Change in serum lipid profiles and glucose after switching from stavudine / lamivudine to zidovudine / lamivudine in non-nucleoside reverse transcriptase inhibitors based anti-retroviral regimens in Southern Ethiopia

Data concerning any difference in serum lipid profiles and glucose level after patients switched from stavudine to zidovudine in Ethiopia is very limited. Seventy eight adults receiving antiretroviral therapy (ART) that included stavudine/lamivudine with either of efavirenz or nevirapine during ART initiation were enrolled. Of these patients, 53 were switched to zidovudine/lamivudine/nevirapine (NVP-group) and the rest 25 were switched to zidovudine/lamivudine/efavirenz (EFV-group). Serum lipid profiles and glucose were determined after overnight fasting. Dyslipidemia and dysglycaemia were assessed according to the United State National Cholesterol Education program-III guideline. Statistical analysis was done using Statistical Package for Social Sciences (SPSS) Version 20. Of the 78 patients, 39.7% were males and 60.3% were females. At the end of the study follow-up, the prevalence of TC ≥ 200 mg/dl, LDL-c ≥ 130 mg/dl, TG ≥ 150 mg/dl, HDL-c < 40 mg/dl and glucose ≥ 110 mg/dl were higher in EFV group when compared with NVP. About 74.4% patients had at least two laboratory abnormities which is compatible with a diagnosis of dyslipidemia at 12 month of post switch. Four lipid profiles abnormal within a single individual was 16% in EFV and 3.8% in NVP group, p = 0.08. Raised HDL-c concentration was observed in NVP group in both periods when compared with EFV. In addition, patients that switched from d4T/3TC/NVP to AZT/3TC/NVP had a significant change in TC and TG (p = 0.001 for both). Also TC ≥ 200 mg/dl was decreased from 49 to 16% (p = 0.04). Furthermore, sex was significantly and negatively associated with raised TC and TGs among patients using NVP based regimen. Raised HDL-c concentration, decreased proportion of abnormal lipid profiles and abnormal glucose was observed in the NVP group. Based on these findings, NVP may be expected to reduce the risk of cardiovascular diseases.


INTRODUCTION
Highly active antiretroviral therapy (HAART) has effectively reduced morbidity and mortality among patients living with human immunodeficiency virus (HIV) infection (Grinspoon et al., 2005;John et al., 2001).In addition, HAART also has a potential to significantly reduce the risk of HIV transmission and the spread of tuberculosis (UNAIDS, 2013).Since its introduction, patients have started to live longer, however co-morbid problems have been emerged.Among these challenges, lipid profile derangements, insulin resistance, and diabetes are some of the metabolic complications of longterm use of HAART (Cohan, 2000;DAD, 2003).Dyslipidemia in HIV-infected patients using HAART includes, elevated level of total cholesterol (TC), LDL cholesterol (LDL-c), triglycerides (TG), and decreased HDLcholesterol (HDL-c), and severe hypertriglyceridemia in several patients (El-Sadr et al., 2005).Stavudine (d4T) from nucleoside reverse transcriptase inhibitors (NRTIs) regimens is the most commonly mentioned antiretroviral agent that is being associated with metabolic syndrome (MS) and HIV related fat accumulation (Matinez et al., 2004;Jemsek et al., 2006).Lipoatrophy and lipohypertrophy are associated with multiple metabolic derangements in the setting of HIV infection and antiretroviral therapy (ART).In addition, visceral adipose tissue (VAT) has been linked to CVD in HIV-infected men irrespective of BMI or waist girth (Haubrich et al., 2009).The protease inhibitors (PIs) in general have been reported to be highly associated with MS (Cerrato et al., 2012).However, non nucleoside reverse transcriptase inhibitors (NNRTIs) based regimens importantly differ from PI-based regimens by being associated with marked increases of HDL-c and slighter raises of LDL-c and TGs ( Van der Valk et al., 2001;Tashima et al., 2003).Furthermore dyslipidemia like hypertriglyceridemia, low level of HDL-c and insulin resistance can occur concurrently in HIV infection, which eventually raises the risk of cardiovascular disease (CVD) (Pao et al., 2008).
Following the recommendations from WHO 2010 guidelines, 2013 d4T was phased-out in Ethiopia from adults' treatment option, due to its metabolic toxicity than other NRTIs (WHO, 2010).The replacing of d4T + lamivudine (3TC) with zidovudine (AZT) + 3TC is not as a result of immunological failure but for the risks of metabolic and anthropometric alterations in HIV-infected patients.Therefore, this cohort study was designed to examine the effect of regimens substitution in lipid profiles, serum glucose and cardiovascular risks after 12 month of post switch, in a resource limited setting.

