Suboptimal Antiretroviral Drug Levels and Virologic Failures among PLHIV at a Rural Referral Hospital in South Western Uganda: A Descriptive Cross-sectional Study.

Background: Achieving favorable HIV treatment outcomes is a major challenge, particularly due to non-adherence and consequent sub-therapeutic plasma antiretroviral drug levels. This is often complicated by the development of resistant strains due to mutations. Monitoring antiretroviral drug levels in the blood of patients enrolled on ART can reveal if levels are too high, enough, or too low. High levels may lead to dose-dependent side effects and sub-therapeutic levels could promote treatment failure and resistance. In Uganda, as part of routine HIV care, plasma antiretroviral drug level is estimated indirectly by clinic-based pill counts and patient self-reported adherence, which give no evidence of ingested medication. This study aimed at exploring steady-state nevirapine and efavirenz drug levels in HIV patients accessing ART at a rural referral hospital in South Western Uganda. Methods: This study was nested into a randomized clinical trial that evaluated the effect of Artemisia annua L. and Moringa oleifera on immunological response and viral load among persons living with HIV (PLHIV). In the parent study, 250 HIV-infected patients with continued immunologic suppression (CD4 count < 350cells/µL) despite a minimum of one-year on ART were enrolled. Out of 250 clinical trial participants, 95 were randomly selected for steady-state efavirenz and nevirapine plasma concentration sampling having taken the last at bedtime. Additionally, CD4 count, HIV load, liver, and renal function tests were determined. Participants were also interviewed for adherence, and factors that affect blood drug levels. Results: Of the 95 participants sampled, 67 (71%) and 28 (30%) were on efavirenz or nevirapine based ART respectively. The median viral load for participants on the efavirenz regimen was 490 copies/mL (IQR 116, 1900) while that for the participants on the nevirapine regimen was 500 copies/mL (IQR 137, 1270). The median plasma level for the participants on the nevirapine regimen (6.56 mg/L IQR 4.50, 9.80) was higher than that for the participants on the efavirenz regiment (2.54 mg/L IQR 1.47, 5.12). The prevalence of virologic failure among participants sampled on the efavirenz regimen was higher (37%) compared to that of the participants on the nevirapine regimen (21%). 30% of all plasma samples tested had sub-therapeutic levels of either efavirenz or nevirapine, including 2% in which no drugs were detected. Conclusion: Periodic therapeutic drug monitoring should be incorporated as one of the components in the monitoring of ART adherence to ensure adequate blood levels among adults accessing ART at MRRH, SW Uganda.


Introduction
Over the past four decades, there has been an unparalleled effort to provide access to antiretroviral therapy (ART) for people living with HIV (PLWH) in sub-Saharan Africa, the region with the highest HIV burden [1]. Despite the signi cant reduction in mortality among PLHIV receiving combination ART, 16% of the patients fail to achieve a sustained virological [2], and immunological response to therapy [1,3]. This could be a result of many factors including poor adherence leading to sub-optimal drug levels and viral mutations among others.
Multiple approaches exist for measuring adherence to ART [4,5]. Self-report of adherence typically overestimates adherence as it may be in uenced by social desirability and recall biases. Clinic-based pill counts may be inaccurate if children or caregivers remove and throw away extra medication to appear more adherent. Moreover, patients may not remember to bring their medication containers to appointments, and bottle openings do not always re ect medication ingestion [3,5]. Achieving favorable HIV treatment outcomes is a major challenge, particularly due to non-adherence and the development of strains harboring resistance-associated mutations. The standard approach for monitoring treatment outcomes in patients on ART depends on the measurement of HIV-load over time [3,6]. According to the WHO guidelines, virologic failure is observed when patients sustain a viral load > 1000 copies/mL after 6 to 12months of ART. The persistent high viral load in most cases is due to non-adherence [3].
Monitoring the levels of antiretroviral drugs in the blood of patients can reveal if levels are too high or too low. High levels may lead to dose-dependent side effects while low levels may not prevent the virus from multiplying. Prevention of viral replication is important for the immune system to recover and to ght opportunistic diseases [6]. There is evidence that monitoring of plasma drug concentrations can be a valuable tool in the treatment of HIV-1 [8,9].
Retrospective data from several cohorts propose an association between nevirapine concentration and a rapid-onset and long-lasting virologic response to ARV therapy. Resistance to nevirapine develops rapidly when the drug is administered in suboptimal regimens, and the emergence of the highly drug-resistant virus has been observed 4 weeks after initiation of monotherapy [6,10]. Efavirenz concentrations have data supporting a correlation with virologic outcomes. In a subgroup analysis in the 2NN study, not having an efavirenz trough plasma concentration of less than 1.1 mg/L had an 89% negative predictive value. That is, participants achieving plasma EFV concentrations above 1.1 mg/L had an 89% likelihood of not failing virologically [6,11]. This study aimed to assess random ARV plasma levels, viral load, CD4 count, liver, and kidney function among PLHIV enrolled on and using efavirenz and nevirapine containing regimens for at least one year.

