Adverse Effects and Regimen Switch among Patients on Antiretroviral Treatment in a Resource Limited Setting in Ethiopia

Abbreviations: ALT: Alanine Amino Transferase; ART: Antiretroviral Therapy; BMI: Body Mass Index; CD4: Cluster of Differentiation 4; CNS: Central Nervous System; DDI: Didanosine (vidax®); D4T: Stavudine (zeri®); EFV: Efavirenz (sustiva®, strocrin®); GI: Gastrointestinal; HAART: Highly Active Antiretroviral Therapy; HIV: Human Immunodeficiency Virus; LPV/r: Lopinavir/ritonavir (kaletra®); NVP: Nevirapine (viramune®); PLWHA: People Living With HIV/ AIDS; SPSS: Stastical Package For Social Sciences; TB: Tuberculosis; ZDV: Zidovudine (retrovir®); 3-TC-Lamuvudine, (epivir®); ZDV/3-TCZidovudine/Lamuvudine (combivir®); ZDV/3:TC/NVP: Zidovudine/ Lamuvudine/Nevirapine (zidolam®)


Introduction
The Human immunodeficiency virus (HIV) infection has created enormous challenges worldwide since recognition of the disease. The highest number of people living with HIV/AIDS (PLWHA) is found in Sub Saharan Africa. Ethiopia is among those countries affected with the epidemic [1,2]. The overall growth of the epidemic has stabilized in recent years. Due to increase in people receiving antiretroviral therapy (ART) the number of AIDS-related deaths has declined [3]. People infected with the HIV are now living longer, healthier live. Since introduction of the ART, the disease has become in developing countries a chronic condition that can be managed for long term [3,4].
The combination ART is the corner stone of management of patients with HIV infection. Current ART regimens are capable of reducing viral load to undetectable levels, with a consequent increase in immune cells and reduction in development of opportunistic infections [4,5]. In recent years there is a growing awareness of the problems accompanying the use of HAART. In addition to drug resistance and difficulty of adhering to complex regimens, adverse effects associated with HAART have become a major concern [6]. Inspite of ART benefits, adverse effects to these drugs have been pointed as one of the main reasons for discontinuation, switch and non adherence [7][8][9].
Some clinical outcomes in AIDS patients are difficult to differentiate whether from the disease itself or from ART side effects. Neurologic complications occur in advanced AIDS disease and may be exacerbated by ART [10]. Anemia in HIV infected patients has multiple possible causes: HIV infection, drugs (zidovudine, cotrimoxazole, anti-tuberculosis), malnutrition, and blood loss [11]. Diarrhea in AIDS patients could be due to drug related or due to HIV invasion of the intestine resulting in enteropathy or due to gastrointestinal opportunistic infections [12]. Viral hepatitis coinfection and alcohol may increase the risk of adverse hepatic effects of ART [13]. Therefore, confounders to ART adverse effects may range from the clinical outcomes of the disease itself (advanced HIV infection) to immune reconstitution inflammatory response, concomitant medications, comorbidities and opportunistic infections.  [14]. Woldemedhin et al. found toxicity as the most common reason for ART regimen change [15]. Understanding the existing HIV treatment strategy is important to assure the efficacy and safety of antiretroviral treatment so as to improve patient adherence and quality of life. The study aimed to assess the prevalence of antiretroviral associated adverse effects and management strategies undertaken among patients taking ART in Jimma University Specialized Hospital.

Methods
A cross-sectional study was conducted by a retrospective review of three years' (2009)(2010)(2011) patient medical records at the ART clinic of Jimma University Specialized Hospital. Data was collected from February to March, 2012.
Patient medical records were used for abstracting information about patients' socio-demography, adverse effects, regimen switch and interruption of treatment. The source population was all HIV patients who started ART and on follow up in the clinic and patients who had more than six month of follow up in the clinic was included in the study.
The sample size was determined using the single population proportion formula, assuming the prevalence of ART adverse effects to be 50%, and 5% margin of error at 95% confidence interval. After considering a 5% increment for non-response, the sample size was determined to be 403.These records were drown from the source population by systematic random sampling, using the list of the card numbers as sampling frame. The sampling interval (k) was determined by dividing the total source population by the sample size (3400/403=8). The first medical record was selected randomly from the first eight records in the sampling frame and then every eighth record was included in the study.
A pre-tested structured data abstraction format was used for data collection. Official letter was written to the hospital medical director from Jimma University. Patient information was kept confidential. The data was collected by trained Nurses under the supervision of the principal investigator. After all data was obtained, it was checked for completeness and clarity. A particular patient record with missed data was excluded Data were entered, coded and analyzed using SPSS for windows version 16 statistical software. Descriptive statistical analysis was used to describe patient demographics, clinical and laboratory characteristics and ART regimens. Since all variables were categorical chi-square test was used to analyze factors associated with increased occurrence of adverse effects. P-value of less than 0.05 was considered as statistical significant.

