The incidence of first-line antiretroviral treatment changes and related factors among HIV-infected sex workers in Nairobi, Kenya

The incidence of first-line antiretroviral treatment changes and related factors among HIV-infected sex workers in Nairobi, Kenya

Frank Ndaks Ndakala^1,2,&, Julius Otieno Oyugi^1,3, Margaret Ng’wono Oluka⁴, Joshua Kimani^1,3, Georg Martin Norbert Behrens^5,6

 

¹University of Nairobi, Institute of Tropical and Infectious Diseases (UNITID), Nairobi, Kenya, ²State Department of Science and Technology, Directorate of Research Management and Development (DRMD), Nairobi, Kenya, ³University of Manitoba, College of Medicine, Department of Medical Microbiology, Winnipeg, Manitoba-Canada, ⁴University of Nairobi, College of Health Sciences, School of Pharmacology and Pharmacognosy, Nairobi, Kenya, ⁵Hannover Medical School, Department of Clinical Immunology and Rheumatology, Hannover, Germany, ⁶German Centre for Infection Research, Germany

 

 

^&Corresponding author
Frank Ndaks Ndakala, University of Nairobi, Institute of Tropical and Infectious Diseases (UNITID), State Department of Science and Technology, Directorate of Research Management and Development (DRMD), Nairobi, Kenya

 

 

 

 

The pressing need to scale-up combined antiretroviral therapy (cART) in resource-poor settings has resulted in a growing number of patients receiving simple standardised doses of the first-line regimens [1]. A majority of the 24.0 to 28.7 million people infected with human immunodeficiency virus (HIV) live in the resource-poor setting of sub-Saharan Africa [2]. Kenya has the fourth-largest HIV epidemic in the world with over 1.6 million HIV-infected people. Out of 880 000 adults eligible for cART, 656 359 Kenyans are on cART [3]. In spite of the current deficit in treatment with cART, several factors remain a hindrance to the improvement of long-term availability of HIV treatment in Kenya.

 

Factors such as adverse drug reactions (ADRs), clinical failure, non-adherence to treatment, and concomitant treatment, contribute to the change of the original first-line cART [4-6]. According to the World Health Organization (WHO), the burden of ADRs and regimen changes in sub-Saharan Africa is further complicated by limited antiretroviral treatment choices [7]. In these settings, patients stay on the first-line regimens as long as possible [8, 9]. This is because consecutive regimens are expensive, less effective and patients may not last long on them due to new possible side effects [9]. Studies conducted in developed countries recorded median time that patients stay on a first-line regimen as 1.6 years and 1.8 years [10, 11]. In the Eastern African region, a higher median time to ART change of 3.1 years was recorded in Kenya [12]. In addition, separate studies conducted in Kenya and Ethiopia revealed incidence rates of 18.6/100 person-years (PY) and 10.11/100 PY respectively [13, 14].

 

Other factors that were observed to influence the change of first-line regimens included WHO stage, low CD4 count, patient’s weight and treatment failure [9, 13]. Since clinicians do not conduct routine viral load measurements to guide the diagnosis of treatment failure, HIV-infected patients in Kenya frequently encounter delayed cART drug changes and are at an increased risk of viral resistance. As a result, the overall financial cost of cART to HIV-infected people could increase, thereby jeopardising long-term prognosis. Due to these limitations, designing strategies aimed at prolonging the use of the original cART regimen among the less privileged HIV-infected populations is essential. However, there is insufficient information that can inform such strategies. This study describes the incidence of and factors influencing the first-line cART regimen changes in a cohort of adult Kenyan sex workers.

 

 

Ethical considerations

 

The study received ethical approval from Kenyatta National Hospital-University of Nairobi-Ethics Review Committee. The sex workers outreach programme (SWOP) administration granted access to the electronic and hard copy medical files in the clinics.