Study design
This sub-study is part of an ongoing randomized clinical trial with three study arms aiming at documenting the effect of Artemisia annua L. and Moringa oleifera on immunological response and viral load in PLWH whose CD4 cell counts remain low despite being on ART for one year. This descriptive cross-sectional sub-study aimed to evaluate random ART drug levels among patients participating in the parent study.

Patient enrolment
From January 2018 to June 2019, the main study had enrolled 250 HIV-infected patients aged 18 years and above with continued immunologic suppression (CD4 count < 350cells/µL) despite a minimum of one-year on efavirenz 600 mg a day in or nevirapine 200 mg twice a day in combination with other antiretroviral agents. The study site was the HIV clinic at Mbarara Regional Referral Hospital (MRRH) located in South Western Uganda (SWU). From the total enrolled participants, 95 participants, 67 on efavirenz, and 28 on nevirapine regimens were randomly selected at enrolment for steady-state efavirenz and nevirapine plasma concentration measurement having taken the last dose at bedtime. Blood samples were also collected from the participants for CD4 count, HIV load, liver, and renal function determination. The participants were asked to state if they were adhering to treatment before enrolment into this study. Participants were also interviewed for factors that affect blood drug levels.

Baseline data collection
Questionnaires were administered to collect socio-demographic data and treatment information including factors that affect blood drug levels. Blood samples for viral-load testing and CD4 count determination, renal, and liver function tests at baseline were drawn from each participant.
We de ned virologic failure as a single viral load measure of > 1000 copies/mL at enrolment [12]. Liver toxicity was de ned as laboratory abnormalities above the normal range. From each study participant single drug concentration blood sample was collected. Samples were labeled for either efavirenz or nevirapine concentration analysis depending on the donor ART history. All samples were centrifuged at 3000 rpm for 10 minutes. Plasma was separated and stored at 80 0 C. Drug quanti cation assays were performed at the Department of Pharmacology and Therapeutics, School of Biomedical Sciences, College of Health Sciences Makerere University Kampala-Uganda. Viral load, CD4 count, renal and liver function tests at were done at Devine Mercy Laboratory Mbarara-Uganda.

Drug bioanalysis
Plasma efavirenz and nevirapine were determined by reverse-phase HPLC with ultraviolet (UV) detection as previously described by [13,14]. The HPLC machine used consisted of a system controller (model SCL-10AVP), a pump (model LC-10ATVP), an auto-injector (model SIL-10ADVP), and a spectrophotometric UV-vis detector (model SPD-10AVP) all supplied by Shimadzu co-operation Kyoto Japan.

Analysis of baseline data
Data collected were entered into an excel sheet version 2016 and was exported to STATA version 13. Descriptive statistics were used to summarise the data. Fisher's exact test was used to determine the association between alcohol use and CD4 count, viral load as well as efavirenz and nevirapine blood levels.

Results
Of the 95 participants whose random nevirapine and efavirenz levels were analyzed, 31 (33%) were male and 64 (67%) were female, with 67 (71%) patients on the efavirenz regimen and 28 (30%) patients on the nevirapine regimen. The age of the participants ranged from 16 to 66 years, with the most representative age group ranging from 16 to 35 years old (Table 1). The median viral load for participants on the efavirenz regimen was 490 copies/mL (IQR 116, 1900) while that for the participants on the nevirapine regimen was 500 copies/mL (IQR 137, 1270 Using a threshold of 3 mg/L for optimal therapeutic nevirapine level [15,16] and 2 mg/L for optimal therapeutic efavirenz level [17], 30% (28/95) of all plasma samples tested had sub-therapeutic levels of either efavirenz or nevirapine, including 2% (2/95) in which no drugs were detected. More females than men had drug concentrations above the threshold i.e. for efavirenz plasma levels was 28 (65%) versus 15 (35%), while for nevirapine plasma level was 18 (75%) Versus 6 (25%) respectively (Table 3).
Liver toxicity (elevated liver enzymes) was observed in 89% (25/28) of the participants on the nevirapine regimen with concentrations higher than 3 mg/L while only 45% (30/67) of the participants on the efavirenz regimen with a concentration above 2 mg/L had liver toxicity. There were no statistically signi cant differences in CD4 count (p = 0.172), Viral load (p = 0.239), efavirenz level (p = 0.955) and nevirapine level (p = 0.472) between alcohol consumers and non-alcohol consumers although generally, participants who did not drink alcohol had more CD4 counts and lower viral loads than those who consumed alcohol (Table 4).

