{"_asset":{"@search.score":1,"Name":"Factors associated with antiretroviral treatment failure among people living with HIV on antiretroviral therapy in resource-poor settings: a systematic review and metaanalysis","AlternateTitleCommaDelimited":null,"Id":"4813","AssetLocationUri":null,"Abstract":"Background\r\nDespite the increase in the number of people accessing antiretroviral therapy (ART), there is limited data regarding treatment failure and its related factors among HIV-positive individuals enrolled in HIV care in resource-poor settings. This review aimed to identify factors associated with antiretroviral treatment failure among individuals living with HIV on ART in resource-poor settings.\r\n\r\nMethods\r\nWe conducted a comprehensive search on MEDLINE (PubMed), Excerpta Medica Database (EMBASE), Cochrane Central Register of Controlled Trials (CENTRAL), World Health Organization’s (WHO’s) library database, and Latin American and Caribbean Health Sciences Literature (LILACS). We included observational studies (cohort, case-control, and cross-sectional studies) where adolescents and adults living with HIV were on antiretroviral treatment regardless of the ART regimen. The primary outcomes of interest were immunological, virological, and clinical failure. Some of the secondary outcomes were mm3 opportunistic infections, WHO clinical stage, and socio-demographic factors. We screened titles, abstracts, and the full texts of relevant articles in duplicate. Disagreements were resolved by consensus. We analyzed the data by doing a meta-analysis to pool the results for each outcome of interest.\r\n\r\nResults\r\nAntiretroviral failure was nearly 6 times higher among patients who had poor adherence to treatment as compared to patients with a good treatment adherence (OR = 5.90, 95% CI 3.50, 9.94, moderate strength of evidence). The likelihood of the treatment failure was almost 5 times higher among patients with CD4 < 200 cells/mm3 compared to those with CD4 ≥ 200 CD4 cells/mm3 (OR = 4.82, 95% CI 2.44, 9.52, low strength of evidence). This result shows that poor adherence and CD4 count below < 200 cells/mm3 are significantly associated with treatment failure among HIV-positive patients on ART in a resource-limited setting.\r\n\r\nConclusion\r\nThis review highlights that low CD4 counts and poor adherence to ART were associated to ART treatment failure. There is a need for healthcare workers and HIV program implementers to focus on patients who have these characteristics in order to prevent ART treatment failure.","PublicationOrPresentationDate":"2020-12-12T00:00:00Z","LastModifiedDate":"2021-07-13T18:49:06.055Z","IsDeleted":false,"DigitalObjectIdentifier":"10.1186/s13643-020-01524-1","ContentTypeName":"Research","ParentContentTypeName":"Journal Article","JournalName":"Systematic Reviews","ContentType":"Journal Article > Research","AvailabilityTypeId":1,"AvailabilityType":"Full Text","JournalIssue":null,"JournalVolume":"9","JournalPages":null,"JournalAbbreviation":"Syst Rev","Conference":null,"ConferenceStartDate":null,"ConferenceEndDate":null,"content":"\nRESEARCH Open Access\n\nFactors associated with antiretroviral\ntreatment failure among people living with\nHIV on antiretroviral therapy in resource-\npoor settings: a systematic review and\nmetaanalysis\nYishak Lailulo1†, Marcel Kitenge1,2†, Shahista Jaffer1, Omololu Aluko1 and Peter Suwirakwenda Nyasulu1,3*\n\nAbstract\n\nBackground: Despite the increase in the number of people accessing antiretroviral therapy (ART), there is limited\ndata regarding treatment failure and its related factors among HIV-positive individuals enrolled in HIV care in\nresource-poor settings. This review aimed to identify factors associated with antiretroviral treatment failure among\nindividuals living with HIV on ART in resource-poor settings.\n\nMethods: We conducted a comprehensive search on MEDLINE (PubMed), Excerpta Medica Database (EMBASE),\nCochrane Central Register of Controlled Trials (CENTRAL), World Health Organization’s (WHO’s) library database, and\nLatin American and Caribbean Health Sciences Literature (LILACS). We included observational studies (cohort, case-\ncontrol, and cross-sectional studies) where adolescents and adults living with HIV were on antiretroviral treatment\nregardless of the ART regimen. The primary outcomes of interest were immunological, virological, and clinical\nfailure. Some of the secondary outcomes were mm3 opportunistic infections, WHO clinical stage, and socio-\ndemographic factors. We screened titles, abstracts, and the full texts of relevant articles in duplicate. Disagreements\nwere resolved by consensus. We analyzed the data by doing a meta-analysis to pool the results for each outcome\nof interest.\n\n(Continued on next page)\n\n© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,\nwhich permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give\nappropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if\nchanges were made. The images or other third party material in this article are included in the article's Creative Commons\nlicence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons\nlicence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain\npermission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\nThe Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the\ndata made available in this article, unless otherwise stated in a credit line to the data.\n\n* Correspondence: pnyasulu@sun.ac.za\n†Yishak Lailulo and Marcel Kitenge contributed equally to this work.\n1Division of Epidemiology and Biostatistics, Department of Global Health,\nFaculty of Medicine and Health Sciences, Stellenbosch University, Cape\nTown, South Africa\n3Division of Epidemiology & Biostatistics, School of Public Health, Faculty of\nHealth Sciences, University of the Witwatersrand, Johannesburg, South Africa\nFull list of author information is available at the end of the article\n\nLailulo et al. Systematic Reviews (2020) 9:292 \nhttps://doi.org/10.1186/s13643-020-01524-1\n\nhttp://crossmark.crossref.org/dialog/?doi=10.1186/s13643-020-01524-1&domain=pdf\nhttp://orcid.org/0000-0003-2757-0663\nhttp://creativecommons.org/licenses/by/4.0/\nhttp://creativecommons.org/publicdomain/zero/1.0/\nmailto:pnyasulu@sun.ac.za\n\n\n(Continued from previous page)\n\nResults: Antiretroviral failure was nearly 6 times higher among patients who had poor adherence to treatment as\ncompared to patients with a good treatment adherence (OR = 5.90, 95% CI 3.50, 9.94, moderate strength of\nevidence). The likelihood of the treatment failure was almost 5 times higher among patients with CD4 < 200 cells/\nmm3 compared to those with CD4 ≥ 200 CD4 cells/mm3 (OR = 4.82, 95% CI 2.44, 9.52, low strength of evidence).\nThis result shows that poor adherence and CD4 count below < 200 cells/mm3 are significantly associated with\ntreatment failure among HIV-positive patients on ART in a resource-limited setting.\n\nConclusion: This review highlights that low CD4 counts and poor adherence to ART were associated to ART\ntreatment failure. There is a need for healthcare workers and HIV program implementers to focus on patients who\nhave these characteristics in order to prevent ART treatment failure.\n\nSystematic review registration: The systematic review protocol was registered with the International Prospective\nRegister of Systematic Reviews (PROSPERO), registration number: 2019 CRD42019136538.\n\nKeywords: HIV, ART, Immunological failure, Virological failure, Clinical failure, Poor outcome\n\nBackground\nHuman immunodeficiency virus (HIV) infections are a\nmajor global public health concern. In 2019, an estimated\n38 million people were living with HIV infection (PLWH)\n[1]. With new infections, an estimated 1.7 million people\nbecame newly infected with HIV in 2019. Sub-Saharan Af-\nrica (SSA) remains the most affected region in the world,\nwith about 20.7 million prevalent cases and 730,000 new\ninfections were recorded in 2019, seconded by Asia and\nthe Pacific region with 5.8 million prevalent cases [1]. Al-\nthough Southern Africa is home to less than 1% of the\nglobal population, the region has more than a fourth of all\nHIV infection in the world, with 300,000 acquired im-\nmune deficiency syndrome (AIDS)-related deaths regis-\ntered in the same year in SSA [1].\nAlthough anti-retroviral therapy (ART) coverage in\n\nthis region has rapidly increased over the past decade\n[2]. The greatest gains in access to ART occurred in SSA\n[3]. In 2019, only 15 million (73%) PLHIV in the region\nwere accessing ART, while 3.5 million (60%) in Asia and\nthe Pacific region [1]. Increasing the use of ART has\ncontributed to a prominent decline in HIV-associated\nmorbidity and death/mortality in SSA [2]. United Na-\ntions program on HIV/AIDS (UNAIDS) has suggested\nuniversal targets for the year 2020 (90-90-90), which\nmeans diagnosing 90% of all PLHIV who should know\ntheir status (PLHIV), initiating antiretroviral treatment\n(ART) for 90% of those diagnosed with HIV infection,\nand attaining an undetectable viral load in 90% of those\non ART [4]. Significant progress has been made in\nachieving that goal. Globally, PLWH accessing ART has\nincreased from 21.7 million in 2015 to 25.4 million in\n2019, an increase from 45 to 67% of all PLHIV [3, 5].