 

Setting

 

The study setting was at the sex workers outreach programme (SWOP) clinics in Nairobi, Kenya. The SWOP cohort of sex workers consists of female commercial sex workers and men who have sex with men (MSM) for commercial reasons. The SWOP primary health care facilities offer treatment to HIV-infected sex workers who live in Nairobi slums. The University of Manitoba from Canada and the University of Nairobi jointly manage the SWOP clinics. Since the start of the SWOP facilities, about 15 000 HIV-infected people have been initiated to cART. The SWOP sites selected for the study were in proximity to each other to enable sampling of patients from the same population.

 

Study design and population

 

This was a retrospective analysis of electronic medical records of HIV-1 infected sex workers receiving cART in five SWOP clinics. The study characterised the cohort as cART-naive adult sex workers aged ≥ 18 years, who started the first-line cART between January 2009 and December 2013 and had a minimum of one follow-up visit record. The SWOP clinics started cART for people with severe or advanced HIV clinical disease (WHO clinical stage III or IV) and in people with CD4 count ≤ 350 cells/mm³. Before 2008, the first-line cART regimens consisted of either stavudine (d4T) or zidovudine (AZT) combined with lamivudine (3TC) and efavirenz (EFV) or nevirapine (NVP). Since 2009, the first-line regimen consisted of AZT or d4T or TDF, with 3TC and either NVP or EFV. However, in 2010, patients already on AZT or d4T containing regimens continued using them unless toxicity developed. All patients had their follow-up visits after two weeks on starting cART, monthly after stabilising and every quarter for monitoring of treatment outcomes. Clinicians measured and reviewed every quarter the CD4 cell counts, haematological, hepatological and renal function defects in patients who had stabilised on cART. Because viral load tests are costly, conducting these tests was rare. Additionally, during each visit, clinicians collected the patients’ details on standardised forms to support data transcription in the Kenyan AIDS Control database.

 

Data collection

 

The study utilised prospectively maintained electronic medical records sourced from Kenya AIDS Control Database. Before extracting electronic files, there was a thorough examination of the information on the medical records. Extraction of any missing information was by referring to the hard copies of medical files accessed from the clinics. Using a simple random sampling technique, the study independently selected about 1 500 medical records from the electronic database as a sampling frame. To conduct a simple systematic random sampling, there was a random selection of the first record from the first four records in the sampling frame and subsequent addition of every fourth record. The study collected independent variables that included demographic and anthropometric characteristics including age, sex, weight (kg) height (m) and body mass index (BMI). Other variables collected in the study were clinical and laboratory variables, including ADRs, CD4 cell counts, pregnancy status, creatinine, alanine aminotransferase (ALT) / aspartate aminotransferase (AST) and haemoglobin (HB); treatment variables, including the type of cART regimen, cART initiation and regimen change dates, reasons for change, new cART regimens, treatment outcomes and dates of outcomes. The clinicians used calibrated scales to measure all the anthropometric measurements. Measurements of current weight (kg) and height (cm) were taken in duplicate based on procedures recommended by WHO [15]. The analysis used the arithmetic means obtained from the two measurements then calculated body mass index (BMI) by dividing the weight in kg by the square of the height in metres.

 

Exposure and outcomes

 

