistent with the different HIVDR rates to NRTI and NNRTI in 2003 (before ART scale-up: 5.6%M184V and 6.1%Y181C) and thereafter (following ART scale-up: 16.3%M184V and 63.7%Y181C) [37]. Despite the higher rate of transmitted HIVDR in Latin America (12-20%), sub-Saharan Africa, though with only 4-9% of transmitted HIVDR, is experiencing a rapid scale-up of ART associated with an increasing rate of transmitted HIVDR (38% increasing risk of HIVDR per year): this important fact urgently needs a regular HIVDR surveillance system [14, 16, 38].

The overall delay in drug pick-up strongly supports a national revision of the community engagement strategy to support patient adherence to ART programmes. This poor/decreasing performance in drug pick-up (EWI₄), together with EWI_{2 and 3,} clearly placed patient non-adherence as the main factor with high risks of HIVDR development and spread within the community. However, it should be noted that EWI₄ itself seems problematic and may not be entirely practical to address adherence issues. Indeed, in our setting, clients may present one-to-two day(s) after their required drug pick-up appointment date and still maintain 100% ART adherent, due to availability of remaining pills from the previous appointment. Thus, as for EWI₃, the overall utility of EWI₄ may also be brought into question, and would likely require expert review or reconsideration. Thus, the low scores found in our study are partly explained by the rigidity of this indicator, which has further being recently modified in the 2012 HIVDR report [16]. Globally, 70% of clinics monitored worldwide met the WHO-recommended target for timely drug keep-up [16]. Detail analysis also reports lower timely drug pick-ups in other African settings (between 17% and 41%) [18–24], against settings out of Africa (73-100%) [23–31]. Thus, in African AIDS programmes, issues such as “time spent by patient for pharmacy service”, “service quality rendered by pharmacy staff”, “pharmacy localization”, may be partly explained by delays in drug pick-up. As suggested by El-Khatib et al. in 2011, adherence to drug-refill (pill count) may also be a useful EWI of virologic and immunologic failure on first-line ART in African settings [23].

The gradual poor performances in drug supply present EWI₅ as the main programmatic setback in the national ART performance. Despite the provision of alternative ART (by replacing shortage of efavirenz with nevirapine or lopinavir/ritonavir) in case of drug discontinuity, stock outs still negatively impact patient adherence with non-negligible risks of HIVDR due to potential suboptimal drug levels. Drug supply machinery should be urgently revised. Our findings could be strengthened by conducting further investigations to as to why these stock outs are occurring; targeting specifically the supply lines, trade agreements specific to Cameroon, land versus air versus sea route issues, or in-country distribution network problems. Globally, 65% of clinics monitored worldwide met the WHO-recommended target for drug supply. Geographically, the procurement systems were successful (100%) in Malawi (cross-sectional study) [18], and with an increasing performance in the Caribbean (31.3%-94%, from 2007–2009) [23, 24]. Aggregated data showed poorer performance in sub-Saharan Africa (42%-47%), as compared to non-African countries (32-90%) [20–31]. Thus, African settings are more in need of further investigations to improve their drug supply system.

The extended use of ART in developing countries warrants a thorough understanding of all factors contributing to the success of national ART programmes. In the Cameroonian setting, patient adherence, drug stock outs, a potential community disengagement and a probable heavy workload, appear as major factors favoring an increasing risk of HIVDR emergence.

Use of EWI is an effective approach to identify both patient and programmatic factors favoring the risks of HIVDR at the clinics and at the national levels. With its simplicity, low cost, and rapid evidence-based policy-making, such surveys would be useful for a successful scale-up of ART in LRS, which in turns contribute to the global HIVDR prevention and surveillance strategy.