Study setting and population
This prospective cohort study was carried out from May 2013 to July 2014 at ART clinic of Hawassa University Teaching Referral

Tadewos and Assegu 19
Hospital.Eligible patients were HIV-infected, immunologically stable adults and initial NRTI backbone was d4T+3TC with one of the NNRTI (either EFV or NVP).The inclusion criteria for this sub study were the patients who had continued with their initial antiretroviral regimens for a minimum of two years of treatment.Any patient who had changed antiretroviral drug in the initial regimen due to any reason before current switching was excluded.All recruited patients were those patients that switched from d4T+3TC to AZT+3TC without changing NNRTI-based regimens.All participants were ≥18 years of age, and have a good ART adherence (adherence rate ≥95%).A good adherence is defined by missing <2 dose of 30 doses or <3 dose of 60 doses; and it was adopted from Ethiopian Federal Ministry of Health (FMOH), HIV Care/ART follow-up form.
Participants receiving statin drugs, pregnant women, known diabetes mellitus patients, patients who took alcohol within 24 h and renal failures were excluded.Subjects were monitored at the time of switching (month 0) and 12 month subsequently for serum lipid profiles and glucose analysis.For all participants, data were collected on the sociodemographic information together with body mass index, medical history including diabetes mellitus, renal failures and use of drugs that alter lipid profiles.CD4+ lymphocyte count was done by K 2 EDTA anticoagulated blood using flow cytometry instrument (Becton Dickinson, CA, USA).After an overnight fast of 8 to 12 h, venous blood was collected in plain tube from each patient and it was centrifuged within 15 to 20 min of collection at 3000 cycles/min for 5 min.Then serum was separated immediately for determination of serum glucose and lipid profiles (TC, HDL-c, LDL-c and TGs) using A25 Random Access Analyzer (BioSystems TM , Spain).Serum glucose level was measured by the glucose oxidase method (GOD-PAP).Enzymatic colorimetric assay method was used for the measurement of TC (CHOD-PAP method) and TG (GPO-PAP method) while direct homogeneous enzymatic colorimetric assay technique was used for the measurement of HDL-c and LDL-c.The required reagents for all these tests were from Human Gesellschaft fϋr Biochemica und Diagnostica mBH (Germany).
Finally, lipid profile derangements was defined as TC ≥200 mg/dl, HDL-c <40 mg/dl, LDL-c ≥130 mg/dl and TG ≥150 mg/dl, while a fasting glucose level ≥110 mg/dl for dysglycaemia according to the United States National Cholesterol Education Program, Adult Treatment Panel (NCEP-ATP) III guidelines (NCEP-III, 2002).To assess cardiovascular risk, Castelli's Index I was calculated, TC/HDL-c ratio: a Castelli's Index I >5.1 for men and >4.4 for women were considered indicative of an elevated CVD risk (Castelli et al., 1983).

Statistical analysis
Analysis included for all participants showing as a minimum of one visit after initiating AZT.Mean (± standard deviation, SD), median (interquartile range at 25 and 75th, IQR) and frequencies were used to describe patients' characteristic as appropriate.Chi-square test and Mann-Whitney U test were used to compare categorical and continuous variables between the two treatment groups (EFV vs. NVP).Pair-samples T test were used to compare measures between the switch baseline and at 12 months after initiating AZT.The independent variables were evaluated with logistic regression to identify the factors that were associated with abnormal lipid profile and glucose.P-value less than 0.05 was considered as statistically significant at 95% confidence interval (CI).
*Corresponding author.E-mail: agetetadewos@yahoo.com.Tel: +251-913-175126.Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License

Ethical clearance
The study was approved by Institutional Review Board (IRB) of Hawassa University College of Medicine and Health Science.Written informed consent was obtained from all participants.

General characteristics of study participants
A total of 80 HIV-infected patients met the inclusion criteria of this study; however, 78 patients (males 31 (39.7%),females 47(60.3%))were followed until the end of this study.Data from patients who were lost to followup were not included in this analysis.Patients switched from the d4T/3TC/EFV to AZT/3TC/EFV and from d4T/3TC/NVP to AZT/3TC/NVP at the time of regimen substitutions.Of the 78 patients, 25 (32%) were in the AZT/3TC/EFV group and 53 (68%) were in the AZT/3TC/NVP group.Baseline age of the individuals was 40.9 months with standard deviation of 7.8.There was a significant differences in age ≥40 years, in CD4+ Tlymphocyte categories and duration of HAART experiences in between NNRTI groups, p=<0.05.However, median CD4+ cells count and the rest demographic characteristics were not significantly different from the switch baseline (0 month) in between NNRTIs, p ≥ 0.05 (Table 1).

