Adherence to Antiretroviral Therapy: Prevalence, Determinants and Impact on Body Weight and Immunological Recovery among People Living with Human Immunodeficiency Virus in Osogbo, Nigeria

Background: antiretroviral therapy has changed the outlook of AIDS. However, identifying factors that will strengthen its maintenance is vital to treatment success. Advocacy is growing on the need for close attention to immunological progress, prevention of excessive body weight gain and associated immunological and metabolic Original Research Article British Journal of Medicine & Medical Research, 4(31): 5003-5018, 2014 5004 consequences for better long-term outcomes among PLWHIV in Africa. Aims: To study prevalence, determinants of adherence, and the existing relationship between body weight and CD4 count among adherents and non-adherent patients on HAART. Methodology: A cross-sectional design for sampling of 270 patients on HAARTS was made and pharmacy based adherence was calculated. Patients were categorized into weight groups according to WHO guideline and CD4 count was determined at baseline, third and sixth months. Result: Calculated overall pharmacy adherence was 62.6% over six months. Disclosure to a close family member (p=0.013) and living outside the city of care (p=0.025) significantly predict adherence. Pretreatment overweight (BMI-2529.9) and obesity (BMI>30.0) were temporary beneficial to CD4 constitution at baseline (p=0.004), while overweight (p=0.041) and obesity (p=0.150) were associated with lower CD4count repopulation at six months postHAART compared to normal body weight (BMI-18.5– 24.9), p˂0.001. Adherent PLWHIV participants had higher body weight increasing effect, but demonstrated lower CD4 lymphocyte count increasing effect compared to the nonadherent at six months post-HAART, (p<0.001). Conclusion: Normal body weight and maintenance during HAART seems beneficial for immune reconstitution at six months postHAART. While emphasizing good adherence to HAART, it becomes necessary for programme implementers to watch against excessive body weight gain and attendants adverse immunological consequences.


INTRODUCTION
The emergence of antiretroviral therapy (ART) approximately one and half decade ago has converted the dreaded human immunodeficiency virus infection (HIV) to a chronic disease [1]. The success recorded with treatment had significantly reduced the morbidity and mortality associated with acquired immunodeficiency deficiency syndrome (AIDS) [2] with consequent improvement in life expectancy [3,4]. This impact of the improved outcome associated with ART is not fully felt in sub-Saharan Africa due to limited accessibility to ART and challenges with adherence to medication. For example, of the 1.3 million Nigerians who required ART, only 600,000 are on medication as at 2012 [5]. Similarly, out of the 277,000 new infections, only 56,000 were commenced on treatment. To achieve the Millennium Development Goal 6 by 2015 [5], studies are needed to identify effective and sustainable evidence-based interventions in the course of standard care. New means of broadening access and maintaining PLWHIV in care are also needed.
Adherence is taking antiretroviral (ARVs) exactly as prescribed by the caregivers [6]. Optimal adherence is necessary for good clinical outcomes, reduction of morbidity and mortality [2,6,7]. The reported adherence rates to ART medication among people living with HIV (PLWHIV) in Nigeria vary from 44% to 98% [8][9][10][11][12] Factors shown to improve or associated with good adherence include text message as reminders, [8] patient selected treatment partners, [9] use of pill box, [10] age and gender [11]. On the other hand, psychiatric morbidity negatively had adverse impacts on adherence [12]. Current reports indicate increasing and growing concerns on HIV -associated multi-morbidity with a potential threat on healthy ageing HIV population with the capacity to overwhelm inadequate health care facilities in the era of ART [1]. Of these is the increasing problem of overweight and obesity, growing burden of risk factors for cardiovascular diseases, [13][14][15][16] chronic inflammation induction [13,14] and cancer [14]. Although reports have been conflicting on the impact of body weight on immunological constitution of HIV patients [15][16][17][18][19]. Systemic hypertension, diabetes, dyslipidemia and cardiovascular diseases are common among obese and overweight PLWHIV [15,17]. Therefore, while optimal adherence has been widely emphasized in studies; concerns are also growing on the need for weight and CD4 count monitoring [15][16][17][18]. Studies have shown a high rate of adherence in the developing countries including Nigeria, but maintenance of adherence and retention in care has been a major concern [6,7,[15][16][17][18].
In view of the foregoing, the present study examined factors associated with adherence, and investigated relationship between body weight and immune status of adherent and nonadherent PLWHIV at baseline and six months after HAART