The exposures in the study were age at cART initiation, sex, CD4 cell counts, HB, cART regimen type, and period on cART. The primary outcome during this analysis was the initial change from the first-line antiretroviral treatment characterised as an individual drug substitution involving d4T, AZT, and TDF regimens or a regimen switch to second-line treatment containing a protease inhibitor-based regimen [16]. Censoring occurred at the time of the first incident of a drug change from the first-line cART with the focus on changes connected to d4T, AZT, and TDF. The reasons for a drug change covered adverse drug reactions (ADRs), treatment failure (considered as an immunologic failure or clinical failure or virologic failure, or any combination of the three), and other reasons. The study defined immunological failure as a fall in CD4 cell count to 50% from the on-treatment peak value or persistent CD4 cell count of < 100 cells or fall of CD4 counts to pre-therapy baseline. Patients with clinical failure showed new or recurrent adverse events indicative of WHO stage IV condition after 6 months of effective treatment. Patients with virologic failure showed plasma viral load of > 1 000 copies/ ml based on two consecutive viral load measurements after 3 months, with adherence support. The reasons for drug change classified as other reasons included drug changes occasioned by defaulted treatment, tuberculosis, pregnancy and drug stoke-out. The analysis used the Kenya National Clinical Manual to define ADRs [17]. The manual defines an ADR as any severe and life-threatening effect of antiretroviral (ARV) drugs presented at normal doses. Polyneuropathy consisted of symptoms that included tingling, burning, pain, numbness, weakness in limbs, paraesthesia, muscle weakness, and inability to walk. Features defining lipodystrophy included a hump on the back, loss of facial fat, thinning or wasting of the limbs, and or accumulation of fat at the abdomen. The study defined hepatological disorder based on the amount of liver transaminase enzymes detected in the blood. Hepatological disorder included patients with > 40 IU/L of alanine aminotransferase (ALT) or aspartate aminotransferase (AST). The amount of haemoglobin (HB) present in the blood helped to characterise patients with the haematological disorder. This group of patients consisted of individuals with < 11 g/dl of HB, without a history of anaemia or > 1 g / dl drop in HB after beginning cART [18]. Patients with a rash had erythema, pruritus, and severe mucous membrane involvement (Stevens-Johnson syndrome). Patients with renal abnormality composed of cases with > 150 µmol / l of serum creatinine after beginning cART. Baseline characteristics included measurements of CD4 cells, HB, ALT, AST, weight, height, BMI and creatinine measured closest to cART starting date (within 6 months before up to 1 week after cART initiation).

 

Statistical analysis

 

The study performed data analysis using SPSS statistics version 17.0. Chicago: SPSS Inc. (USA). Medians and interquartile ranges described variables with a skew distribution while percentages described categorical variables. The calculations of drug-specific incident rates were in the form of rates per individuals’ initiating the precise drug. Estimating survival time to first drug change was through Kaplan-Meier survival analysis [19]. Censoring of patients targeted the time of the first incident of a drug change from the first-line cART. Modelling for the risk factors related to drug changes employed the Cox Proportional Hazards analysis. The model covered the following variables: sex, age at cART initiation, height, weight, BMI, HB, CD4 cell count, AST, ALT, creatinine, types of cART regimens and calendar year of cART initiation. The analysis considered significant factors with a p - value of 0.05.

 

 

Out of the eligible 1 500 medical records sampled, the analysis excluded 50 medical records of patients with duplicate data sets. The demographic characteristics of the remaining 1 450 eligible medical records showed that 77.8% of the patients in the study were female commercial sex workers. The rest of the study population consisted of male commercial sex workers (men who have sex with men (MSM)). Overall, most of the patients started combination antiretroviral therapy (cART) in 2009 and 2010 (Table 1). About 38% of the patients started on AZT-based regimens, 36% on d4T-based regimens, and 26% on TDF-based regimens. Compared to patients who started either AZT or TDF-based cART, patients who initiated to d4T containing cART had a lower median CD4 cell count, which reflected a greater likelihood of HIV-1 disease progression.

 

Incidence of first-line cART drug changes

 