\n\nAntiretroviral treatment failure\nPatients with ART failure are increasingly encountered in\nresource-limited settings, while recent estimates suggest\nonly 2% of those currently on ART are on second-line [6],\n\na far greater number is likely to be failing virologically but\nhave not switched to an alternative regimen. Furthermore,\nan increase in the coverage of ART use among PLHIV,\nwhich has resulted in an increase in the number of indi-\nviduals failing first-line ART, and therefore, the magnitude\nincreases with prolonged use of ART. The WHO pre-\ndicted earlier on that 500,000 and 800,000 PLWH on the\nfirst-line combination of ART will require a switch to the\nsecond-line therapy by 2010 [2]. However, the burden of\ntreatment failure is not well-documented, while there is a\nlarge scale of ARV in resource-limited countries. Meta-\nanalysis data showed that the rate of the treatment failure\nfor the first-line was 6.08% globally; however, the study\nnoted a substantial heterogeneity across regions with\n7.10% in Africa and 2.55% in Asia [3].\nA retrospective cohort study done in South Africa\n\nfound that among patients on non-nucleotide reverse\ntranscriptase inhibitor (NNRTI)-based ART, after a me-\ndian of 15 months on ART treatment, 19% had failed\nvirologically and immunologically [6]. Studies in East Af-\nrica have shown a high prevalence of immunologic fail-\nure ranging from 8 to 57% among clients on the first-\nline ART [7–9].\nTreatment failure is typically measured in three ways in\n\npoor-resource settings: (i) clinically, as evidenced by disease\nprogression; (ii) immunologically, as evidenced by trends in\nCD4 counts over time; and (iii) virologically, as evidenced by\nmeasurement of HIV RNA levels. In 2013, WHO recom-\nmended viral load testing as the preferred monitoring ap-\nproach to diagnose and to confirm ARV treatment failure [10].\n\nFactors associated with treatment failure\nEarlier studies have emphasized a number of factors that\nmay be associated with virological suppression in ART;\nthese are reasons for testing: routine testing, suspected\ntreatment failure, and repeat testers after suspected failure\n[9–11]. While a significant number of studies have found\nthat treatment failure is significantly associated with young\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 2 of 17\n\n\n\nage, unsatisfactory adherence, low hemoglobin, history of\nlost to follow-up, being male and educational status, and\ntreatment regimen [12–14], some studies have recognized\nlow baseline CD4 cell count, rate of CD4 decline, prior ex-\nposure to ART and treatment interruptions, and non-\nadherence as determinants of treatment failure [15, 16].\nIn 2016, WHO most recent guideline defined a clinical\n\nfailure as a new or recurrent clinical event indicating se-\nvere immunodeficiency (WHO clinical stage 4 condition)\nafter 6 months of effective treatment. Immunological fail-\nure is defined as CD4 count at or below 250 cells/mm3\n\nfollowing a clinical failure or persistent CD4 levels below\n100 cells/mm3, and virological failure is defined as viral\nload above 1000 copies/mL based on two consecutive viral\nload measurements in 3 months, with adherence support\nfollowing the first viral load test [17]. The results from a\nprevious study have confirmed that low baseline CD4 cell\ncount, particularly < 100 cells/mm3, and history of loss to\nfollow-up are risk factors for immunological discordance\n[18]. Independent risk factors associated with virological\nfailure were being followed-up at the semirural center,\nhaving experienced unstructured treatment interruptions,\nand having low CD4 counts at enrolment [19].\nGender, time on ART, baseline CD4 T cell count, WHO\n\nstage, ART regimen, adherence, and TB co-infection were\nassociated with viral suppression [20]. The history of the\nantiretroviral use before starting ART, change of antiretro-\nviral therapy due to toxicity, opportunistic infections while\non ART treatment, level of CD4 + lymphocytes below 100\ncells/ml at start of ART, adherence, and clinical stage were\nindependently associated with virological failure [21]. Age\nyounger than 40 years was also associated with virologic\nfailure [22]. The relative contribution of the main predic-\ntors to virological failure may differ across settings and\npopulation groups and context. Thus, specific data are\ncritical to the carrying out of corrective measures.\n\nImportance of the review\nViral load testing provides early and accurate indications\nof the treatment failure and the need to switch from the\nfirst-line to second-line drugs, thereby reducing the ac-\ncumulation of the drug-resistant mutations and improv-\ning clinical outcomes [23].\nHowever, regular access to routine viral load testing re-\n\nmains a challenge due to the high cost. In such a situation,\nclinical and immunological monitoring is used for detecting\ntreatment failure [24–27]. The number of people accessing\nART has significantly increased in many poor resource set-\ntings [28]. Hence, it is significant to sustain treatment suc-\ncess and limit the development of treatment failure. For the\ntimely detection of treatment failure, WHO reconfirmed\nthe use of viral load testing as the gold standard test to\nmonitor patients’ response to ART [29]. Where the viral\nload is not routinely available, CD4 count and clinical\n\nmonitoring should be used to diagnose treatment failure. In\nspite of a large number of patients receiving ARTs in low-\nand middle-income countries (LMICs) and poor settings,\nthere are few reports on ART outcomes in these settings.\nIdentifying baseline predictors of the first-line ART out-\ncome among PLWH on ART in LMICs where access to\nviral load testing is limited is of paramount importance.\nThe technique and accuracy of identifying treatment\n\nfailure in poor settings are important but challenging.\nDelayed detection of ART failure may increase drug tox-\nicity may lead to the increase of drug resistance related\nwith mutations (further controlling treatment choices)\nand may result in increased morbidity and mortality.\nEarly detection of treatment failure is crucial to ensure\nthe effectiveness of the first-line therapy [6].\nThe main objective of this review was to identify fac-\n\ntors associated with antiretroviral treatment failure\namong PLWH on ART in resource-poor settings.\n\nObjective\nPrimary objective\nThe primary objective of the study was to determine the\nclinical, immunological, and virological factors associ-\nated with antiretroviral treatment failure among PLWH\nin resource-poor settings.\n\nSecondary objective\nThe secondary objective of the study is to identify the\nsocio-demographic and economic factors associated with\nantiretroviral treatment failure among PLWH among\nPLWH in resource-poor settings.\n\nMethods\nThe methods of this systematic review and meta-analysis\nwere reported as per the Preferred Reporting Items for\nSystematic Review and Meta-Analysis Protocols (PRIS\nMA-P) checklist [30]. We registered the protocol for this\nsystematic review on the International Prospective Regis-\nter of Systematic Reviews (PROSPERO) with a registra-\ntion number: CRD42019136538.\n\nCriteria for considering studies for review\nTypes of studies\nWe included all types of observational studies including\nprospective/retrospective or ambi-directional cohort studies,\ncase-control studies, population-based/nested or hospital-\nbased case-control studies, and cross-sectional studies.\nInterventional studies were excluded from this review.\n\nTypes of participants\nAdolescents and adults living with HIV who were on\nART for ≥ 6 months, regardless of the regimen. Only\nparticipants with documented baseline CD4 and VL\nwere considered for this systematic review.\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 3 of 17\n\n\n\nType of outcome\nPrimary outcome\nTreatment failure was defined as follows:\n\nVirological failure\nVirological failure is defined as a plasma viral load above\n1000 copies/ml based on two consecutive viral load mea-\nsurements after 3 months, with adherence support. A\nviral load test is a measurement of the amount of HIV\nin a sample of the blood. This is usually reported as the\nnumber of copies per milliliter (copies/mm3) [17].\n\nImmunological failure\nImmunological failure is defined as a fall in CD4 count\nto the baseline (or below) or persistent CD4 levels below\n100 cells/mm3. The CD4 lymphocyte count is an excel-\nlent indicator of how healthy the immune system is.\nThese are a type of white blood cells, called T cells,\nwhich move throughout the human body to find and\ndestroy bacteria, viruses, and other invading germs. The\nCD4 cell count is indicated in cells per mm3, and it is\nmeasured by taking a blood sample [17].\n\nClinical failure\nClinical failure is defined as the occurrence of new op-\nportunistic infections (excluding immune reconstitution\ninflammatory syndrome [IRIS]) and/or other clinical evi-\ndence of HIV disease progression during therapy. AIDS-\ndefining illnesses (opportunistic infections) are those\nwhich the Centers for Disease Control and Prevention\n(CDC) have classified as being directly associated with\nadvanced HIV infection. We considered the common\ndiseases, which are pneumonia, TB, lymphoma, and\ncryptococcosis [17].