Study Location/Design
This cross-sectional study involved every alternate patients seen by caregivers at the dedicated clinic to PLWHIV of the Ladoke Akintola University of Technology Teaching Hospital, Osogbo, Osun State, Nigeria from February 2011 to January 2012. The HIV clinic is being supported through collaboration programme of Federal Government of Nigeria through the Institute of Human Virology Nigeria (IHV-N), President Emergency Fund for Aid Relief (PEPFAR) and other non -Governmental donors at the time of the study. Free drugs, laboratory support, technical assistance and social support were being offered to the clinic attendees by the collaborators.

Subject Selection
The study participants included 270 HIV-seropositive adults diagnosed by two antibody screening tests. The participants were adult (age ≥18years), non-pregnant and using highly active antiretroviral therapy (HAART) for not less than three months before the commencement of the study. The participants signed informed written consent. The study excluded PLWHIV, who were on ART less than three months, and those who refused to be part of the study. Research Ethical Committee of the LAUTECH Teaching Hospital approved this study.

General data
The attending physicians or the adherence counselors interviewed all consenting eligible participants using a semi-structured questionnaire. The questionnaire captured sociodemographic data such as the participant's age, the gender, ethnicity, the highest level of education attained, occupation, place of residence and approximate distance travelled before getting to the clinic, HAART, duration on HAART, and side effect profile of drugs. The side effects of drugs were grouped to gastrointestinal (GIT), skin, and central nervous system (CNS). Gastrointestinal symptom included nausea, vomiting and abdominal pain; skin side effects included presence of a new rash after commencement of HAART and pruritus; and central nervous system complications included headache, paraesthesia, dizziness and vivid dreaming. The clinical stages of the HIV infection at commencement of HAART were noted. The participants' knowledge of the use of different evidence-based intervention or strategies to enhance adherence in the low-and middle-income countries was assessed irrespective of whether patient was using them presently. The study documented knowledge of interventions such as the pill box, drug partner, timed pill, and patient's education module by affirmative direct questioning (yes, or no).

Assessments of medication adherence
Medication refill takes place at the pharmacy attached to the clinic. In the early phase of treatment, the refill periods were typically short, initially every two weeks in the first month of therapy, subsequently monthly until the sixth month of therapy. The trained adherence counselors gave group and individual adherence counseling at every contact. The HAART was defined according to treatment guidelines of World Health Organization (WHO) for HIV treatment as applicable to adults and adolescents in the developing countries as adopted by the Federal Government of Nigeria [20]. Two nucleoside reverse transcriptase inhibitors (NRTIs) plus one NNRT were used as first-line therapy. The HAART combination were (Zidovudine (AZT)/Lamivudine, (3TC)/Efavirenz (EFV) and (AZT/3TC/Nevirapine(NVP) essentially. Very few participants were taking Tenofovir-based therapy [20]. Adherence to HAART was assessed over the last 3 to six months using the pharmacy-based adherence measure (PAM). Using the pill count category of PAM, the quantity of medication used was determined between two antiretroviral pickup date as a proportion of the number of pills dispensed expressed in percentage [21]. The PLWHIV with adherence rate equal or greater than 95% were categorized as been adherent while those with less than 95% calculated adherence categorized as non-adherent. All data related to medication intake were extracted from duplicate of pharmacy refill forms at the back of the patients' medical files.

Anthropometric assessments
Routinely in the clinic, the weight and height of participants were measured using a stadiometer to the nearest 0.1cm and 0.1kg respectively. Investigators extracted data of baseline height and weight, weight at the third and sixth months from the participants medical files. The body mass index (BMI) was calculated using the equation weight (kg)/ height 2 (m 2 ) and categorized according to WHO criteria as follows: underweight when BMI is <18.5kg/m 2 , normal, BMI=18.5 -24.9kg/m 2 , overweight, BMI=25 -29.9kg/m 2 and obese, BMI ≥30.0kg/m 2 [21].

CD4 count
All available CD4 count results done in the course of treatment: at baseline, at the third and sixth months of therapy were copied from participant's folders.