Tuberculosis treatment interaction
Interaction antiretroviral therapy with treatment of TB (mainly nevirapine and rifampcin), reducing the efficacy of anti-retroviral drugs and of anti-TB drugs while increasing the risk of drug toxicity.

Toxicity
Severe and life threatening adverse effect to ARV drugs in a dose normally used for treatment and requiring change in therapy.

Results
Four hundred three (403) patient records were reviewed in the study. Patient demographic characteristics showed that 264 (65.5%) patients were females and 139 (34.5%) males. Majority of the patients, 376 (93.3%) were in the age group of 15-49 years. Regarding patient source, most of them (96%) were outpatients and 16 (4%) had history of hospitalization. Fourteen (5.3%) of the females had history of pregnancy. The nutritional status of patients was estimated using body mass index (BMI) considering the weight and height at the start of treatment. One hundred seventy one (42.4%) had BMI less than 17 and 98 (33.3%) and 134 (24.3%) had BMI 17-18 and greater than 18, respectively. Regarding the CD4 counts at the start of ART, 299 (74.2%) patients had CD4 count less than 200cells/mm 3 .
Out of the 403 patient records reviewed 264 (65.5%) patients had at least one adverse effect related to the antiretroviral treatment. Two hundred three (50.3%) patients had a recorded history of three or more adverse effects. Mild but commonly encountered adverse effects were gastrointestinal; nausea (28%), vomiting (23.8%) diarrhea (20.6%); and central nervous system (CNS) side effects: nightmare (24.6%), dizziness (15.6%) and insomnia (20.5%). Moderate to severe adverse effects recorded were anemia characterized by hemoglobin less than 11mg/ dl and fatigue in 81 (20.1%), peripheral neuropathy in 91 (22.7%), hepatotoxicity characterized by elevated liver enzyme and jaundice in 37 (9.2%), and rash in 121 (30%). Six patients reported a very severe rash. Some 79 (19.7%) patients had recorded history of cough which was difficult to identify it as drug or disease related. The rarely observed side effect was body fat accumulation (lipodystrophy) in 7 (2%) patients. Table 1 and Figure 1 shows regimen based prevalence of major adverse effects. High frequency of regimen discontinuation and switch drug regimen which could be complicated and commonly associated with adverse effects. In this study, the prevalence of adverse effect was about 65.5%. A higher prevalence of adverse effects (92.2%) was reported from a study in Brazil [9]. This difference could be due to differences in methodology, where the study in Brazil was based on patient self report and our study was based on patient records which could be affected by adverse effect recording and reporting practice of the health professionals. Similar to a study conducted in two public referral HIV/ AIDS centers in Brazil [8], the most commonly encountered side effects were gastrointestinal and CNS side effects. Other common but more severe adverse effects were rash, peripheral neuropathy, anemia and hepatotoxicity.
Parallel to findings from pivotal trials [16], CNS side effects were common among patients on efavirenz based regimens (D4T/3TC/ EFV, ZDV/3TC/EFV). Though studies on pharmacogenomics data of Ethiopians are lacking, some reported variability in the CYP2B6*6/*6 and ABCB13435TT genotype are common in Ethiopians resulting in elevated plasma efavirenz and increased side effects [17]. As compared to EFV based regimen, patients on NVP based regimen experienced higher rates of hypersensitivity reactions (rash, hepatotoxicity). Very severe rash, which seems Steven Johnson Syndrome (SJS), occurred in seven patients on NVP and cotrimoxazole; it was difficult to identify which drug caused this syndrome. Slightly higher rate of efavirenz associated hepatotoxicity (12.5%) was observed in a prospective cohort study by Maggiolo et al. [18]. High occurrence of hepatotoxicity was significantly associated with long duration of use, higher hepatitis virus coinfection, and high baseline Alanine amino transferase (ALT). In our case, though no record of hepatitis virus coinfection was found, the high rate of hepatotoxicity and rash could be due to overlapping toxicities with anti-tuberculosis drugs. Patients on antituberculosis were at higher risk of developing adverse effects as compared to patients on cotrimoxazole (OR=4, CI, 0.12-1.7, p=0.04). Such overlapping toxicities could be serious and challenging in the management of patients with TB-HIV co-infection.
Stavudine (D4T) based regimens were the commonly used regimens in our setting despite occurrence of adverse effects like peripheral neuropathy (21.5%), lipodystrophy (1.4%) and some gastrointestinal events. A study from an outpatient clinic in Kenya also reported peripheral neuropathy (20.7%) as most commonly encountered side effect [19]. More peripheral neuropathy was reported in patients who switched from D4T (30) to D4T (40) due to weight gain [19][20][21]. was observed in patients who experienced anemia, rash, hepatotoxicity, and peripheral neuropathy.
When we see the time occurrence of adverse effects, rash hepatotoxicity, most GI, and CNS adverse effects occurred early after starting treatment at mean interval of 2 weeks. Lipodystrophy, anemia, and peripheral neuropathy were among the long term adverse effects observed with mean time occurrence of 2, 1.6, and 0.5 years respectively. One hundred forty (34.7%) patients had co-morbid tuberculosis and 43 patients changed previous ART regimen due to Nevirapine drug interaction with antituberculosis drugs. About 303 (75%), and 108 (27%) patients were on cotrimoxazole and isoniazid prophylaxis, respectively.
ART switch was done in about 257 (62.8%) patients while 27 patients switched two times during the three years of duration. The main reason for ART switch was weight gain (129 (50.2%) patients); where patients previously on 30 mg stavudine had to switch to 40mg when the patients' weight is above 60 kg. Patients who switched to 40 mg stavudine reported more adverse effects such as peripheral neuropathy. The other reasons for ART switch was toxicity (30%), tuberculosis treatment interaction (17.1%), pregnancy (3.3%) and treatment failure (0.38%).
Regarding the clinical measures taken to manage the adverse effects, 77 patients switched to other regimen (20 of them discounted treatment for days and switched later) while 10, 22, and 12 patients were made to discontinue the concomitant anti-tuberculosis, isoniazid and cotrimoxazole prophylaxisis, respectively. Supportive medications prescribed for management of adverse effects were amitriptyline, diclophenac, diphenhydramine, iron sulfate, predinisolone, methclopropamide, cemetidine and metronidazole. Except metronidazol, which was used for treatment amebiasis infection, the supportive medications were given for management of the adverse effects. Amitriptyline -peripheral neuropathy; iron sulfate for anemia; predinisolone for rash; methclopropamide and cemetidine for mangment of gastrointestinal events. Table 2 summarizes the association of variables with occurrence of adverse effects. Variables found with clinically significant association were age, body mass index, CD4 count, regimen type, ART switch and presence of concurrent drug.  Unlike to low rate of ZDV induced anemia (12%) reported from tertiary general hospital, in London, anemia was significant in our study (26.6%). Underweight might be the reasons behind in our study subjects where underweight patients were 1.96 times more to develop adverse effects [15].