\n\nSecondary outcome\nSecondary outcomes for this study are all the predictors’\nvariables that contribute to treatment failure. The fol-\nlowing information was collected if measured at base-\nline: CD4 cells (cells/mm3), viral load (copies/ml), WHO\nclinical, tuberculosis, opportunistic infection, treatment\nregimen (NRTI or NNRTI), BMI, weight, study site\n(rural versus urban), gender, age, educational status, em-\nployment status, marital status, and spouse HIV sero-\nstatus.\n\nInclusion and exclusion criteria\nIncluded studies\nParticipants in the study were (1) those who had been\non ART for ≥ 6 months and (2) those who had docu-\nmented CD4 cell count and viral load measurement at\nbaseline and 6 months.\n\nExcluded studies\nAll studies with participants who had pregnancy history\nthe past 6 months while on treatment and at 6 months’\nvisit or had missing values of CD4 cell count and viral\nload at baseline and 6 months’ visit were excluded.\n\nSearch methods for identification of the studies\nWe conducted a comprehensive search on 5 databases\nfrom December 1, 2000, to November 2019. With assist-\nance from an information specialist, we searched in the\nfollowing databases: MEDLINE (Pubmed), EMBASE\n(OVID), LILACS (BIREME), Science Citation Index Ex-\npanded (SCI-EXPANDED, Web of Science), Social Sci-\nences citation index (SSCI, Web of Science), Conference\nProceedings Citation Index-Social Science & Humanities\n(CPCI-SSH, Web of Science), and Cinahl (EBSCOHost).\nA detailed search strategy is provided in Appendix 1. A\nhand search of citations from selected studies was con-\nducted to identify additional studies missing from the\noriginal electronic searches.\n\nScreening and assessments of study eligibility\nAll potential studies were imported into Covidence\n(Covidence systematic review software, Veritas Health\nInnovation, Melbourne, Australia), and two review au-\nthors (YL and SJ) independently screened the titles and\nabstracts. Both authors also assessed full-text eligibility.\nAll published full-text articles, abstracts, and brief re-\n\nports were included, and provided/available complete\ndata were elicited from them. The disagreements be-\ntween the two authors who assessed study eligibility\nwere resolved by discussion and consensus.\n\nData extraction, management, and analysis\nData from the full-text articles were extracted by two in-\ndependent review authors (YL, SJ) using a standardized\npre-piloted data extraction form. A third reviewer (MK,\nPN) checked whether the extracted data were correct.\nExtracted data were categorized into four main headings:\ngeneral information, socio-demographic and economic\ncharacteristics of participants, and clinical and immuno-\nlogical information of the participant. In case of missing\ninformation, we clarified the conducted study or the\nstudies that had relevant data, which were not reported\nin the published manuscript, and we contacted the au-\nthors for additional information.\n\nRisk of bias and quality of evidence\nTwo authors independently assessed the risk of bias in\neach study by examining the study population, study at-\ntrition, prognostic factor measurement, outcomes meas-\nurement, study confounding, and statistical reporting\n(YL and OA). They coded studies as at high, medium,\nlow, or unclear risk of bias for each of these features\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 4 of 17\n\n\n\nusing the Quality in Prognosis Studies tool (QUIPS tool)\n[31]. Finally, we assessed the quality of the evidence\nusing the Grading of Recommendations Assessment De-\nvelopment and Evaluations (GRADE) approach using\nthe five criteria of the GRADE system.\n\nStatistical analysis\nFor the studies that were relatively homogeneous in terms\nof methodology and outcomes, a meta-analysis of the data\nwas performed. Sufficiently, similar data was pooled using\nthe inverse variance approach to accommodate crude and\nadjusted odds ratios, where possible. Additionally, the\nmeta-analysis was summarized using pooled estimates, the\n95% confidence interval, and the between-study variance\nwas estimated using Tau2. We extracted all unadjusted\nand adjusted measures of the association from all included\nstudies and converted effect sizes as necessary to possible\nselection bias, thus allowing us to use the data from as\nmany studies as possible. We anticipated that results from\nmultivariate analyses would have been reported as odds\nratios (ORs), risk ratios (RRs), and hazard ratios (HRs), if\nso, we would use ORs as the common measure of the as-\nsociation, using RRs and HRs to estimates ORs at a\n\nparticular time point [32]. Furthermore, measures of effect\nwere analyzed using RevMan statistical software for sys-\ntematic reviews. Statistical heterogeneity was quantified\nusing the I2 statistic [33]. If the I2 statistic is high (75 to\n100%—as suggested by Higgins et al.) indicating high het-\nerogeneity [33], a random effect model was used.\n\nResults\nPRISMA flow chart\nWe retrieved 2418 articles regarding treatment failure\namong ART users in poor resource setting as identified\nin MEDLINE (PubMed); EMBASE (OVID); LILACS (BI-\nREME); Science Citation Index Expanded (SCI-EX-\nPANDED, Web of Science), Social Sciences citation\nindex (SSCI, Web of Science), and Conference Proceed-\nings Citation Index-Social Science & Humanities (CPCI-\nSSH, Web of Science), and CINAHL (EBSCOHost).\nThese are shown in Fig. 1.\nOf these initial articles, 3 articles were duplicates; 2158\n\narticles were excluded after reviewing their titles and ab-\nstracts and confirmed irrelevant to this review. Thus,\n237 potential full-text articles were assessed for eligibil-\nity, which resulted in further exclusion of 100 articles.\n\nFig. 1 PRISMA flowchart of included studies\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 5 of 17\n\n\n\n57 had wrong outcomes, 19 assessed HIV drug-resistant\nmutations, 12 had the wrong study design, 7 had a\nwrong patient population, 2 were not in English 1 and\nwas a duplicate, 1 had a wrong setting, and 1 was\npediatric population. Finally, 137 studies met the eligibil-\nity criteria. These are shown in Table 1.\n\nMeta-analysis\nThe association between adherence and treatment fail-\nure was based on six cross-sectional studies [14, 35, 37,\n40, 42, 47]. The results as presented in Fig. 2 showed a\n\nstrong relationship between treatment failure and poor\ntreatment adherence. The odds of treatment failure were\nnearly 6 times higher among patients who had poor ad-\nherence (OR = 5.90, 95% CI 3.50, 9.94, moderate strength\nof evidence). The test statistics, however, showed a sub-\nstantial heterogeneity (I2 = 65% and p = 0.02).\nSimilarly, the association between poor adherence and\n\ntreatment failure was examined using four cohort studies\n[36, 39, 41, 46]. The results as presented in Fig. 3 showed\nthat the hazard ratio of treatment failure was nearly 2.5\nhigher among patients who had poor adherence (HR =\n\nTable 1 Characteristics of included studies\n\nReferences Year of\npublication\n\nStudy design Country Patients groups ART used Sample\nsize\n\nNumber of\nTreatment failure\n\nBabo et al. [34] 2017 Case-control study Ethiopia Adult Stavudine vs. Zidovudine\nNevirapine vs. Efavirenz\n\n307 230\n\nBayu et al. [35] 2017 Case-control study Ethiopia Adults aged ≥\n15 years\n\nD4T-based\nAZT-based\nTDF-based\n\n306 160\n\nBilcha et al. [36] 2019 Retrospective\ncohort study\n\nEthiopia Adult Nevirapine-based\nEfavirenz-based\n\n396 47\n\nBisson et al. [37] 2008 Case-control study Botswana Adults older than\n18 years\n\nNR 302 247\n\nFatti et al. [38] 2019 Prospective cohort\nstudy\n\nSouth\nAfrica\n\nAdults aged ≥\n18 years\n\nNRTI and NNRTI 1901 60\n\nFord et al. [39] 2010 Observational\ncohort\n\nSouth\nAfrica\n\nAdult EFV, NVP, and other 207 32\n\nGunda et al. [40] 2019 Case-control study Tanzania Adult AZT/3TC/EFV, AZT/3TC/NVP,\nD4T/3TC/NVP, TDF/3TC/EFV\n\n197 24\n\nHaile et al. [41] 2016 Retrospective\ncohort study\n\nEthiopia Adult (≥ 15 years\nold)\n\n1a(d4T + 3TC + NVP), 1b(d4T\n+ 3TC + EFV),\n1c(AZT + 3TC + NVP),\n1d(AZT + 3TC + EFV),\n1e(TDF + 3TC + EFV), 1f(TDF\n+ 3TC + NVP)\n\n4809 113\n\nHailu et al. [42] 2018 Retrospective\nfollow-up study\n\nEthiopia Adults (≥ 20\nyears)\n\nTDF 3TCEFV/NVP, AZT 3TC\nNVP/EFV,\nD4T 3TC NVP/EFV, ABC 3TC\nEFV\n\n260 30\n\nHassan et al. [14] 2014 Cross-sectional\nstudy\n\nKenya Adult Zidovudine-based and\nStavudine-based\n\n232 57\n\nIzudi et al. [43] 2016 Retrospective\ncohort\n\nUganda Adult 383 28\n\nKarade et al. [44] 2016 Cross-sectional\nstudies\n\nIndia Adult AZT + 3TC + NVP, AZT +\n3TC + EFV\nTDF + 3TC + NVP, TDF +\n3TC + EFV\nd4T + 3TC + NVP/EFV\n\n844 104\n\nLay et al. [45] 2017 Retrospective\ncohort study\n\nCambodia Adult (≥ 18 years\nold)\n\nd4T/3TC/EFV, d4T/3TC/NVP\nAZT/3TC/EFV, AZT/3TC/NVP\nOther\n\n3581 137\n\nNdahimana et al.\n[46]\n\n2016 Retrospective\ncohort\n\nRwanda 15 years and\nolder\n\nNRTIs, NNRTIs, and PIs 828 70\n\nAhmed et al. [47] 2019 Case-control study Ethiopia Adult d4t + 3TC + NVP, AZT + 3TC\n+ NVP\nAZT + 3TC + EFV, TDF + 3TC\n+ EFV\nTDF + 3TC + NVP\n\n308 199\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 6 of 17\n\n\n\n2.46, 95% CI 1.72, 3.51, high strength of evidence). The re-\nsult of test statistics showed no heterogeneity (I2 = 0% and\np = 0.90). Here too, a random effect meta-analysis model\nwas used to determine the association with the outcome.