Statistical analysis
All the continuous and categorical variables of interest were documented as means ± SD and percentages respectively. Chi-square was employed to find the gender differences in the categorical variables of the participants. Differences between two means were assessed using Student's t test. Pharmacologic adherence was dichotomized to < or ≥95%, and the relationship with socio-demographic and clinical variables was estimated using chi-square, while strength of association for significant variables were tested using the odd ratio statistics. All significant factors associated with adherence in the bivariate analysis were entered into binary logistic regression model to identified factors that predict adherence. One way analysis of variance (ANOVA) was used to determine significant difference in CD4 count across BMI categories at baseline, the pair means differences in CD4 count at three and six months were determined using student's t-tests. The pair changes in means of weight, BMI, and CD4 count for the adherent and non-adherent patients were tested using Wilcoxon signed rank test; effect size was calculated using Z/square root of N, where N is the number of paired data. P-values were two-tailed, values <0.05 were taken as significant. Statistical Programme for Social Sciences (SPSS) version 17 (SPSS Chicago Inc, IL.USA) was used for analysis.

RESULTS
The study participants consisted of 270 adults (202 females and 68 males) with a mean age of 38.0±8.8years, with the majority [208 (77%)] under 45 years of age. The men were significantly older and taller than the women. There was no gender difference in the educational status, mean weight and BMI of study participants at the baseline. Most participants had normal body weight 154 (59.0%), 16.1% were underweight, 17.2% were overweight, while 7.7% of participants were obese. The women had higher mean CD4 count than men at the baseline (255.17±168.60 vs. 181.71±168.60 cells/mm 3 (p=0.002), Table 1. Fifty-two percent of the participants were in the advanced (18.0%) and severe (34.0%) immunosuppression stages at the time of diagnosis, while 27.6% and 20.4% were in stages 1 and 2 respectively.
Majority of the study participants (85.6%) were placed on Nevirapine-based HAART, while 8.5% were on Tenofovir-based and 2.9% were on Efavirenz-based HAART. Two hundred and nineteen (82%) participants had been on HAART for at least six months before the commencement of the study while 48(18.0%) had been on the HAART for less than six months. The median duration of HAART use was 20months (Range: 3-84).
The calculated overall pharmacologic adherence over last six months was 62.6%. Table 2 shows the relationship between adherence and socio-demographic and clinical variables. There was no age, gender, ethnic, educational, and occupational difference between the adherent and non-adherent participants. However, the study members residing outside Osogbo (care center) metropolis were more adherent than those living within the city (p=0.01). Dichotomized travel distance (less or greater than 10kilometer) to the source of care did not influence the adherence pattern (p=0.266). No significant association was found between adherence, categorical CD4 count, extent of disease progression at the baseline, presence or absence of wasting and clinical staging, ( Table 2).  Table 3 shows the pattern of pharmacologic adherence in relation to the knowledge of adherence enhancers, medication side effects and serological status disclosure. The participants demonstrated good knowledge of the adherence enhancer instruments with an average of greater than 90% for all enhancers, except the pillbox that was 85.6%. However, there was no association between the knowledge of these instruments and adherence. Seventy-seven participants (59.7%) experienced one form of side effect or the other with gastrointestinal side effect being the commonest (61.5%). There was no relationship between different side effects and adherence. In relation to disclosure, a greater percentage (92.2%) of participants had disclosed to at least one individual. Of all the forms of disclosure, disclosure to close family member was found to be significantly associated with pharmacologic adherence. Patients who disclosed to close family members were about 2.0 times more likely to be adherent than their counterparts who did not disclose. (OR: 2.0; CI=1.2 -3.3, p=0.007), Table 3.  The pair means differences in CD4 count, weight, and BMI of adherent and non-adherent participants were examined from baseline to 6 months post-HAART period using the Wilcoxon signed rank test for two related data. The adherent participants had lower mean baseline weight, BMI, but higher CD4 cell count compared to the non-adherent participants at baseline. Both medication adherent and non-adherent participants had a significant increase in weight, BMI and CD4 count after six months of HAART, p<0.

