Background
Methods
Search strategies
Inclusion criteria
Exclusion criteria
Study selection
Quality assessment tools
Data extraction process
Pretesting the data extraction tool
Outcome measures
Data synthesis and statistical analysis
Results
Quality assessment
Characteristics of included studies
Author | Country | Design | Sample size | Follow-up Period | Female | Age Median ± IQR | CD4 count Median ± IQR | Viral load (log) Median ± IQR | BMI Median ± IQR |
---|---|---|---|---|---|---|---|---|---|
Stringer et al. [50] | Zambia and Kenya | Prospective Cohort | 509 152 | 4 years | 661 | Zambia 32 (28–36) Kenya 32 (28–36) | 148 (88–211) 147 (76.5–212.5) | 5.0 (4.4–5.4) 5.3 (4.8–5.9) | 19.7 (18.3–21.9) 19.9 (17.7–22.6) |
Kwobah et al. [51] | Kenya | Case-control | 3233 | 5 years | 1992 | Case: 36.3(30.6–43.2) Control: 36.5(30.7–43.1) | Case: 80 (32–177) Control: 194 (112–324) | NA | NA |
Nachega et al. [45] | South Africa | Cohort | 1817 | 10 years | 1771 | NVP: 33.2 (7.68) EFV: 36(8.0) | NVP: 171(78–243) EFV: 136(58–216) | NVP: 5.1(4.6–5.6) EFV: 5.2(4.7–5.7) | NA |
Boulle et al. (18, 48) | South Africa | Cohort | 2679 | 4 years | 1896 | 32(28–38) | 83(34–140) | 5.5(4.6–5.5) | NA |
Shearer et al. ([47]) | South Africa | Cohort | 12,840 | 8 years | 7962 | EFV: 37.2(31.9–43.9) NVP: 33.0(28.4–39.1) | EFV: 98(36–169) NVP: 100.5(42–160) | NA | EFV: 21.7 (19.2_24.9) NVP: 22.0 (19.7_25.1) |
Sarfo et al. [52] | Ghana | Retrospective Observational | 3990 | 7 years | 2717 | EFV: 40 (35–47) NVP: 35 (30–41) | EFV: 127 (45–213) NVP: 140 (56–220) | NA | EFV: 19.6 (17.4–22.3) NVP: 19.9 (17.6–22.9) |
Shearer et al. [46] | South Africa | Cohort | 2385 | 1 year | 1485 | EFV: 37.8 (31.8–44.3) NVP: 33.1 (28.1–40.4) | EFV: 132 (58–194) NVP: 132 (66–179) | NA | EFV: 22.5 (19.7–25.9) NVP: 23.0 (20.2–25.6) |
Barth et al. [49] | South Africa | Retrospective observational cohort | 735 | 2–4 years | 526 | Mean = 35.3 | 68 (20–140) | Mean = 4.9 | NA |
Gsponer et al. [54] | Sub-Saharan Africa | Collaborative analysis | 2404 | 6 years | 1276 | Switched: 29 (25–33) Not switched: 29 (25–34) | Switched: 77 (41–133) Not switched: 146 (82–232) | NA | NA |
Sarfo et al. [10] | Ghana | Retrospective cohort study | 3999 | 7 years | 2727 | 38 (32–45) | 132 (50–217) | NA | 19.8(17.5–22.7) |
Keiser et al. [55] | Sub-Saharan Africa | Case control | 4281 | 5 years | 3022 | Switch: 35 (30–41) Not switch: 34 (30–40) | Switch: 73 (23–133) Not switch: 67 (21–161) | Switch: 5.0 (4.4–5.6) Not switch: 5.3 (4.6–5.8) | NA |
Anlay et al. [53] | Ethiopia | Retrospective follow-up | 410 | 5 years | 265 | 33.3 ± 8.7 | 162.5 (90.5–235.5) | NA | NA |
van Zyl et al. [43] | South Africa | Retrospective | 167 | 4 years | NA | NA | NA | NA | NA |
Abah et al. [42] | Nigeria | Cohort | 6309 | 3 years | 4156 | 35 (30–42) | 154 (76–259) | NA | NA |
Bock et al. [44] | South Africa | Multicentre cohort | 27,350 | 4 years | 16,828 | EFV: 35.8 (30.7–42.6) NVP: 31.1 (27.1–35.8) | EFV: 107 (51–160) NVP: 132 (76–177) | NA | NA |
Tirfe et al. [25] | Ethiopia | Retrospective cohort study | 492 | 3 years | 289 | EFV: 35(10) NVP: 33(11) | NA | NA | NA |
Treatment failure
Author | Outcomes | Intervention | Confounder adjusted | Main findings | Newcastle-Ottawa scale |
---|---|---|---|---|---|
Stringer et al. [50] | Treatment failure |