\nFurthermore, the association between CD4 and treat-\n\nment failure was examined by using three cross-\nsectional studies [35, 40, 47]. The results as presented in\nFig. 4 showed that treatment failure was strongly associ-\nated with CD4 count. The odds of treatment failure were\nnearly 5 times higher among patients who had a CD4\ncell count of 200 cells/mm3 (OR = 4.82, 95% CI 2.44,\n9.52, low strength of evidence). However, the test statis-\ntics showed substantial heterogeneity (I2 = 71% and p =\n0.03). Hence, a random effect meta-analysis model was\nused to determine the association with the outcome.\nLikewise, the association between low CD4 count and\n\ntreatment failure was also observed using four cohort stud-\nies [36, 38, 45, 46]. Results presented in Fig. 5 showed that\nthe hazard ratio of treatment failure was nearly 3\ntimes higher among patients who had CD4 lower\nthan 200 cells/mm3 (HR = 2.98, 95% CI 2.23, 4.00,\nmoderate strength of evidence). The result of the test\nstatistics showed no evidence of heterogeneity (I2 =\n0% and p = 0.55). A random effect meta-analysis\nmodel was used to determine the association with the\noutcome.\nOur study also demonstrated similar findings to the\n\nabove through data abstracted from two cross-sectional\n\nstudies [34, 44]. We also found that treatment failure\nwas significantly associated with low CD4 count, where\nthe odds of treatment failure were 1.14 times higher\namong patients with CD4 lower than 100 cells/mm3\n\n(OR = 1.14, 95% CI 0.52, 2.47, low strength of evidence).\nThe test statistics showed moderate heterogeneity (I2 =\n49% and p = 0.75), see Fig. 6. Consequently, a random\neffect meta-analysis model was computed to determine\nthe association.\n\nRisk of bias assessment\nMost of the studies had a low risk of bias on prognostic\nfactors that accounted for 125/137, followed by study\nparticipants (123/135), statistical analysis and reporting\n(116/137), and outcome measurement (115/137). More-\nover, 109/137 studies had a low risk of bias on study\nconfounding and 103/137 studies had a low risk of bias\non study participant attrition. The full table of results is\nshown in Appendix 3: risk of bias assessment.\n\nDiscussion\nThis review was aimed at identifying factors associated\nwith antiretroviral treatment failure among individuals\nliving with HIV and showed that low CD4 T cell count\n(≤ 200 cells/mm3) and poor adherence to ART were sig-\nnificantly associated with virological failure.\nIn this review, the odds of virological failure were\n\nhigher among those who had a CD4 cell count of ≤\n\nFig. 2 Pooled odds ratio between adherence and treatment failure. Comparison: poor versus good adherence (outcome: virological failure)\n\nFig. 3 Pooled odds ratio between adherence and treatment failure. Comparison: poor versus good adherence (outcome: virological failure)\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 7 of 17\n\n\n\n200cells/mm3 in both case-control and cohort studies.\nThe finding is supported by the studies conducted in\nSSA [35, 43], while a retrospective analysis of a large\nART program in Cambodia showed that previous ART\nexperience, nevirapine-based regimen, and CD4 count ≤\n200 cells/mm3 were independently associated with an in-\ncreased risk of treatment [48]. Similar findings were re-\nported in a meta-analysis data from India, where CD4\ncount ≤ 200 had a significantly greater risk of treatment\nfailure [49]. As CD4 cell count increases, viral replication\ndecreases, which means it has an inverse relationship\nwith viral load. As patients’ immune status drops, and\nthe rate of viral load increases compared to the\nimmuno-competent individuals with HIV infection. In\naddition, users with compromised immunity are more\nsusceptible to different opportunistic infections that en-\ndure the cruel cycle of immunity depletion and viral rep-\nlication [50].\nMoreover, the results found from case-control studies\n\nshown that the odds of virological failure were 6 times\nmore among those who had poor adherence compared\nwith those who had good adherence to antiretroviral\ntreatment. Likewise, the finding from cohort studies\nshowed that the odds of virological failure were higher\namong those who had poor adherence compared with\nthose who had good adherence to antiretroviral treat-\nment. This finding is supported by findings from pri-\nmary studies conducted in African countries [11, 51, 52],\nbut also consistent with the finding from a study con-\nducted in Vietnam and other developed countries [53–\n\n55]. It is obvious that poor adherence to medication\ncompromises treatment response due to suboptimal\ndrug concentration hence creates a conducive environ-\nment for viral replication leading to virological failure\n[56, 57]. This reaffirms the need for reinforcement of\ndrug adherence counseling for HIV patients before and\nduring their life course of taking ART.\nPoor adherence may lead to a number of adverse con-\n\nsequences on both individual and public HIV healthcare\nlevels. Therefore, the measured efforts are immediately\nneeded in HIV care by responsive bodies like ART case\nmanagers, adherence counselors in the hospitals on pa-\ntients with low current CD4 count through improving\npoor adherence to ART treatment by strengthening en-\nhanced adherence counseling. Each low-income country\nnational HIV program should give attention to improv-\ning HIV services to strengthen adherence among pa-\ntients on ART in order to reduce the proportion of\npatients who are failing the treatment.\nOur systematic review has some strengths. We\n\nplanned the review a priori with clearly defined selection\ncriteria. We conducted a comprehensive and exhaustive\nsearch, using many additional sources to identify rele-\nvant studies, including reference searches of other HIV/\nAIDS conferences (IAS and CROI) for the past 20 years.\nThis review had several limitations mainly related to\n\nthe quality of the evidence available. To our knowledge,\nwe suspect publication or reporting biases, or both, sug-\ngesting that our results may be overestimated. Positive\nstudy bias is likely to be problematic in this review. Our\n\nFig. 4 Pooled odds ratio between CD4 and treatment failure. Comparison: CD4 < 200 cells/mm3 versus CD4 ≥ 200 cells/mm3 (outcome:\nvirological failure)\n\nFig. 5 Pooled odds ratio between CD4 and treatment failure. Comparison: CD4 < 200 versus CD4 ≥ 200 (outcome: virological failure)\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 8 of 17\n\n\n\nliterature search for relevant and potential studies in-\ncluded focused searches, i.e., including search terms re-\nlated to the “less CD4 count,” “viral load” in our\nelectronic search. Studies that report a relationship be-\ntween the prognostic factors and common outcomes are\ntherefore more likely to have been identified in these\nsearches due to reporting of positive results in the study\nabstract.\nIn addition, we also observed that some studies re-\n\nported positive unadjusted association of factors with\noutcomes of interest, but did not report the associ-\nation adjusted for other important covariates. This\nmay contribute to a likely overestimation of the ad-\njusted results. Therefore, future research is required\nto investigate the impact and potential strategies to\nalleviate reporting and publication bias, as well as ini-\ntiatives to require registration of protocols and publi-\ncation of prognostic studies.\nFurthermore, our review was the pooling of the ad-\n\njusted the results despite studies did not include identi-\ncal sets of covariates. Studies included in this review\nwere homogenous; therefore, pooling of the adjusted re-\nsults was feasible. However, comparison and interpret-\nation may be challenging in this case. Our review only\nfocused on studies conducted in poor resource settings\nlimiting its generalizability to high-income settings.\n\nStrength of evidence\nThe strength of evidence contributing to several outcomes\nin this review was graded as low, moderate, or high. We\nused the GRADE approach to assess the strength of evi-\ndence as shown in the summary of the finding table, Ap-\npendix 4. The certainty of evidence was downgraded in\nmost instances due to a high risk of bias as well as\ninconsistency.\n\nConclusion\nART failure among individuals living with HIV is a pub-\nlic health concern; the timing and accuracy of identifying\ntreatment failure in resource-limited settings are funda-\nmental but challenging. The findings of this review\nhighlighted that low CD4 counts and poor adherence to\nART were associated to ART treatment failure. There is\n\nan urgent need that health professionals and HIV pro-\ngrams should focus on novel approaches for patients\nwho have these characteristics in order to prevent ART\nfailure. Further review is required to be done in multiple\nART centers and a broader community as well as the\ndifferent factors associated with treatment failure to de-\ncide whether there are discrepancies in virological and\nimmunological responses to antiretroviral therapy at dif-\nferent stages of HIV infection.