The study observed an overall incident rate of total drug changes of 11.1 per 100 person-years (P/Ys) within a total follow-up period of 3 427.9 person-years (Table 2). Out of 380 patients who changed cART, 370 (97%) had a drug substitution and 10 (3%) switched regimens. Compared to patients using TDF (n = 12) and AZT (n = 25), 63.8% of d4T users (n = 333) substituted regimens, reflecting a higher proportion of total drug changes. Treatment switches from first to second-line drugs accounted for 2.6% (n = 10) of drug changes. The cumulative P/Ys of exposure for patients receiving TDF and AZT-based cART were 783.9 P/Ys and 1 393 P/Ys respectively (Table 2). The rates of drug changes were highest among patients initiating d4T-based regimens [27.0/100 PYs (95% CI: 22.8-27.4)] compared to either AZT [1.9/100 PYs (95% CI: 0.4-2.0)] or TDF [2.0/100 PYs (95% CI: 0.9-4.5)]. Table 3 shows the reasons for the first incident of drug changes from the first-line cART as captured by the clinicians (Table 3). The clinicians cited ADRs as the main reason for a drug change, accounting for 75.5% of the captured drug changes. The overall prevalence rate of ADRs was 34.9%. The prevalence rates of specific ADRs that included lipodystrophy, polyneuropathy, haematological disorder, hepatological disorder, rash and renal abnormality were 41.7% (n = 211), 29.4% (n = 149), 15.4% (n = 78), 9.3% (n = 47), 2.8% (n = 14) and 1.4% (n = 7) respectively. Most patients who substituted d4T had lipodystrophy and polyneuropathy. Twelve patients out of the 78 patients with the haematological disorder had a drug substitution (Table 3). Individual drug substitutions from one first-line to a different first-line regimen were the key drivers of cART drug changes attributable to ADRs. However, the reason cited for treatment switches from a first-line cART regimen to a second-line cART regimen was treatment failure.

 

Risk factors of antiretroviral drug changes

 

A Kaplan-Meier plot presented in Figure 1 illustrates the time to the first incident of drug changes from the first-line cART stratified by backbone first-line antiretroviral drug regimens (Figure 1). Throughout follow-up time, there was a gradual rise in cART drug changes attributable to ADRs, treatment failure and other reasons that included pregnancy, tuberculosis and defaulted treatment. Since d4T accounted for up to 79.3% of cART drug changes, the survival time on d4T significantly related to a high risk of drug changes compared to TDF and AZT (p < 0.001). However, the graph representing AZT regimen was close to that of TDF regimen, implying the difference in survival time to the first incident of a drug change between the two regimens was insignificant (p = 0.657). After 12 months of cART, the percentage of patients who changed either TDF, AZT or d4T were 0.6%, 0.3%, and 4.6% respectively. However, after 60 months of cART use, about 1.8% of patients on AZT had a drug change compared to 23.3% of d4T users. Further, the study performed both univariate and multivariate analyses of time to the first incident of drug changes from first-line cART (Table 4). In a univariate analysis, an increased hazard of drug changes related to CD4 cell counts of ≥ 200 cells/mm³ ((hazards ratio 4.51; 95% CI: 3.63-5.60), creatinine ((hazards ratio 1.01; 95% CI: 1.00-1.01), d4T ((hazards ratio 12.4; 95% CI: 7.51-20.5), and the calendar year 2010 ((hazards ratio 6.24; 95% CI: 2.3-16.9). To introduce BMI, HB, AST and ALT as continuous variables in the Cox model, the SPSS analysis used standardised residuals to check for linearity before transforming the data by Log₁₀ transformation. To avoid over-adjustment, the multivariate model did not adjust for height and weight. Instead, after adjusting for BMI, a statistically insignificant increased risk of cART drug changes related to BMI (hazards ratio 1.0; 95% CI: 0.99-1.00), AST (adjusted hazards ratio 1.34; 95% CI: 0.70-2.57) and ALT [adjusted hazards ratio 1.17; 95% CI: 0.66-2.60]. In addition, using AZT as a comparator, d4T regimen independently associated with a significantly increased risk of cART drug changes [adjusted hazards ratio 10.16; 95% CI: 6.02-17.16], while the risk among TDF users was low (adjusted hazards ratio 0.83; 95% CI: 0.44-1.55).

 

 

The present findings report an overall incident rate of total drug changes of 11.1 per 100 person-years (P/Ys) within a total follow-up period of 3 427.9 PYs. Specifically, the incident rates of total drug changes for patients on AZT, TDF and d4T were 1.9 per 100 P/Ys, 2.0 per 100 PYs and 27 per 100 PYs respectively. Most patients who substituted d4T had symptoms of lipodystrophy and polyneuropathy. In addition, an increased risk of drug substitution was independently associated with d4T-based regimens (adjusted hazards ratio 10.16; 95% CI: 6.02-17.16).