Discussion
In contrast to other studies [8,19] there were no sex difference in rate and severity of adverse effects. Variables statistically associated with occurrence of adverse effects were age>60, body mass index <18, CD4 count >400, regimen on D4T(40)/3-TC/NVP, ART switch and presence of concurrent antituberculosis drugs. Related study by Menezes [8] found sex, ART switch, and high CD4 count as predictor of ART toxicity. Age older than 40 years was independent predictor of clinical toxicity in another study conducted in outpatient clinic [19]. CD4 count in this study indicates improved therapeutic outcome since many patients had increment in CD4 counts as compared to baseline. Patients with CD4 count 200-400 had higher occurrence of adverse effects, though no significant difference was obtained in this study, many studies report that female patients with high CD4 count (>250 cells/mm 3 ) are at high risk of Nevirapine associated rash and hepatotoxicity [22]. Currently there is increasing interest that genetic variability can significantly affect individuals' response adverse effects. In Ethiopian population pharmacogenomic data of antiretrovirals are lacking and this limits us to explain the existing difference in response to ART medication between Ethiopians and other population across the world.
In this study the clinical measures undertaken to manage adverse effects were ART switching, discontinuation of co-medications and prescribing supportive medications. Adverse effects that required high rate of regimen switch were rash, anemia, peripheral neuropathy and hepatotoxicity. Similarly, studies found that hypersensitivity (rash and hepatitis), hematologic and liver function abnormalities were common toxic effects leading to change in ART [19,23].
Many studies reported toxicity as the main reason for ART switch [18,19,21]. In this study the main reason for ART switch was weight gain followed by toxicity and TB treatment interaction. Likewise, in a study in Kenya the ART related toxicity and tuberculosis treatment interaction were the main reasons for ART switch [19]. ART switch by itself may result in development of new and severe side effects and further complications and loss of future treatment options.
Some of the limitation of the study was poor adverse effect recording and reporting by the health professionals. Some effects of the ARV drugs are also found in people with HIV infection which are characteristics advanced HIV infection, or might be related to concomitant medications and comorbidities). Being a retrospective study differentiation of ART adverse effect from AIDS related outcome was difficult. However AIDS related outcomes occur commonly in untreated patients or at early stage of treatment. The adverse effects are likely due to ART where there was strong temporal relationship of occurrence after initiating ART, pharmacologically related, and there was improvement up on discontinuation and switch to other agents.

Conclusion
In this study the prevalence of adverse effects associated with antiretroviral treatment were found common (65.5%). Mild but  commonly reported adverse effects were gastrointestinal and CNS adverse effects while anemia, peripheral neuropathy, rash and hepatotoxicity were severe effects resulted in high rate of regimen switch. ART switch could be an option for management of toxicities but it should be undertaken considering the risk of loss of future treatment options. Therapeutic drug monitoring may not be feasible in resource limited setting; therefore health providers working in the ART clinic should monitor patients both clinically and with laboratory for the occurrence of side effects. Particularly patients on ZDV, NVP and those on concomitant medication need close follow up. The health system should develop ADR database so as to easily record and report adverse effect.