\n\nAppendix 1\nSearch strategy—database\n#1 Search ((HIV OR hiv-1 OR hiv-2* OR hiv1 OR hiv2\nOR hiv infect* OR human immunodeficiency virus OR\nhuman immune deficiency virus OR human immuno-\ndeficiency virus OR human immune-deficiency virus OR\n((human immun*) AND (deficiency virus)) OR acquired\nimmune deficiency syndromes\nOR acquired immune deficiency syndrome OR ac-\n\nquired immuno-deficiency syndrome OR\nacquired immune-deficiency syndrome OR ((acquired\n\nimmun*) AND (deficiency syndrome)) OR HIV/AIDS))\n#2 Search ((HIV infections [MeSH] OR HIV [MeSH]))\n#3 Search (#1 OR #2)\n#4 Search ((Antiretroviral* OR ((anti) AND (retro-\n\nviral*)) OR ARV* OR ART OR “antiretroviral therapy”\nOR HAART OR ((highly) AND (active) AND (anti-\n\nretroviral*) AND (therap*)) OR ((anti) AND (hiv)) OR\n((anti) AND (acquired immunodeficiency)) OR ((anti)\n\nAND (acquired immuno-deficiency)) OR ((anti)\nAND (acquired immune-deficiency)) OR ((anti) AND\n\n(acquired immun*) AND (deficienc*))))\n#5 Search ((antiretroviral agents [Mesh] OR antiretro-\n\nviral therapy, highly active [Mesh]))\n#6 Search (#4 OR #5)\n#7 search #3 AND #6\n#8 Search (virological failure OR Immunological fail-\n\nure OR less CD4 count OR viral load)\n#9 Search (low-income setting OR disadvantaged com-\n\nmunities OR resource limited setting OR Sub-Saharan\nAfrica)\n#10 Search (#7 AND #8 AND #9)\n\nFig. 6 Pooled odds ratio between CD4 and treatment failure. Comparison: CD4 < 100 versus CD4 ≥ 100. (outcome: virological failure)\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 9 of 17\n\n\n\nAppendix 3\nRisk of bias assessment\n\nTable 2 Risk of bias assessment\n\n# Study ID Study Prognostic Statistical\nanalysis\nand\nreporting\n\nparticipant Attrition Factor measurement Outcome measurement Study confounding\n\n1 Abah 2018 Low High Low Low Low Low\n\n2 Ahmed 2019 Low Low Low Low Low Low\n\n3 Ahn 2019 Low High Low Low Low Low\n\n4 Ahoua 2009 Low Low Low Low Low Low\n\n5 Assefa 2014 Low High Low Low Low Low\n\n6 Ayalew 2016 Low Low Low Low Low Low\n\n7 Ayele 2018 Low Low Low Low Low Low\n\n8 Babo 2017 Low Low Low Low Low Low\n\n9 Bayou 2015 Low High Low Low Low Low\n\n10 Bayu 2017 Low Low Low Low Low Low\n\n11 Billioux 2015 Low Low High High Low Low\n\n12 Biscione 2014 Low Low High High High Low\n\n13 Bisson 2008 Low Low High Low High Low\n\n14 Boender 2016a Low Low Low Low Low Low\n\n15 Boender 2016b Low Low Low Low Low Low\n\n16 Boettiger 2016c Low Low Low Low Low Low\n\n17 Boettiger 2015 Low Low Low Low Low Low\n\n18 Boettiger 2016d Low Low Low Low Low Low\n\n19 Boettiger 2014 Low Low Low Low Low Low\n\n20 Boulle 2015 Low Low Low Low Low Low\n\n21 Braun 2017 Low Low Low Low High High\n\n22 Brooks 2016 Low Low Low Low High High\n\n23 Bulage 2017 Low High Low Low Low Low\n\n24 Byabene 2017 Low Low Low Low Low Low\n\n25 Cao 2018 Low Low Low Low Low Low\n\n26 Carriquiry 201 Low Low Low Low Low Low\n\n27 Caseiro 2018 Low Low Low High High Low\n\n28 Castelnuovo 2016 Low Low Low Low Low Low\n\n29 Cesar 2015 Low Low Low Low Low Low\n\n30 Cesar 2014 Low Low Low Low Low Low\n\n31 Chaiwarith 2011 Low Low Low Low Low Low\n\n32 Chaiwarith 2007 Low Low High Low Unclear Low\n\n33 Chakravarty 2015 Low Low Low Low Unclear Low\n\n34 Charles 2013 Low Unclear Low Low Unclear Low\n\n35 Chawana 2014 Low Low Low Low Unclear Low\n\n36 Chen 2014 Low High Low Low High Low\n\n37 Chhim 2018 Low High Low Low Low Low\n\n38 Chkhartishvili 2014 Low Low Low Low Unclear Low\n\n39 Collier 2017 Low Low Low Low Low Low\n\n40 Costiniuk 2014 High Unclear Low Low High High\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 10 of 17\n\n\n\nTable 2 Risk of bias assessment (Continued)\n\n# Study ID Study Prognostic Statistical\nanalysis\nand\nreporting\n\nparticipant Attrition Factor measurement Outcome measurement Study confounding\n\n41 Court 2014 Low Low Low Low Low Low\n\n42 Datay 2010 Low Low Low Low Unclear Low\n\n43 DeBoni 2018 Low Low Low Low Low Low\n\n44 deLaHoz 2014 Low Low Unclear Unclear Unclear Low\n\n45 Dolling 2017 Low Low Low Low Low Low\n\n46 Dray-Spira 2007 Low Low Low Low Unclear Low\n\n47 Ekstrand 2011 Low Low High Unclear Unclear High\n\n48 Rusine 2013 Low Low Low Low Low Low\n\n49 Sadashiv 2017 Low Low Low Low Low Low\n\n50 Safren 2014 Low Low Low Low Low Low\n\n51 Saracino 2014 Low Low Low Low Low Low\n\n52 Singini 2016 Low Low Low Low Low Low\n\n53 Sithole 2018 Low Low Low Low Low Low\n\n54 Sovershaeva 2019 Low Low Low Low Low Low\n\n55 Syed 2016 Low Low Low Low Low Low\n\n56 Telele 2018 Low Low Low Low Low Low\n\n57 Teshome 2014 Low Low Low Low Low Low\n\n58 Thiha 2016 High Low Low Low Low Low\n\n59 Tran 2014 Low High Low Low Low Low\n\n60 Tsegaye 2016 Low High Low Low Low Low\n\n61 vandenBerg 2005 Low Low Low Low Low Low\n\n62 Vanobberghen 2015 Low Low Low Low Low Low\n\n63 Wang 2011 High High Low Low Low Low\n\n64 Yimer 2015 Low Low Low Low Low Low\n\n65 Yirdaw 2015 Low Low Low Low Low Low\n\n66 Zhao 2017 Low High Low Low Low Low\n\n67 Zoufaly 2015 Low Low Low Low Low Low\n\n68 Elema 2009 Low Low Unclear low Unclear Low\n\n69 Enderis 2009 Low Low Low Low Low Low\n\n70 Eshleman 2017 Low Low Low Low Low Low\n\n71 Evans 2018 Low High Low Low Low Low\n\n72 Evans 2013 Low High Low Low Low Low\n\n73 Fatti 2019 Low Low Low Low Low Low\n\n74 Fatti 2014 Low Low Low Low Low Low\n\n75 Ferradini 2007 Low Low Low Low Low Low\n\n76 Ferreyra 2012 Low Low Low Low Low Low\n\n77 Fibriani 2013 Low Low Low Unclear Unclear Low\n\n78 Flynn 2017 Low Low Low Low Low Low\n\n79 Fogel 2017 unclear Unclear Low Low Low Unclear\n\n80 Ford 2010 Low Low Low Low Low Low\n\n81 Fox 2012 Low Low Low Low Low Low\n\n82 Fox 2010 Low Low Low Low Low Low\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 11 of 17\n\n\n\nTable 2 Risk of bias assessment (Continued)\n\n# Study ID Study Prognostic Statistical\nanalysis\nand\nreporting\n\nparticipant Attrition Factor measurement Outcome measurement Study confounding\n\n83 Goldman 2008 Low Low Low Low Unclear Low\n\n84 Gross 2017 Low Low Low Low Low Low\n\n85 Gunda 2019 Low Low Low Low Low Low\n\n86 Haggblom 2016 Low Low Low Low Low Low\n\n87 Haile 2016 Low High Low Low High Low\n\n88 Hailu 2018 Low Low Low Low Low Low\n\n89 Hamers 2012 Low Low Low Low Low Low\n\n90 Hare 2014 Low Low Low Low Low Low\n\n91 Hassan 2014 Low Low Low Low Low Low\n\n92 Hawkins 2015 Low High Low Low Low Low\n\n93 Hawkins 2016 Low Low Low Low Low Low\n\n94 Hermans 2018 Low High Unclear Low Unclear Low\n\n95 Huang 2015 Low High Low Low Low Low\n\n96 Hunt 2017 Low Low Low High Unclear Low\n\n97 Huong 2011 Low Low Low low Unclear Low\n\n98 Inzaule 2018 Low Low Unclear low Low Low\n\n99 Izudi 2016 Low Low Unclear low Low Low\n\n100 Jiamsakul 2016 Low High Low low Low Low\n\n101 John 2016 Low Low Low low Low Low\n\n102 Joram 2017 Low High Low High Low Low\n\n103 JosephDavey 2018 Low Low Low low Low Low\n\n104 Kamya 2007 Low Low Unclear low Low Low\n\n105 Kan 2017 Low Low Low High Low Low\n\n106 Karade 2016 Low Low Low High Low Low\n\n107 Kazooba 2018 Low High Low Low Low Low\n\n108 Khienprasit 2011 Low Low Low High Low High\n\n109 Kyaw 2017 Low High Low High Low Low\n\n110 Lay 2017 Low Low Low High Low High\n\n111 Leng 2014 Low Low High High Low Low\n\n112 Lenjisa 2015 Low High Low Low Low High\n\n113 Levison 2011 Low Low Low Low High High\n\n114 Liegeois 2013 Low Low Low High Low High\n\n115 Masikini 2019 High Low Low High Low Low\n\n116 Meloni 2016 Low Low Low High High Low\n\n117 Mpawa 2017 High High Low High Low High\n\n118 Mujugira 2016 Low Low Low low Low Low\n\n119 Mungwira 2018 Low High Low High Low Low\n\n120 Musa 2015 Low Low Low Low Low Low\n\n121 Nachega 2008 High Low Low Low High Low\n\n122 Ndahimana 2016 High Low Low Low Low High\n\n123 Negi 2018 Low High Low Low Low HIgh\n\n124 Nsanzimana 2019 High High Low Low Low High\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 12 of 17\n\n\n\nAppendix 2\nRisk of bias criteria and justifications\n\nAssessment for risk of bias\n\nFirst author Reviewer........................\n\nBiases Issues to\nconsider\nfor judging\noverall\nrating of\n“risk of\nbias”\n\nStudy\nmethods\nand\ncomments\n\nRating of risk of bias\n\nAssess the risk\nof each\npotential bias\n\nThese issues\nwill guide\nyour\nthinking and\njudgment\nabout the\noverall risk\nof bias\nwithin each\nof the 6\ndomains.\n\nProvide\ncomments\nor excerpts\nto facilitate\nthe\nconsensus\nprocess that\nwill follow\n\nHigh, moderate, low\n\n1) Study\nparticipation\n\nThe study sample\nadequately represents the\npopulation of interest\n\nSummary\n\na. Adequate participation in\nthe study by eligible persons\n(> 80%)\n\nHigh bias: The\nrelationship between the\nPF and outcome is very\nlikely to be different for\nparticipants and eligible\nnonparticipantsModerate\nbias:The relationship\nbetween the PF and\noutcome may be different\nfor participants and\neligible\nnonparticipantsLow bias:\nThe relationship between\n\nb. Description of the source\npopulation or population of\ninterest\n\nc. Description of the baseline\nstudy sample\n\nd. Adequate description of\nthe sampling frame and\nrecruitment.