 

The overall incident rate of cART drug changes reported in the current study is comparable to findings of other studies conducted in Kenya. Studies conducted in Western Kenya reported incident rates of cART drug changes as 18.6 per 100 PYs and 13.3 per 100 PYs [5, 13]. However, the Swiss cohort study which had a higher incident rate of 41.5 per 100 PYs [20] is indicative of higher incident rates of cART drug changes observed in developed countries [10, 21]. The observed variations in incident rates of cART changes could be due to several reasons including variations in study methodologies, treatment monitoring, and selection bias. In addition, the difference could be because the present study only considered ADRs leading to cART change. Many studies have reported ADRs, as the main reason for changing cART in several resource-poor settings [22, 23]. The current findings support two Kenyan studies that cited cART-related toxicity as the main reason for changing treatment [13, 24]. This study further reports that patients who substituted cART had lipodystrophy and polyneuropathy which accounted for 41.7% and 29.4% of the total ADRs recorded. However, comparisons between these studies remain a challenge because of variations in ADR reporting ways, study periods, types of the cART regimens in use [25, 26] and the use of concurrent medication [27]. There were fewer incident rates connected to the switch of a first-line cART regimen to a second-line regimen, suggesting limited treatment options available in this setting [9]. Alternatively, since there was no data on patients who failed to switch to second-line therapy due to loss to follow-up, this study may have underestimated the number of patients who switched regimens.

 

Despite the high prevalence of ADRs including lipodystrophy and polyneuropathy, the use of d4T-based cART was common in the SWOP clinics. Compared to AZT and d4T, TDF showed signs of a well-tolerated drug regimen with fewer drug changes. Studies with comparable findings exist in Zambia and South Africa [28-31]. In Uganda and Swaziland, the main causes of d4T-based drug substitutions included lipodystrophy and polyneuropathy, whilst anaemia caused most of the AZT-based drug substitutions [32, 33]. Similarly, the present study had a significantly higher proportion of patients who needed d4T substitutions. Most substitutions were in line with the WHO’s recommendations of 2010 [7]. Though different studies have associated lower rates of TDF substitutions with renal impairment [34, 35], these findings did not confirm the association.

 

In a multivariate analysis, an increased hazard of drug changes related to the d4T regimen. This further supports evidence in the literature that relates TDF to a good safety record [29, 32, 36]. Older age is a well-known risk factor in patients with polyneuropathy, lipodystrophy and renal toxicity [7]. However, these findings found an insignificant association between older age and an increased hazard of drug substitution. Similar contrasting data exist in the literature, but they have been less consistent [29, 32]. Study limitations included the poor recording of ADRs and inadequate reporting by the health care practitioners. It is possible that some ADRs were mistaken with effects of advanced HIV-1 infection, concomitant medications, or co-morbidities. In addition, by restricting the analysis to the first incident of cART drug changes alone, the study may have underestimated the true number of the overall drug changes. However, this study had a robust data set, with adequate sample size and minimal missing data in the descriptive variables.

 

 

This study reports a moderate incidence of cART changes among HIV-infected sex workers from Nairobi, Kenya. The most common reason identified for changing cART was the presentation of adverse drug reactions. Most individuals using stavudine were at a higher risk of experiencing a change in their cART due to lipodystrophy and polyneuropathy, mostly presenting within 20 months after the start of treatment.

 

 

The authors declare no competing interest.

 

 

F.N.N. (Directorate of Research Management and Development), was responsible for data acquisition, analysis and drafting of the manuscript. J.O.O. (University of Nairobi) and M.N.O. (University of Nairobi) critically revised the manuscript. J.K. (University of Nairobi) facilitated data abstraction and critically reviewed the manuscript. G.M.N.B. (Hannover Medical School) contributed to study conception and design, interpretation of data and critically reviewed the manuscript. All authors have read and agreed to the final version of this manuscript.