n
NVP = 355
n
EFV = 523 | Month on ART, CD4 cell count, viral load, WHO stage, age, Hgb, BMI, Weight, NNRTI, TB at baseline | -Month on ART, CD4 cell count, viral load, WHO stage, age, Hgb, and BMI were significantly associated factors. -In total, 724 women (82%) completed 48 weeks of follow-up on an NNRTI-containing regimen | Selection 2 stars Comparability 1 star Outcome 2 stars |
Kwobah et al. [51] | Treatment failure |
n
EFV = 427
n
NVP = 2633 | Education level, CD4 category, WHO stage, BMI, Hemoglobin, Adherence, disclosure, travel time, NNRTI, and NRTI | -No association between the choice of NNRTI used (Nevirapine or Efavirenz) and treatment failure -Low baseline CD4, AZT based NRTI, imperfect adherence are associated with first line ART failure | Selection 3 stars Comparability 2 stars Outcome 1 star |
Nachega et al. [45] | Virologic failure |
n
EFV = 212
n
NVP = 103 | Age, sex, race, baseline CD4, baseline VL, NRTI, year of ART, adherence | -Nevirapine was associated with greater risk of virologic failure compared to efavirenz -NNRTI, age, sex, baseline viral load, year on ART, and adherence were significantly associated with failure | Selection 2 stars Comparability 2 stars Outcome 2 stars |
NNRTI substitution |
n
EFV = 1612
n
NVP = 1067 | Weight, age, WHO stage per increment, CD4 Count, district | -Substantial difference in the tolerability of commonly used first line ART drugs. -Baseline weight, and Age, for NVP and Weigh and WHO stage for EFV | Selection 2 stars Comparability 2 stars Outcome 2 stars | |
Shearer et al. [47] | Treatment failure |
n
EFV = 11,962
n
NVP = 878 | NNRTI, ART year, sex, age, baseline CD4, WHO stage, BMI, NRTI, baseline anemia | -Patients with NVP-are more likely to experience virologic failure -NNRTI, years of ART initiation, age, and baseline CD4 cell counts | Selection 3 stars Comparability 2 stars Outcome 1 star |
Sarfo et al. [10] | Composite endpoint |
n
EFV = 2369
n
NVP = 1621 | Sex, age, NNRTI, NRTI, Baseline CD4, baseline BMI, WHO stage, adherence | -Treatment outcomes were comparable whether EFV or NVP is used -There was a 36% lower risk of all-cause discontinuation of EFV compared with NVP -NRTI, age, baseline CD4 counts, BMI, WHO stage, and adherence were associated factors of treatment failure | Selection 3 stars Comparability 2 stars Outcome 2 stars |
Shearer et al. [46] | Treatment failure |
n
EFV = 2254
n
NVP = 131 | NNRTI, sex, age, baseline CD4, WHO stage, Anemia, BMI | -Given TDF as NRTI, Nevirapine has higher risk of treatment failure as compared to EFV. -Regimen, and Baseline CD4 cell counts were significantly associated with failure | Selection 3 star Comparability 1 star Outcome 1 star |
Barth et al. [49] | Treatment failure |
n
EFV = 426
n
NVP = 309 | Gender, age, BMI, Karnofsky score, CD4 counts, time since start ART, NNRTI, Occupation | -No difference between EFV and NVP in treatment failure −60% of patients showed virological failure; only few of them were switched to second-line treatment -Gender, mean BMI, and baseline CD4 counts were associated in the univariate | Selection 2 star Comparability 1 star Outcome 2 star |
Gsponer et al. 2012 [54] | Treatment failure |
n
EFV = 186
n