\n\nRisk of bias criteria and justifications (Continued)\n\nthe PF and outcome is\nunlikely to be different for\nparticipants and eligible\nnonparticipants\n\ne. Adequate description of\nthe period and place of\nrecruitment\n\nf. Adequate description of\ninclusion and exclusion\ncriteria\n\n2) Study\nattrition\n\nThe study data available\n(i.e., participants not lost to\nfollow-up) adequately rep-\nresent the study sample\n\nSummary\n\na. Adequate response rate\nfor study participants (> 80%)\n\nHigh bias: The\nrelationship between the\nPF and outcome is very\nlikely to be different for\ncompleting and\nnoncompeting\nparticipantsModerate\nbias: The relationship\nbetween the PF and\noutcome may be different\nfor completing and\nnoncompeting\nparticipantsLow bias: The\nrelationship between the\nPF and outcome is\nunlikely to be different for\ncompleting and\nnoncompeting\nparticipants\n\nb. Description of attempts to\ncollect information on\nparticipants who dropped\nout\n\nc. Reasons for loss to follow-\nup are provided\n\nd. Adequate description of\nparticipants lost to follow-up\n\ne. There are no important\ndifferences between\nparticipants who completed\nthe study and who did not\n\n3) Prognostic\nfactor\nmeasurement\n\nThe PF is measured in a\nsimilar way for all\nparticipants\n\nSummary\n\na. A clear definition or\ndescription of the PF is\nprovided\n\nHigh bias: The\nmeasurement of the PF is\nvery likely to be different\nfor different levels of the\noutcome of\ninterestModerate bias:\n\nb. Method of PF\nmeasurement is adequately\nvalid and reliable (i.e., direct\n\nTable 2 Risk of bias assessment (Continued)\n\n# Study ID Study Prognostic Statistical\nanalysis\nand\nreporting\n\nparticipant Attrition Factor measurement Outcome measurement Study confounding\n\n125 Ntamatungiro 2017 Low High Low Low Low High\n\n126 Ongubo 2017 High High Low Low Low High\n\n127 Onoya 2016 Low High Low Low Low Low\n\n128 Palladino 2013 High Low Low Low Low Low\n\n129 Patrikar 2017 Low Low Low High Low High\n\n130 Penot 2014 High Low Low High Low High\n\n131 Raimondo 2017 Low Low Low low Low Low\n\n132 Rajasekaran 2007 Low Low Low High High Low\n\n133 Ramadhani 2007 High Low Low low Low High\n\n134 Rangarajan 2016 Low High Low low Low Low\n\n135 Rohr 2016 Low High Low low Unclear Low\n\n136 Ruperez 2014 High Low Low low Low High\n\n137 Ruperez 2015 High Low Low low Low High\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 13 of 17\n\n\n\nRisk of bias criteria and justifications (Continued)\n\nThe measurement of the\nPF may be different for\ndifferent levels of the\noutcome of interestLow\nbias: The measurement of\nthe PF is unlikely to be\ndifferent for different\nlevels of the outcome of\ninterest\n\nascertainment; secure record,\nhospital record)\n\nc. Continuous variables are\nreported or appropriate cut-\npoints are used\n\nd. The method and setting\nof measurement of PF is the\nsame for all study\nparticipants\n\ne. Adequate proportion of\nthe study sample has\ncomplete data for the PF (>\n80%)\n\nf. Appropriate methods of\nimputation are used for\nmissing PF data\n\n4) Outcome\nmeasurement\n\nThe outcome of interest is\nmeasured in a similar way\nfor all participants\n\nSummary\n\na. A clear definition of the\noutcome of interest is\nprovided (including the time\nof death)\n\nHigh bias: The\nmeasurement of the\noutcome is very likely to\nbe differently related to\nthe baseline level of the\nPFModerate bias: The\nmeasurement of the\noutcome may be\ndifferently related to the\nbaseline level of the\nPFLow bias: The\nmeasurement of the\noutcome is unlikely to be\ndifferently related to the\nbaseline level of the PF\n\nb. Method of outcome\nmeasurement used is\nadequately valid and reliable\n(i.e. independent blind\nassessment, hospital record\nor record linkage)\n\nc. The method and setting of\noutcome measurement is\nthe same for all study\nparticipants\n\n5) Study\nconfounding\n\nImportant potential\nconfounder is appropriately\naccounted for\n\nSummary\n\na. Most important\nconfounders are measured\n\nHigh bias: The observed\neffect of the PF on the\noutcome is very likely to\nbe distorted by another\nfactor related to PF and\noutcomeModerate bias:\nThe observed effect of the\nPF on outcome may be\ndistorted by another\nfactor related to PF and\noutcomeLow bias: The\nobserved effect of the PF\non the outcome is\nunlikely to be distorted by\nanother factor related to\nPF and outcome\n\nb. Clear definitions of the\nimportant confounders\nmeasured are provided\n\nc. Measurement of all\nimportant confounders is\nadequately valid and reliable\n\nd. The method and setting\nof confounding\nmeasurement are the same\nfor all study participants\n\ne. Appropriate methods are\nused if imputation is used for\nmissing confounder data\n\nf. Important potential\nconfounders are accounted\nfor in the study design (by\nlimiting the study to specific\npopulation groups, or by\nmatching)\n\ng. Important potential\nconfounders are accounted\nfor in the analysis (by\n\nRisk of bias criteria and justifications (Continued)\n\nstratification, multivariate\nregression)\n\n6) Statistical\nanalysis and\npresentation\n\nThe statistical analysis is\nappropriate, and all primary\noutcomes are reported\n\nSummary\n\na. Sufficient presentation of\ndata to assess the adequacy\nof the analytic strategy\n\nHigh bias: The reported\nresults are very likely to\nbe spurious or biased\nrelated to analysis or\nreportingModerate bias:\nThe reported results may\nbe spurious or biased\nrelated to analysis or\nreportingLow bias: The\nreported results are\nunlikely to be spurious or\nbiased related to analysis\nor reporting\n\nb. Strategy for model\nbuilding is appropriate and is\nbased on a conceptual\nframework or model\n\nc. The selected statistical\nmodel is adequate for the\ndesign of the study\n\nd. There is no selective\nreporting of results (based on\nthe study protocol, if available,\nor on the “Methods” section)\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 14 of 17\n\n\n\nAbbreviations\nAIDS: Acquired immune deficiency syndrome; ART: Antiretroviral therapy;\nBMI: Body mass index; CDC: Centers for Disease Control and Prevention;\nCENTRAL: Cochrane Central Register of Controlled Trials; EMBASE: Excerpta\nMedica Database; HRs: Hazard ratios; HIV: Human immunodeficiency virus;\nIRIS: Immune reconstitution inflammatory syndrome; LILACS: Latin American\nand Caribbean Health Sciences Literature; LMICs: Low- and middle-income\ncountries; MEDLINE: Medical Literature Analysis and Retrieval System Online;\nNNRTI: Non-nucleotide reverse transcriptase inhibitors; NRTI: Nucleotide\nreverse transcriptase inhibitors; OR: Odds ratio; PLHIV: People living with\nhuman immunodeficiency virus; PRISMA-P: Preferred Reporting Items for\nSystematic Reviews and Meta-Analysis Protocols; PROSPERO: Prospective\nRegister of Systematic Reviews; PubMed: Public/Publisher MEDLINE; RRs: Risk\nratios; SSA: Sub-Saharan Africa; TB: Tuberculosis; UNAIDS: United Nations\nProgramme on HIV; WHO: World Health Organization\n\nAcknowledgements\nNot applicable.\n\nAuthors’ contributions\nYL, MK, and PN contributed to the conceptualization of the project. YL, MK,\nSJ, OA, and PN designed the search strategy, study selection process, and\ndrafting of the manuscript. YL, MK, SJ, OA, and PN contributed to critically\nreviewing the manuscript. PN is the guarantor. The authors gave the final\napproval of the manuscript for publication.\n\nFunding\nNo funding was received for this systematic review.\n\nAvailability of data and materials\nAll data generated or analyzed during this study are included in this\npublished article and its additional files.\n\nEthics approval and consent to participate\nNot applicable because no primary data were collected.\n\nConsent for publication\nNot applicable\n\nCompeting interests\nThe authors declare that they have no competing interests.\n\nAuthor details\n1Division of Epidemiology and Biostatistics, Department of Global Health,\nFaculty of Medicine and Health Sciences, Stellenbosch University, Cape\nTown, South Africa. 2Médecins Sans Frontières (MSF), Eshowe, KwaZulu Natal,\nSouth Africa. 3Division of Epidemiology & Biostatistics, School of Public\nHealth, Faculty of Health Sciences, University of the Witwatersrand,\nJohannesburg, South Africa.\n\nReceived: 5 May 2020 Accepted: 8 November 2020\n\nReferences\n1. UNAIDS. UNAIDS Report on the Global AIDS Epidemic 2019; 2018. p. 1–376.\n2. Renaud-théry AF, Duncombe C, Kerr S, Thierry S. Adult antiretroviral therapy\n\nin resource limited settings : a systematic review of first-line failure and\nattrition rates. WHO Tech Br. :1–18.\n\n3. Boyd MA. Current and future management of treatment failure in low- and\nmiddle-income countries. Curr Opin HIV AIDS. 2010;5(1):83–9.\n\n4. Joint United Nations Programme on HIV/AIDS (UNAIDS). Ambitious\ntreatment targets: writing the final chapter of the AIDS epidemic. Geneva:.\n\n5. Joint United Nations Programme on HIV/AIDS. Global AIDS update 2016.\nGeneva: UNAIDS.\n\n6. El-Khatib Z, Katzenstein D, Marrone G, Laher F, Mohapi L, Petzold M, Morris\nL, Ekström AM. Adherence to drug-refill is a useful early warning indicator\nof virologic and immunologic failure among HIV patients on first-line ART in\nSouth Africa. PloS one. 2011;6(3):e17518. https://doi.org/10.1371/journal.\npone.0017518.\n\n7. Vanobberghen FM, Kilama B, Wringe A, Ramadhani A, Zaba B, Mmbando D,\nTodd J. Immunological failure of first-line and switch to second-line\nantiretroviral therapy among HIV-infected persons in Tanzania: analysis of\nroutinely collected national data. Tropical Med Int Health. 