 

 

The authors would really like to acknowledge the technical support provided by Dr. Joshua Kimani, the clinical director and the employees in SWOP clinics. We would additionally wish to give thanks to Dr. Miriam Poirier, of the United States National Cancer Institute, Carcinogen-DNA Interactions Section for her comments and suggestions. This study received valuable support from the University of Manitoba’s International Infectious Diseases & Global Health Training Programme (IID & GHTP) and was part of a Ph.D. thesis at the University of Nairobi, Institute of Tropical and Infectious Diseases (UNITID).

 

 

Table 1: characteristics of HIV-infected patients initiated on ART in SWOP clinics from January 2009-December 2013 (n = 1 450)

Table 2: rates of d4T, AZT and TDF 1st drug changes during follow-up

Table 3: reasons for changing treatment from d4T, AZT, and TDF during ART in HIV-infected Kenyan patients

Table 4: cox hazards model of risk factors for duration to first cART drug changes in SWOP clinics

Figure 1: Kaplan-Meier analyses of time to first AZT, d4T and TDF drug change among HIV-infected patients in SWOP clinics

 

 

1. Calmy A, Pinoges L, Szumilin E, Zachariah R, Ford N, Ferradini L. Generic fixed-dose combination antiretroviral treatment in resource-poor settings: multicentric observational cohort. AIDS. 2006; 20(8): 1163-9. Google Scholar


2. United Nations Programme on HIV/AIDS. The Gap Report. Geneva: UNAIDS. 2014; 1-83. Google Scholar


3. National AIDS Control Council. Kenya AIDS Response Progress Report Progress towards Zero. NACC, Ministry of Health, Government of Kenya. 2014; 37. Google Scholar


4. Pádua CA, César CC, Bonolo PF, Acurcio FA, Guimarães, MD. High incidence of adverse reactions to initial antiretroviral therapy in Brazil. Braz J Med Biol Res. 2006 Apr; 39(4): 495-505. PubMed | Google Scholar


5. Braitstein P, Ayuo P, Mwangi A, Wools-Kaloustian K, Musick B, Siika A, & Kimaiyo, S. Sustainability of first-line antiretroviral regimens: findings from a large HIV treatment program in western Kenya. J Acquir Immune Defic Syndr. 2010 Feb; 53(2): 254-9. PubMed | Google Scholar


6. Woldemedhin B, Wabe NT. The Reason for regimen change among HIV/AIDS patients initiated on first-line Highly Active Antiretroviral Therapy in Southern Ethiopia. N Am J Med Sci. 2012 Jan; 4(1): 19-23. PubMed | Google Scholar


7. World Health Organization. Antiretroviral Therapy for HIV Infection in Adults and Adolescents. WHO, Geneva, Switzerland. 2010. Google Scholar


8. Sungkanuparph S, Manosuthi W, Kiertiburanakul S, Piyavong B, Chumpathat N, Chantratita W. Options for a second-line antiretroviral regimen for HIV type 1-infected patients whose initial regimen of a fixed-dose combination of stavudine, lamivudine, and nevirapine fails. Clin Infect Dis. 2007 Feb 1;44(3):447-52. PubMed | Google Scholar


9. Orrell C, Harling G, Lawn SD, Kaplan R, McNally M, Bekker L-G, Wood R. Conservation of first-line antiretroviral treatment regimen where therapeutic options are limited. Antivir Ther. 2007; 12(1): 83-8. PubMed | Google Scholar


10. The Australian HIV Observational Database. Rates of combination antiretroviral treatment change in Australia, 1997-2000. HIV Med. 2002 Jan; 3(1): 28-36. PubMed | Google Scholar


11. Chen RY, Westfall AO, Mugavero MJ, Cloud GA, Raper JL, Chatham AG, Saag MS. Duration of highly active antiretroviral therapy regimens. Clin Infect Dis. 2003 Sep 1; 37(5): 714-22. PubMed | Google Scholar