NVP = 2218 | Age, Sex, baseline CD4, WHO stage | -Mortality was lower among patients who switched compared to patients remaining on failing first-line ART -Current CD4 count was associated | Selection 2 star Comparability 1 star Outcome 1 star |
Sarfo et al. [10] | NNRTI Substitution |
n
EFV = 2378
n
NVP = 1621 | NNRTI, gender, age, BMI, WHO stage, CD4 counts, hepatitis B surface antigen status, ALT | -Patients starting nevirapine are more likely to develop rashes and then more likely to discontinue therapy than those starting efavirenz. -NNRTI, gender, and low baseline CD4 counts were associated factors | Selection 3 star Comparability 2 star Outcome 1 star |
Keiser et al. [55] | Treatment failure |
n
EFV = 1951
n
NVP = 2325 | Not controlled | -Compared to patients who remained on non-failing first-line therapy, mortality and loss from follow-up was higher in patients who switched, and substantially higher in patients who remained on failing first-line therapy | Selection 2 star Comparability 1 star Outcome 1 star |
Anlay et al. [53] | NNRTI Substitution |
n
EFV = 289
n
NVP = 121 | Weight, WHO stage, TB on initial regimen, NRTI, NNRTI, Co-medication, and side effect | -There is no difference in regimen change between NVP and EFV -WHO stage, TB status, co-medication, and side effect were associated factors | Selection 3 star Comparability 2 star Outcome 2 star |
van Zyl et al. [43] | NNRTI resistance |
n
EFV = 82
n
NVP = 85 | Age, gender, ART regimen, most recent CD4 count, concurrent viral load, and genotypic resistance information | -Failure on NVP therapy may result in cross-resistance to ETV. NNRTI and estimated period of failure were associated | Selection 2star Comparability 2 star Outcome 1 star |
Abah et al. [42] | NNRT substitution |
n
EFV = 558
n
NVP = 5751 | Sex, age, HBV, CD4 count, WHO stage, NNRTI, NRTI | -Drug substitutions of efavirenz (EFV) were more likely than of nevirapine (NVP) -Age, greater immunosuppression, EFV compared to NVP, and drug toxicity were significant predictors | Selection 2 star Comparability 1 star Outcome 2 star |
Bock et al. [44] | NNRT substitution |
n
EFV = 19,441
n
NVP = 7909 | NNRTI, NRTI (AZT & D4T), gender, age, baseline CD4, WHO stage, TB treatment at baseline, year of ART initiation, level of care | -Superior efficacy of EFV compared with NVP for first-line ART -NNRTI, gender, year of initiation, and province were associated factors | Selection 3 stars Comparability 2 stars Outcome 2 stars |
Tirfe et al. [25] | Treatment failure |
n
EFV = 246
n
NVP = 246 | NNRTI, facility type, age, sex, marital status, education status, religion, NRTI, presence of OIs, eligibility criteria, WHO stage, functional status, BMI, provision of IPT, and baseline CD4 counts | -NVP and EFV based HAART regimens were effective and comparable, in term of immunological responses. -Gender, eligibility criteria, WHO stage, provision of IPT, and baseline CD4 counts were associated factors | Selection 3 stars Comparability 2 stars Outcome 2 stars |
Meta-analysis results
Evaluation for publication bias
Covariate | Estimate | S.E | 95%CI |
---|---|---|---|
Length of follow-up | − 0.0138 | 0.0666 | (− 0.1322–0.1046) |
Median CD4 count | − 0.0034 | 0.0066 | (− 0.0150–0.0079) |
Median age | 0.0415 | 0.0503 | (− 0.1013–0.0755) |
Year of publication | 0.0637 | 0.0815 | (− 0.1736–0.0980) |
Female proportion | − 0.5305 | 1.4575 | (− 3.6792–2.6183) |