2015;20(7):880–92.\n\nAppendix 4\nStrength of evidence\n\nTable 3 Summary of findings of included studies using the GRADE methodology (Grading of Recommendations Assessment,\nDevelopment and Evaluation)\n\nFactors assessed Number of studies\n(SD)\n\nMain findings Strength of evidence (high, moderate, low,\nvery low)\n\nAdherence (poor versus good) 6 (cross-sectional) Odds ratio: 5.90 (95%CI, 3.50–\n9.94)\n\nModeratea\n\nAdherence (poor versus good) 4 (cohort studies) Hazar ratio: 2.46 (95% CI, 1.72–\n3.51)\n\nHigh\n\nCD4 cell count (< 200 versus ≥ 200\ncells/mm3)\n\n3 (cross-sectional) Odd ratio: 4.82 (95% CI, 2.44–\n9.52)\n\nLowb\n\nCD4 cell count (< 200 versus ≥ 200\ncells/mm3)\n\n4 (cohort studies) Hazard ratio: 2.98 (95% CI,\n2.23–4.0)\n\nModeratec\n\nCD4 cell count (< 100 versus ≥ 100\ncells/mm3)\n\n2 (cross-sectional) Odds ratio: 1.14 (95% CI, 0.52–\n2.47)\n\nLowd\n\nSD study design\naDowngraded once to indirectness, the final sample of some of the included studies only represents the population of interest\nbImprecision and inconsistency were major concerns, imprecision due to a limited number of studies and wide confidence intervals, and there was a substantial\nheterogeneity statistical heterogeneity (heterogeneity: Tau2 = 0.25; chi2 = 6.25, df = 2(P = 0.03), I2 = 71%) and marked clinical heterogeneity\ncDowngraded once due to a risk of bias, bias to statistical analysis and reporting, and potential confounding factors\ndImprecision due to a limited number of participants and studies included. Inconsistency as there was a moderate statistical heterogeneity (heterogeneity: Tau2 =\n0.18; chi2 = 1.95, df = 1 (P = 0.16); I2 = 49%)\nGRADE Working Group grades of evidence\nHigh certainty: We are very confident that the true effect lies close to that of the estimate of the effect\nModerate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility\nthat it is substantially different\nLow certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 15 of 17\n\nhttps://doi.org/10.1371/journal.pone.0017518\nhttps://doi.org/10.1371/journal.pone.0017518\n\n\n8. Ncaca LN, Kranzer K, Orrell C. Treatment interruption and variation in tablet\ntaking behaviour result in viral failure: a case-control study from Cape Town,\nSouth Africa. PloS One. 2011;6(8):e23088. https://doi.org/10.1371/journal.\npone.0023088.\n\n9. Yayehirad AM, Mamo WT, Gizachew AT, Tadesse AA. Rate of immunological\nfailure and its predictors among patients on highly active antiretroviral\ntherapy at Debremarkos Hospital, Northwest Ethiopia: a retrospective follow\nup study. J AIDS Clin Res. 2013;4(5). https://doi.org/10.4172/2155-6113.\n1000211.\n\n10. World Health Organisation. Consolidated guidelines on the use of\nantiretroviral drugs for treating and preventing HIV infection:\nrecommendations for a public health approach. WHO Guidelines. 2013; 272.\n\n11. Bulage L, Ssewanyana I, Nankabirwa V, Nsubuga F, Kihembo C, Pande G,\nArio AR, Matovu JK, Wanyenze RK, Kiyaga C. Factors associated with\nvirological non-suppression among HIV-positive patients on antiretroviral\ntherapy in Uganda, August 2014–July 2015. J BioMed Central Infectious Dis.\n2017;17(1):326.\n\n12. Palladino C, Briz V, Bellón JM, Bartolo I, Carvalho P, Camacho R, et al.\nPredictors of attrition and immunological failure in HIV-1 patients on highly\nactive antiretroviral therapy from different healthcare settings in\nMozambique. PLoS One. 2013;8(12):e82718. https://doi.org/10.1371/journal.\npone.0082718.\n\n13. Anude CJ, Eze E, Onyegbutulem HC, Charurat M, Etiebet MA, Ajayi S, Dakum\nP, Akinwande O, Beyrer C, Abimiku AL, Blattner W. Immuno-virologic\noutcomes and immuno-virologic discordance among adults alive and on\nanti-retroviral therapy at 12 months in Nigeria. BioMed Central Infectious\nDis. 2013;13(1):113.\n\n14. Hassan AS, Nabwera HM, Mwaringa SM, Obonyo CA, Sanders EJ, de Wit TF,\nCane PA, Berkley JA. HIV-1 virologic failure and acquired drug resistance\namong first-line antiretroviral experienced adults at a rural HIV clinic in\ncoastal Kenya: a cross-sectional study. BioMed Central AIDS Research\nTherapy. 2014;11(1):9.\n\n15. Meriki HD, Tufon KA, Afegenwi MH, Nyindem BA, Atanga PN, Anong DN,\nCho-Ngwa F, Nkuo-Akenji T. Immuno-haematologic and virologic responses\nand predictors of virologic failure in HIV-1 infected adults on first-line\nantiretroviral therapy in Cameroon. BioMed Central Infectious Diseases\nPoverty. 2014;3(1):5.\n\n16. Greig JE, du Cros PA, Mills C, Ugwoeruchukwu W, Etsetowaghan A, Grillo A,\nTayo-Adetoro A, Omiyale K, Spelman T, O’Brien DP. Predictors of raised viral\nload during antiretroviral therapy in patients with and without prior\nantiretroviral use: a cross-sectional study. PLoS One. 2013;8(8):e71407.\n\n17. WHO. Consolidated guidelines on the use of antiretroviral drugs for treating\nand preventing HIV infection: recommendations for a public health\napproach– 2nd ed. Geneva, WHO, 2016.\n\n18. Kwobah CM, Mwangi AW, Koech JK, Simiyu GN, Siika AM. Factors associated\nwith first-line antiretroviral therapy failure amongst HIV-infected African\npatients: a case-control study. World J AIDS. 2012;2(4):271–8.\n\n19. Mingbunjerdsuk P, Asdamongkol N, Sungkanuparph S. Factors associated\nwith immunological discordance in HIV-infected patients receiving\nantiretroviral therapy with complete viral suppression in a resource-limited\nsetting. Japanese J Infectious Dis. 2015;68(4):301–4.\n\n20. Liégeois F, Vella C, Eymard-Duvernay S, Sica J, Makosso L, Mouinga-Ondémé\nA, Mongo AD, Boué V, Butel C, Peeters M, Gonzalez JP. Virological failure\nrates and HIV-1 drug resistance patterns in patients on first-line\nantiretroviral treatment in semirural and rural Gabon. J Int AIDS Society.\n2012;15(2):17985.\n\n21. Kassa D, Gebremichael G, Alemayehu Y, et al. Virologic and immunologic\noutcome of HAART in Human Immunodeficiency Virus (HIV)-1 infected\npatients with and without tuberculosis (TB) and latent TB infection (LTBI) in\nAddis Ababa, Ethiopia. AIDS Res Ther. 2013;10:18. https://doi.org/10.1186/\n1742-6405-10-18.\n\n22. JORGE, Alave R., et al. Risk factors associated with virologic failure in HIV-\ninfected patients receiving antiretroviral therapy at a public hospital in Peru.\nRevista chilena de infectologia: organo oficial de la Sociedad Chilena de\nInfectologia, 2013, 30.1: 42.\n\n23. SANG, R. K. A.; MIRUKA, F. O. Factors associated with virologic failure\namongst adults on antiretroviral therapy in Nyanza Region, Kenya. IOSR J\nDental Med Sci, 2016, 15.7: 108-121.\n\n24. Putting HIV and HCV to the test: a product guide for point-of-care CD4 and\nlaboratory-based and point-of-care virological HIV and HCV tests. Geneva:\nMédecins Sans Frontieres. 2015. 2nd edition –reported on 14 july 2015.\n\n25. Braithwaite RS, Nucifora KA, Toohey C, Kessler J, Uhler LM, Mentor SM,\nKeebler D, Hallett T. How do different eligibility guidelines for antiretroviral\ntherapy affect the cost–effectiveness of routine viral load testing in sub-\nSaharan Africa? AIDS (London, England). 2014;28(1):S73-23.\n\n26. Estill J, Egger M, Johnson LF, Gsponer T, Wandeler G, Davies MA, Boulle A,\nWood R, Garone D, Stringer JS, Hallett TB, Keiser O. IeDEA Southern Africa\nCollaboration. Monitoring of antiretroviral therapy and mortality in HIV\nprogrammes in Malawi, South Africa and Zambia: mathematical modelling\nstudy. PLoS One. 2013;8(2):e57611. https://doi.org/10.1371/journal.pone.\n0057611.\n\n27. Keebler D, Revill P, Braithwaite S, Phillips A, Blaser N, Borquez A, Cambiano\nV, Ciaranello A, Estill J, Gray R, Hill A. Cost-effectiveness of different strategies\nto monitor adults on antiretroviral treatment: a combined analysis of three\nmathematical models. Lancet Global Health. 2014;2(1):e35–43.\n\n28. World Health Organization. Antiretroviral therapy (ART) coverage among all\nage groups. World Health Organization. Available from https://www.who.\nint/data/gho/hiv/epidemic_response/ART/en/.\n\n29. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and\nPreventing HIV Infection: Recommendations for a Public Health Approach.\n2nd edition. Geneva: World Health Organization; 2016. Available from:\nhttps://www.ncbi.nlm.nih.gov/books/NBK374294/.\n\n30. Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P,\nStewart LA. Preferred reporting items for systematic review and meta-\nanalysis protocols (PRISMA-P) 2015: elaboration and explanation. Brit Med J.\n2015;349:g7647.\n\n31. Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M,\nHenry D, Altman DG, Ansari MT, Boutron I, Carpenter JR. ROBINS-I: a tool for\nassessing risk of bias in non-randomised studies of interventions. Brit Med J.\n2016;355:i4919.\n\n32. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in\nmeta-analyses. Brit Med J. 2003;327(7414):557.\n\n33. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected\nby a simple, graphical test. British Medical Journal. 1997 Sep 13;315(7109):\n629–34.\n\n34. Babo YD, Alemie GA, Fentaye FW. Predictors of first-line antiretroviral\ntherapy failure amongst HIV-infected adult clients at Woldia Hospital,\nNortheast Ethiopia. PloS one. 2017;12(11):e0187694. https://doi.org/10.1371/\njournal.pone.0187694.