Discussion
Clinicians are often confronted with treatment failure or side-effects and need methods to evaluate drug exposure in patients [11]. Adherence to antiretroviral medications prescribed for the treatment of HIV is central to the effective management of the disease and can predict plasma drug levels in HIV patients who are on ART if adherence is high. Efavirenz and nevirapine are extensively used in Uganda as part of the rst-line therapy, and viral load monitoring for ART management is not always accessible and affordable. Moreover, drug level and toxicity monitoring are not included in the clinical management of HIV patients. Our ndings indicate that 30% of all the patients had sub-therapeutic drug levels of either efavirenz or nevirapine, including those in whom no drugs were detected implying poor adherence to treatments by the patients. These ndings are in agreement with a study carried out in Northwestern Tanzania [18] that also reported 28.3% of patients had sub-therapeutic levels of efavirenz and nevirapine.
In a study conducted in China to assess the relationship between mean efavirenz (EFV) plasma concentration and clinical effect, EFV plasma concentrations above 2 mg/L appeared to suppress HIV replication more effectively than concentrations below 2 mg/L [17]. The same study concluded that the emergence of EFV-resistant strains is likely to be facilitated by exposure to sub-therapeutic drug levels, and treatment failure is more frequent in patients with levels lower than 2 mg /L compared with those with higher levels. These ndings are in agreement with our study.
We observed that more patients on the efavirenz regimen had a virologic failure (37%) than those on the nevirapine regimen (21%). This is probably because efavirenz is associated with CNS side effects causing less adherence to patients on this regimen.
It is well known that virtually all of the antiretroviral drugs available, except tenofovir, lamivudine, and abacavir, can produce a mild elevation in liver function tests. As observed by [19], the number of patients with elevated ASP who had elevated nevirapine levels (> 3 mg/L) was double that of the people on efavirenz (> 2 mg/L). This is in agreement with [20], who stated that during therapy with NVP or EFV increases of liver enzymes were not unusual. Our ndings are in agreement with the above studies. In our study, more patients on the nevirapine regimen had liver toxicity (89%) than the patients on the efavirenz regimen (45%).
HIV management in Uganda follows the existing national guidelines that include periodic monitoring of viral load and CD4 levels do not include Therapeutic drug level (TDL) monitoring as is often the case in most advanced countries. If ARV drug plasma levels and viral load are used in the routine management of HIV positive patients, treatment failure due to toxicities, concomitant clinical conditions, and development of virologic failure can be detected early leading to early regimen switching preventing the deterioration of the immune system hence giving better treatment outcomes.
It has been reported that individuals with HIV infection may be nonadherent to ART when consuming alcohol due to beliefs regarding adverse interactions between alcohol and ART or because their providers have advised them not to consume alcohol while taking ART [21]. Our ndings though not statistically signi cant show that, generally, patients who did not consume alcohol had slightly lower viral loads (p = 0.239) and slightly higher CD4 counts (p = 0.172) than those who consumed (Table 4). This is probably because alcohol consumption may reduce adherence to ART, leading to decreased ART effectiveness and, ultimately, increased HIV-related mortality [22]. The difference in efavirenz drug levels among alcohol drug users was not statistically signi cant (p = 0.955). Efavirenz is known to induce the liver enzyme cytochrome P450 (CYP 450) 3A4 that might contribute to the decreased alcohol plasma concentrations. Since CYP 3A4 is a minor pathway for alcohol metabolism, medications (such as efavirenz) that induce this enzyme function might result in small corresponding changes in the overall metabolism of alcohol [21]. This might explain why alcohol consumption did not signi cantly affect efavirenz plasma concentration.

Conclusion And Recommendation
Our study concludes that a large number of patients on nevirapine and efavirenz-based regimen in South Western Uganda have sub-therapeutic drug levels and viral load copies > 1000 copies mL. This situation presents a brewing pot for resistant viral strain selection which in the long run will lead to the two regimens being rendered useless. Regular plasma drug and viral load monitoring are highly recommended in ART care in South Western Uganda.