Characteristics of the lipid profiles and glucose abnormalities at the switched 0 month
Lipid profile tests (TC, TG, HDL-c, and LDL-c), and glucose were performed for a total of 78 participants.The mean TC, and TC/HDL-c ratio were higher among patients using NVP based regimen when compared with patients using EFV, but not significant.The median (IQR) of TG/HDL-c ratio was 4.3 (3.0 to 10.6) in EFV group and 4.4 (2.6 to 6.9) in NVP group (p=0.52).TG/HDL-c ratio (≥2.4), is an indicator of insulin resistance disorder, which was 84% in EFV group and 81% in NVP group (p=0.76).The mean LDL-c and median value of TG was insignificantly higher in EFV group when compared with NVP group (Table 1).Also, the prevalence of lipid profiles TC ≥ 200 mg/dl and TG ≥150 mg/dl were higher in NVP group when compared with EFV group, but the difference was not significant.
Moreover, the prevalence of HDL-c is below 40 mg/dl, glucose ≥110 mg/dl Castelli's index I and LDL-c ≥130 mg/dl were higher in EFV group when compared with those on NVP (Table 1).Derangement of four lipid profiles within a single individual (TC, TG, HDL-c, LDL-c) according to NCEP-ATP III was 12% in EFV group and 15.1% in NVP group, (p=0.71).The proportion of patients with serum glucose level >180 mg/dl were 12% in EFV group whereas 3.7% in NVP group.

Characteristics of lipid profiles and glucose after 12 month of AZT/3TC replacement
The mean TC, TC/HDL-c ratio, LDL-c, glucose and median TG were higher among patients using EFV when compared with patients using NVP, but not significant.However, the mean HDL-c value was higher among patients using NVP when compared with those using EFV (Figure 1).
In addition, the prevalence of TC ≥200 mg/dl, LDL-c ≥130 mg/dl, TG ≥150 mg/dl, HDL-c <40 mg/dl, Castelli's index I and glucose ≥110 mg/dl, were insignificantly higher in EFV group when compared with those on NVP (Figure 2).About 74.4% patients had at least two lipid profile laboratory abnormalities, which is compatible with a diagnosis of dyslipidemia.However, derangement of four lipid profiles within a single individual was higher in EFV group (16%) when compared with NVP group (3.8%), p=0.08.The prevalence of patients with serum glucose level >180 mg/dl was 8.0% in EFV group, whereas 7.5% in NVP.Furthermore, the percentage change of parameters between EFV versus NVP is depicted as shown in Table 2.

Lipid profiles and glucose between switch 0 month versus 12 month of post switch within the same NNRTI based regimen
Patients switched from d4T/3TC/EFV to AZT/3TC/EFV showed insignificant differences in mean and median value of lipid profiles (TC, TG, LDL-c and TC/HDL-c ratio), p≥0.05.The proportions of TC ≥ 200 mg/dl, TG ≥150 mg/dl, HDL-c <40 mg/dl, Castelli's index I and glucose ≥110 were increased at the 12 month of post switch.In contrast, those patients switched from d4T/3TC/NVP to AZT/3TC/NVP showed a significant decrease in mean TC and median TG (p = 0.001 for both), and also TC ≥200 was significantly decreased from 49 to 16% (p=0.04).Moreover, the proportion of Catelli's index I, an indicator of the cardiovascular diseases risk, was decreased from 33.9 to 28.3% (Table 3).
Univariate and multivariate analysis were applied to assess possible predicted factors associated with each abnormal lipid profile among patients within NVP regimen.In both models, sex was significantly and negatively the associated risk factor of raised TC and TGs (Table 4).