12. Kwobah CM, Mwangi AW, Koech JK, Simiyu GN, Siika AM. Factors Associated with First-Line Antiretroviral Therapy Failure amongst HIV-Infected African Patients: A Case-Control Study. World Journal of AIDS. 2012; (2): 271-278. Google Scholar


13. Inzaule S, Otieno J, Kalyango J, Nafisa, L, Kabugo C, Nalusiba J, Karamagi C. Incidence and predictors of first-line antiretroviral regimen modification in western Kenya. PLoS One. 2014 Apr 2; 9(4): e93106. PubMed | Google Scholar


14. Anlay DZ, Alemayehu ZA, Dachew BA. The rate of initial highly active antiretroviral therapy regimen change and its predictors among adult HIV patients at University of Gondar Referral Hospital, Northwest Ethiopia: a retrospective follow-up study. AIDS Res Ther. 2016 Feb 17;13:10. PubMed | Google Scholar


15. World Health Organization. Physical status: the use and interpretation of anthropometry. WHO, Geneva, Switzerland. 1995. Google Scholar


16. World Health Organization. HIV / AIDS Programme Antiretroviral Therapy for HIV infection in adults and adolescents: Recommendations for a public health approach 2006 revision important: Addendum to 2006 WHO guidelines on antiretroviral therapy for HIV infection. WHO, Geneva, Switzerland. 2006. PubMed | Google Scholar


17. National AIDS and STIS Control Programme. Kenya national clinical manual for ART providers: a concise and practical guide to ART provision. Nascop, Ministry of Health, Government of Kenya. 2007; 1-70. Google Scholar


18. van der Klauw MM, Goudsmit R, Halie MR, van’t Veer MB, Herings RM, Wilson JH, Stricker BH. A population-based case-cohort study of drug-associated agranulocytosis. Arch Intern Med. 1999 Feb 22; 159(4): 369-74. PubMed | Google Scholar


19. Kaplan EL, Meier P. Nonparametric Estimation from Incomplete Observations. Journal of the American Statistical Association. 1958; 53(282): pp 457. Google Scholar


20. Elzi L, Marzolini C, Furrer H, Ledergerber B, Cavassini M, Hirschel B, Swiss HIV Cohort Study et al. Treatment modification in human immunodeficiency virus-infected individuals starting combination antiretroviral therapy between 2005 and 2008. Arch Intern Med. 2010 Jan 11; 170(1): 57-65. PubMed | Google Scholar


21. Dieleman JP, Jambroes M, Gyssens IC, Sturkenboom MC, Stricker BH, Mulder WM, ATHENA Study Group. Determinants of recurrent toxicity-driven switches of highly active antiretroviral therapy: The ATHENA cohort. AIDS (London, England). 2002; 16(5): 737-45. Google Scholar


22. Maggiolo F, Arici C, Airoldi M, Ripamonti D, Quinzan G, Gregis G, Ravasio V, Bombana E, Suter F. Reasons for discontinuation of nevirapine-containing HAART: results from an unselected population of a large clinical cohort. J Antimicrob Chemother. 2007 Mar; 59(3): 569-72. PubMed | Google Scholar


23. Forna F, Liechty CA, Solberg P, Asiimwe F, Were W, Mermin J, Behumbiize P, Tong T, Brooks JT, Weidle PJ. Clinical toxicity of Highly Active Antiretroviral Therapy in a home-based AIDS care program in rural Uganda. J Acquir Immune Defic Syndr. 2007 Apr 1; 44(4): 456-62. PubMed | Google Scholar


24. Hawkins C, Achenbach C, Fryda W, Ngare D, Murphy R. Antiretroviral durability and tolerability in HIV-infected adults living in urban Kenya. J Acquir Immune Defic Syndr. 2007 Jul 1; 45(3): 304-10. PubMed | Google Scholar