DISCUSSION
The present study shows adherence to HAART is significantly predicted by disclosure to close family member and residing outside the locality of the source of care. The pretreatment overweight and obesity were beneficial and associated with initial higher CD4 count constitution. However, after six months of HAART, the benefit of high BMI appears eroded as normal body weight (BMI, 25 -29.9) category demonstrated highest CD4 count reconstitution effect compared to other weight categories. These tendencies to higher weight gain with concomitant reduce CD4 count reconstitution was found among HAART adherent study members, thereby suggesting a negative impact of obesity on immune reconstitution in HAART era.
The overall pharmacy-based adherence rate of 62.6% documented in this study falls within quoted 45 -100% range for most developing nations [6,10,11,12]. Sero-status disclosure to a close family member significantly predicts good adherence to HAART in the present study as documented by Stirratt et al. and other workers [22][23][24]. While some workers attributed adherence and tendencies of PLWHIV to disclose to their close family members and friends to help with medical care and counseling, [23] others did not find or assessed the types of practical support been rendered to PLWHIV by these acquaintances [21,24]. Family support could independently explain adherence while non-disclosure to a family member was associated with non-adherence in a Uganda study [25]. All initiatives that will encourage PLWHIV to disclose should be strengthened in the course of standard care in limited resource settings. Other studies are necessary to identify the content and extent of family support been rendered to PLWHIV in sub-Sahara Africa. Contrary to findings of other studies, characteristics like age, gender, education, occupation are not associated medication adherence in the present study [11].
Residing outside the city where the source of care is located predicts good adherence as also shown by previous studies [34,35]. Carlucci et al. [35] found optimal adherence among HIV-positive participants that had to travel long to the source of care. In contrast to the present finding, long distance travelling has been associated with missed doses, poor adherence, and family financial constraints [34]. Long distance journey to the source of care was commoner among white males, elderly, those on multi-therapy and affluent rural dwellers in England [34]. We opined that having to traverse other nearby care centers to a distant location may be associated with stigma feeling and confidentiality protection. Further analysis of this cohort showed majority (78.0%) of this mobile participants were selfemployed and were incurring extra travel expenses which their low socio -economic status may not be able to sustain for a long time. Sustainability due to financial constraint that may arise anytime in the course of treatment may reverse this primary trend. While current evidence-based decentralization of care to bring treatment close to PLWHIV is been pursued because of its documented advantages, more effort should be directed at means of reducing social stigma to prevent excessive mobility among PLWHIV [36].
Many of the participants showed good understanding of adherence enhancers such as the pill boxes, timed pill, treatment partners, and education module. Although none of the enhancers was predictive of adherence to HAART, participants with good knowledge of enhancer's demonstrated better adherence compared to those without knowledge. Opinions have been diverse in published studies on the impact of enhancers in Africa [37][38][39][40]. The use of patient's selected treatment partners [16] and "adherence club" were associated with superior adherence and better clinical outcomes respectively [37]. Kunutsor et al. [38] also found the combination of enhancers favored long term sustenance on ARVs better than systematic adherence monitoring.
The present study showed that highly active antiretroviral therapy induce weight gain among PLWHIV and body weight impacts CD4 lymphocyte count as previously shown in studies [15][16][17][18][19]. We found association between been underweight BMI ˂18.5 and lower pre HAART CD4 count, while overweight, and obesity were beneficial to CD4 count constitution at baseline and 3months post HAART only. Although the observation was not statistically significant, obesity seems less beneficial to the CD4 count reconstitution while normal weight category (BMI: 18.5-24.9) experienced significant, and highest CD4 count repopulation effect compared to the overweight and obesity at six months post HAART. Our finding is similar to the findings of other workers that showed lower gain in CD4 cell count among obese PLWHIV over a longer time frame, as CD4 was noticed to have progressively risen from baseline but to plateau in the overweight category [15,16,18,19]. This study is however different from other studies that showed obesity did not compromise immune cell response [17,41,43]. However, despite lower gain in CD4 count compared to normal weight category (BMI, 18-24.9), obese and overweight PLWHIV (BMI≥25) in contrast to a previous study [17], showed a higher likelihood to exceed CD4 count gain threshold of ˃100cells/mm 3 at six months post HAART. Much greater likelihood threshold (˃350cells/µL) was observed by Womack et al. in their study among white women [44]. Womack et al. [44] suggested discrepancies observed in these studies may be attributed to analysis methods.