DISCUSSION
The aim of this cohort study carried out in a resource limited setting was to assess the trend of lipid profile derangements and dysglycaemia among HIV-infected patients switched from d4T/3TC to AZT/3TC based regimens without changing NNRTIs.Most HAART drugs have been found to induce moderate to severe toxic effects after long-term use and therefore pose a tackle to chemotherapy (Sharma, 2011).Patients are switched to a new regimen because of the adverse effects of d4T therapy, but not due to immunological failure.It was found that the CD4+ cells count in the EFV group was not significantly different from the NVP group, thus shows the EFV is similar in immunological properties in HIV-infected patients when compared with NVP.
In the present study, majority of the patients (74.4%) had at least two laboratory abnormalities, which is compatible with a diagnosis of dyslipidemia according to NCEP-ATP III criteria at the end of the study follow-up.HIV-1 infection itself or use of HAART may induce oxidative stress and it predisposes for further pathogenesis in HIV infected patients (Sharma, 2014).The association between dyslipidemia and HAART has been mainly described for PIs based regimens (Anastos et al., 2007;Nery et al., 2011).Lipid derangements were higher among patients who received d4T when compared with other NRTIs (Kalyanasundaram et al., 2012;Ceccato et al., 2011).NNRTIs derange lipid profiles during therapy (Van der Valk et al., 2001;Young et al., 2005); however, evidences in support of the characterization of dyslipidemia regarding NNRTIs after HAART switching in sub-Saharan African countries are scarce.Similar with our findings: a cohort study report from  Germany revealed that a substitution of d4T with other NRTIs may reverse lipid derangements (Claas et al., 2007).It was found out that the proportions of TC ≥200 mg/dl, TG ≥150 mg/dl, LDL-c ≥130 mg/dl, TC/HDL-c ratio and HDL-c <40 mg/dl were higher in EFV group (48, 76, 32, 48 and 32%) when compared with NVP group (30.2, 62.3, 24.5, 28.3 and 30.2%), respectively.These described abnormal lipid profiles in EFV group are atherogenic (NCEP-III, 2002;Asztalos et al., 2006;Sudano et al., 2006), and suggest a potential risk for the development of cardiovascular diseases in a significant proportion among HIV-infected patients in the near future.The randomized trial report of 2 non-nucleosides (2NN) indicated that patients on NVP group had significantly improved HDL-c concentration and had relatively low lipid profile derangements when compared with those on EFV (van Leth et al., 2004).Also, the finding of the present study indicates HDL-c concentration was slightly higher in the NVP group when compared with EFV, hence it confirms that ART regimen which contains NVP has antiatherogenic effects (Van der Valk et al., 2001;van Leth et al., 2004), in addition to restoration of patients' health.
Furthermore, the trends of lipid profile within a single group (EFV in NRTI switch 0 month versus EFV at the 12 month of post switch; and NVP in NRTI switch 0 month versus NVP at the 12 month of post switch) was checked.Patients in the EFV group had raised TC/HDL-c, TG, TC and TC/HDL-c when compared with the switch baseline.However, a significant decreasing trend was observed in the mean TC, median TG and TC ≥200 mg/dl in NVP group at the end of the study follow-up when compared with EFV.Similarly, it has been stated that EFV has a deleterious effect on lipids when compared with NVP (Erdembileg et al., 2009;Tungsiripat et al., 2005).These variations may be due to patient characteristics such as life style, gender and race/ethnicity, drug metabolism polymorphism which affect differences in lipid profile between populations taking the same antiretroviral drug (Armstrong et al., 2011).Also treatment duration may contribute to these differences (Tomazic et al., 2004;Jevtovic et al., 2009).
In the present study, univariable and multivariable analysis were applied to assess possible predicted factors associated with each abnormal lipid profile among patients within NVP group.In both models, sex was a significantly and negatively associated risk factor for raised TC and TGs; but the findings are not in line with the cross-sectional study conducted in Cameroon (Pefura Yone et al., 2011).Antiretroviral-specific risk factors for glucose abnormalities include the exposure to PIs and to certain NRTIs.Of these, the use of PIs has emerged as the strongest risk factor, with studies from the early HAART era suggesting a prevalence of 8 to 46% for the spectrum of glucose metabolism abnormalities, which includes, impaired glucose tolerance, insulin resistance and diabetes in patients receiving PIs (Behrens et al., 1999;Mauss et al., 1999).The present study finding indicated no significant differences between EFV and NVP regarding glucose abnormality (≥110 mg/dl); however, the abnormal rate was decreased in both groups after 12 months of post switch.A substitution of d4T to AZT may reverse lipid derangements and it may provide a good opportunity to improve dysregulation of glucose in HIVtreated patients.A high TG/HDL-C ratio ≥2.4 is a strong indicator of the insulin resistance syndrome (NCEP-III, 2002;Einhorn et al., 2003;McLaughlin et al., 2003), which was higher in EFV group when compared with NVP but not significant (84% versus 71.7%; p=0.27).So, a wide-ranging analysis of the accessible literature on the toxicity of ARV drugs, their mechanisms of action and possible management strategies are mandatory to combat such complications (Sharma, 2011).

Conclusion
Our study indicates a decreased proportion of abnormal TC, TG, HDL-c, LDL-c, TC/HDL-c ratio and glucose among patients using NVP when compared with EFV at the end of the study follow-up according to NCEP-ATP III.Also raised HDL-c was seen in NVP group.Based on these findings, HAART regimens which contain NVP may be expected to reduce the risk of cardiovascular diseases.Therefore, it is recommended that lipid profiles should be monitored periodically for the maximum benefit of patients' health management.

Figure 1 .PercentageFigure 2 .
Figure 1.Comparing of mean (standard deviation) and median (interquartile range) measures between EFV vs. NVP groups at the 12 month of NRTI switching.

Table 1 .
Baseline characteristics at the time of NRTI switching (month 0) in between EFV versus NVP.

Table 2 .
Mean percentage change of parameters in between NNRTIs after 12 months.

Table 3 .
Comparison of parameters between switch baseline and 12 months with in a group.

Table 4 .
Determinants of abnormal lipid profiles in the NVP group at the 12 month of NRTI switch.