25. Wester CW, Kim S, Bussmann H, Avalos A, Ndwapi N, Peter TF et al. The initial response to highly active antiretroviral therapy in HIV-1C-infected adults in a public sector treatment program in Botswana. J Acquir Immune Defic Syndr. 2005 Nov 1; 40(3): 336-43. PubMed | Google Scholar


26. Pádua CA, César CC, Bonolo PF, Acurcio FA, Guimarães MD. Self-reported adverse reactions among patients initiating antiretroviral therapy in Brazil. Braz J Infect Dis. 2007 Feb; 11(1): 20-6. PubMed | Google Scholar


27. Westreich DJ, Sanne I, Maskew M, Malope-Kgokong B, Conradie F, Majuba P, Funk M, Kaufman JS, Van Rie A, Macphail P. Tuberculosis treatment and risk of stavudine substitution in first-line antiretroviral therapy. Clin Infect Dis. 2009 Jun 1; 48(11): 1617-23. PubMed | Google Scholar


28. Chi BH, Mwango A, Giganti M, Mulenga LB, Tambatamba-Chapula B, Reid SE and al. Early clinical and programmatic outcomes with tenofovir-based antiretroviral therapy in Zambia. J Acquir Immune Defic Syndr. 2010 May 1; 54(1): 63-70. PubMed | Google Scholar


29. Boyle A, Sonecha S, Mandalia, S, Nelson M. An investigation into frequency and reasons why patients switch antiretroviral therapy and which antiretrovirals are commonly implicated in toxicity. J Int AIDS Soc. 2012; 15(6(Suppl 4)). Google Scholar


30. Brennan AT, Maskew M, Ive P, Shearer K, Long L, Sanne I, Fox MP. Increases in regimen durability associated with the introduction of tenofovir at a large public-sector clinic in Johannesburg, South Africa. J Int AIDS Soc. 2013 Nov 19; 16: 18794. PubMed | Google Scholar


31. Njuguna C, Orrell C, Kaplan R, Bekker LG, Wood R, Lawn SD. Rates of switching antiretroviral drugs in a primary care service in South Africa before and after the introduction of tenofovir. PLoS One. 2013 May 22; 8(5): e63596. PubMed | Google Scholar


32. Castelnuovo B, Kiragga A, Kamya MR, Manabe Y. Stavudine toxicity in women is the main reason for a treatment change in a 3-year prospective cohort of adult patients started on first-line antiretroviral treatment in Uganda. J Acquir Immune Defic Syndr. 2011 Jan 1;56(1):59-63. PubMed | Google Scholar


33. Evans D, Takuva S, Rassool M, Firnhaber C, Maskew M. Prevalence of peripheral neuropathy in antiretroviral therapy naive HIV-positive patients and the impact on treatment outcomes: a retrospective study from a large urban cohort in Johannesburg, South Africa. J Neurovirol. 2012 Jun; 18(3): 162-71. PubMed | Google Scholar


34. Cassetti I, Madruga JV, Suleiman JM, Etzel A, Zhong L, Cheng AK, Enejosa J. The safety and efficacy of tenofovir DF in combination with lamivudine and efavirenz through 6 years in antiretroviral-naïve HIV-1-infected patients. HIV Clin Trials. 2007 May-Jun; 8(3): 164-72. PubMed | Google Scholar


35. Bygrave H, Kranzer K, Hilderbrand K, Jouquet G, Goemaere E, Vlahakis N, Trivino L, Makakole L, Ford N. Renal safety of a tenofovir-containing first line regimen: experience from an antiretroviral cohort in rural Lesotho. PLoS One. 2011 Mar 2; 6(3): e17609. PubMed | Google Scholar


36. Prosperi MC, Fabbiani M, Fanti I, Zaccarelli M, Colafigli M, Mondi A, D’Avino A, Borghetti A, Cauda R, Di Giambenedetto S. Predictors of first-line antiretroviral therapy discontinuation due to drug-related adverse events in HIV-infected patients: a retrospective cohort study. BMC Infect Dis. 2012 Nov 12; 12: 296. PubMed | Google Scholar