Low CD4 cell count as observed in pre HAART and post HAART underweight category in this study may be attributed to ongoing malnutrition among HIV patients [45]. Associated with malnutrition is apoptosis induced thymocyte depletion that has particular depleting effect on CD4 count and attenuation of cell proliferation. Studies have also confirmed proteinenergy malnutrition (PEM) is associated with a reduction of fat cell mass and consequently decrease in circulating leptin, impaired production of pro-inflammatory cytokines (IFN γ, and TNF α) and low T cell activation as a possible link to decline of CD4 cell repopulation in this group [46][47][48][49][50][51][52][53]. Serum leptin levels showed a positive correlation with the extent of adiposity and higher in women than men [48]. The administration of leptin to congenitally deficient children led to the reversal of immunodeficient state: restoring IFN γ secretion, CD4 cells development and functions [47]. Although malnutrition is still a common problem in HAART era especially in Africa, certain conditions associated with reduced weight gains such as inadequate viral suppression and on-going intercurrent infections should be investigated as a possible cause of lower repopulation of CD4 count among underweight PLWHIV [43].
Researchers have also proposed the mechanism of lower CD4 count and increasing cardiovascular morbidity and mortality [1,[13][14][15][16] among obese PLWHIV. Obesity has been associated with hyperlectinemia [47][48][49]53]. Studies have suggested tolerance or resistance to the effect of leptin as the cause of "leptin-deficient" state among obese individuals. High circulating leptin has been associated with alteration in expression of leptin signal molecules at the receptor level [48]. Deficient leptin signaling which is due to "leptin deficiency" state results in increasing obesity, lipid storage in muscle, liver and other human tissues, dysfunction of neuro-endocrine function, as well as immune function. High level of leptin has been linked with direct induction of an inflammatory state and is an underlying factor for insulin resistance and development of metabolic syndrome. Similarly, HAART-treated PLWHIV with HIV-associated lipodystrophy syndrome showed a significant reduction (40%) in leptin and reduced expression of transcription factors within their fat cells. They also had increased production of TNF-α, interleukin 6 (IL-6) and IL-8 expression [50][51][52]. The low adiponectin, shift in pro inflammation factors, profibrotic and dysregulated metabolic state are the factors that contributed to the reduction of cell differentiation including reduce CD4 cell activation, reduced fatty acid uptake and consequent dyslipidemia [1,[50][51][52][53]. The elevation of many of the inflammatory markers has been attributed to increase disease and all-cause mortality [1]. However, in this study, lower level of CD4 count in the obese category six months post HAART may be partly explained by a lower starting CD4 count at baseline compared to the overweight category. Other studies may be needed to examine the contributors to the chronic inflammation induction and reduce immune activation among obese PLWHIV.
The prevalence of pre HAART overweight and obesity (24.9% ) found in the present study falls within 8.1 -35.1% and 20 -30% documented among Nigerian and American general population respectively, but lower than 63% documented at last visit in a study among HIVpositive military men in United States [15,19,54]. However, the multiplication effect of high national HIV prevalence in most African countries portends grave consequences for already burdened health care system. The cardiovascular and metabolic consequences among PLWHIV are a time bomb that may shape the future needs of HIV as a chronic disease [1,[13][14][15][16]41].
From this study, excessive weight gain and the consequent complications are more likely to occur among the ARV-adherent populations. PLWHIV and programme implementers must pay close attention to clients that are "doing well" as evidence by increasing weight and other physical parameters. Close attention to immunological progress, weight monitoring, and frequent assessment for cardiovascular risk factors are important implementable programme needed now in Africa. Studies are needed to examine allowable limits of weight necessary for optimal immunological outcomes for PLWHIV on HAART. Similarly, a change that physical therapy would make on weight reduction, weight maintenance and immunological recovery should be examined in a longitudinal study among PLWHIV.
This study was fraught with some limitations. Viral load (VL) assessment is not routinely done in our centre and due to constraint of funds VL assessments were not done. Similarly, attrition during treatment, morbidity and mortality were not examined as end point since our focus is on adherence as related to weight and immune constitution. The assessment of BMI in these patients only suggests weight gain but not a good estimate of body fat composition.

CONCLUSION
In Conclusion, disclosure to close relative and residing outside the source of care predict adherence to HAART. Stigma and confidentiality protection are still influencing the decision of PLWHIV to travel far to source for care. The HAART adherent PLWHIV demonstrated higher increasing weight effect but lower CD4 reconstitution effect. Normal body weight and its maintenance during HAART seems beneficial for immune reconstitution at six months post-HAART. While emphasizing adherence, careful emphasis should be placed on weight and immunological monitoring.