\n\n35. Bayu B, Tariku A, Bulti AB, Habitu YA, Derso T, Teshome DF. Determinants of\nvirological failure among patients on highly active antiretroviral therapy in\nUniversity of Gondar Referral Hospital, Northwest Ethiopia: a case–control\nstudy. HIV/AIDS (Auckland, NZ). 2017;9:153.\n\n36. Enderis BO, Hebo SH, Debir MK, Sidamo NB, Shimber MS. Predictors of time\nto first line antiretroviral treatment failure among adult patients living with\nHIV in public health facilities of Arba Minch Town, Southern Ethiopia.\nEthiopian Journal of Health Sciences. 2019;29(2):175.\n\n37. Bisson GP, Rowh A, Weinstein R, Gaolathe T, Frank I, Gross R. Antiretroviral\nfailure despite high levels of adherence: discordant adherence–response\nrelationship in Botswana. J Acquir Immune Defic Syndr (1999). 2008;49(1):107.\n\n38. Fatti G, Grimwood A, Nachega JB, Nelson JA, LaSorda K, van Zyl G,\nGrobbelaar N, Ayles H, Hayes R, Beyers N, Fidler S. Better virological\noutcomes among people living with human immunodeficiency virus (HIV)\ninitiating early antiretroviral treatment (CD4 counts≥500 cells/mm3) in the\nHIV prevention trials network 071 (PopART) trial in South Africa. Clin\nInfectious Dis. 2020;70(3):395–403.\n\n39. Ford N, Darder M, Spelman T, Maclean E, Mills E, Boulle A. Early adherence\nto antiretroviral medication as a predictor of long-term HIV virological\nsuppression: five-year follow up of an observational cohort. PloS one. 2010;\n5(5):e10460. https://doi.org/10.1371/journal.pone.0010460.\n\n40. Gunda DW, Kilonzo SB, Mtaki T, Bernard DM, Kalluvya SE, Shao ER.\nMagnitude and correlates of virological failure among adult HIV patients\nreceiving PI based second line ART regimens in north western Tanzania; a\ncase control study. BMC infectious diseases. 2019;19(1):235.\n\n41. Haile D, Takele A, Gashaw K, Demelash H, Nigatu D. Predictors of treatment\nfailure among adult antiretroviral treatment (ART) clients in Bale zone\nhospitals, south eastern Ethiopia. PloS One. 2016;11(10):e0164299. https://\ndoi.org/10.1371/journal.pone.0164299.\n\n42. Hailu GG, Hagos DG, Hagos AK, Wasihun AG, Dejene TA. Virological and\nimmunological failure of HAART and associated risk factors among adults\nand adolescents in the Tigray region of Northern Ethiopia. PloS One. 2018;\n13(5).\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 16 of 17\n\nhttps://doi.org/10.1371/journal.pone.0023088\nhttps://doi.org/10.1371/journal.pone.0023088\nhttps://doi.org/10.4172/2155-6113.1000211\nhttps://doi.org/10.4172/2155-6113.1000211\nhttps://doi.org/10.1371/journal.pone.0082718\nhttps://doi.org/10.1371/journal.pone.0082718\nhttps://doi.org/10.1186/1742-6405-10-18\nhttps://doi.org/10.1186/1742-6405-10-18\nhttps://doi.org/10.1371/journal.pone.0057611\nhttps://doi.org/10.1371/journal.pone.0057611\nhttps://www.who.int/data/gho/hiv/epidemic_response/ART/en/\nhttps://www.who.int/data/gho/hiv/epidemic_response/ART/en/\nhttps://www.ncbi.nlm.nih.gov/books/NBK374294/\nhttps://doi.org/10.1371/journal.pone.0187694\nhttps://doi.org/10.1371/journal.pone.0187694\nhttps://doi.org/10.1371/journal.pone.0010460\nhttps://doi.org/10.1371/journal.pone.0164299\nhttps://doi.org/10.1371/journal.pone.0164299\n\n\n43. Izudi J, Alioni S, Kerukadho E, Ndungutse D. Virological failure reduced with\nHIV-serostatus disclosure, extra baseline weight and rising CD4 cells among\nHIV-positive adults in north-western Uganda. BMC infectious diseases. 2016;\n16(1):614.\n\n44. Karade SK, Ghate MV, Chaturbhuj DN, Kadam DB, Shankar S, Gaikwad N,\nGurav S, Joshi R, Sane SS, Kulkarni SS, Kurle SN, Paranjape RS, Rewari BB,\nGangakhedkar RR. Cross-sectional study of virological failure and\nmultinucleoside reverse transcriptase inhibitor resistance at 12 months of\nantiretroviral therapy in Western India. Medicine (Baltimore). 2016;95(37):\ne4886. https://doi.org/10.1097/MD.0000000000004886.\n\n45. Lay P, An S, Soeung S, Srey PS, Thai S, Lynen L, Griensven J. Ten year\nexperience with antiretroviral treatment in Cambodia: trends in patient\ncharacteristics and treatment outcomes. PloS One. 2017;12(11):e0185348.\nhttps://doi.org/10.1371/journal.pone.0185348.\n\n46. Ndahimana JD, Riedel DJ, Mwumvaneza M, Sebuhoro D, Uwimbabazi JC,\nKubwimana M, Mugabo J, Mulindabigwi A, Kirk C, Kanters S, Forrest JI. Drug\nresistance mutations after the first 12 months on antiretroviral therapy and\ndeterminants of virological failure in Rwanda. J Tropical Med Int Health.\n2016;21(7):928–35.\n\n47. Ahmed M, Merga H, Jarso H. Predictors of virological treatment failure\namong adult HIV patients on first-line antiretroviral therapy in Woldia and\nDessie hospitals, Northeast Ethiopia: a case-control study. BMC Infectious\nDiseases. 2019;19(1):305.\n\n48. Lay P, An S, Soeung S, Srey PS, Thai S, Lynen L, et al. Ten year experience\nwith antiretroviral treatment in Cambodia: trends in patient characteristics\nand treatment outcomes. PLoS One. 2017;12(11):1–16.\n\n49. Chakraborty A, Hershow RC, Qato DM, Stayner L, Dworkin MS. Adherence to\nantiretroviral therapy among HIV patients in India: a systematic review and\nmeta-analysis. AIDS Behav. 2020;24(7):2130–48.\n\n50. AIDSinfo. Guidelines for the use of antiretroviral agents in HIV-1-infected\nadults and adolescents. 2017.\n\n51. Matare T, Shambira G, Gombe N, Tshimanga M, Bangure D. Factors\nassociated with human immunodeficiency virus first-line treatment failure in\nZvishavane District, Zimbabwe, 2014. Austin J HIV/AIDS Research. 2015;2(1):\n1010.\n\n52. Assemie MA, Alene M, Ketema DB, Mulatu S. Treatment failure and\nassociated factors among first line patients on highly active antiretroviral\ntherapy in Ethiopia: a systematic review and meta-analysis. Global Health\nResearch Policy. 2019;4(1):1-0.\n\n53. Thao VP, Quang VM, Wolbers M, Anh ND, Shikuma C, Farrar J, et al. Second-\nline HIV therapy outcomes and determinants of mortality at the largest HIV\nreferral Center in Southern Vietnam. Medicine. 2015;94(43).\n\n54. Zaragoza-Macias E, Cosco D, Nguyen ML, Del Rio C, Lennox J. Predictors of\nsuccess with highly active antiretroviral therapy in an antiretroviral-naive\nurban population. AIDS Research and Human Retroviruses. 2010;26(2):133–8.\n\n55. Krüsi A, Milloy MJ, Kerr T, Zhang R, Guillemi S, Hogg R, Montaner J, Wood E.\nOngoing drug use and outcomes from highly active antiretroviral therapy\namong injection drug users in a Canadian setting. Antiviral therapy. 2010;\n15(5):789.\n\n56. Nachega JB, Marconi VC, van Zyl GU, Gardner EM, Preiser W, Hong SY, et al.\nHIV treatment adherence, drug resistance, virologic failure: evolving\nconcepts. Infect Disord-Drug Targets (Formerly Current Drug Targets-\nInfectious Disorders). 2011;11(2):167–74.\n\n57. Olowookere SA, Adewole IF. Effects of adherence to antiretroviral therapy\non body mass index, immunological and virological status of Nigerians\nliving with HIV/AIDS. Alexandria J Med. 2016;52(1):51–4.\n\nPublisher’s Note\nSpringer Nature remains neutral with regard to jurisdictional claims in\npublished maps and institutional affiliations.\n\nLailulo et al. Systematic Reviews (2020) 9:292 Page 17 of 17\n\nhttps://doi.org/10.1097/MD.0000000000004886\nhttps://doi.org/10.1371/journal.pone.0185348\n\n\tAbstract\n\tBackground\n\tMethods\n\tResults\n\tConclusion\n\tSystematic review registration\n\n\tBackground\n\tAntiretroviral treatment failure\n\tFactors associated with treatment failure\n\tImportance of the review\n\n\tObjective\n\tPrimary objective\n\tSecondary objective\n\n\tMethods\n\tCriteria for considering studies for review\n\tTypes of studies\n\tTypes of participants\n\n\tType of outcome\n\tPrimary outcome\n\tVirological failure\n\tImmunological failure\n\tClinical failure\n\tSecondary outcome\n\n\tInclusion and exclusion criteria\n\tIncluded studies\n\tExcluded studies\n\tSearch methods for identification of the studies\n\tScreening and assessments of study eligibility\n\tData extraction, management, and analysis\n\tRisk of bias and quality of evidence\n\tStatistical analysis\n\n\n\tResults\n\tPRISMA flow chart\n\tMeta-analysis\n\tRisk of bias assessment\n\n\tDiscussion\n\tStrength of evidence\n\n\tConclusion\n\tAppendix 1\n\tSearch strategy—database\n\n\tAppendix 3\n\tAppendix 2\n\tRisk of bias criteria and justifications\n\tAbbreviations\n\n\tAcknowledgements\n\tAuthors’ contributions\n\tFunding\n\tAvailability of data and materials\n\tEthics approval and consent to participate\n\tConsent for publication\n\tCompeting interests\n\tAuthor details\n\tReferences\n\tAppendix 4\n\tPublisher’s Note\n\n","UrlSlug":"factors-associated-antiretroviral-treatment-failure-among-people-living-hiv-antiretroviral-therapy-resource-poor-settings-systematic-review","collections":[],"countries":[],"asset_languages":[{"AlternateTitle":null,"language":[{"Id":1,"Name":"English"}],"contentTranslation":[],"contentLocation":[]}],"creators":[{"Name":"Kitenge M","Id":19691,"ac":[{"Order":1}]},{"Name":"Nyasulu P","Id":19695,"ac":[{"Order":4}]},{"Name":"Lailulo Y","Id":26863,"ac":[{"Order":0}]},{"Name":"Jaffer S","Id":26864,"ac":[{"Order":2}]},{"Name":"Aluko O","Id":26865,"ac":[{"Order":3}]}],"keywords":[{"Name":"HIV/AIDS","Id":73}]}}

Science Portal

Factors associated with antiretroviral treatment failure among people living with HIV on antiretroviral therapy in resource-poor settings: a systematic review and metaanalysis