We conducted a comprehensive search to identify all relevant studies related to parent to child transmission of HIV, treatment and survival of CLHIV (Box 1). We searched the following databases: Medline, Pubmed, IndMED, and used search engines like Web of Science and Google Scholar. To locate additional studies and reports, we screened the table of contents of journals that publish studies related to HIV/AIDS and scanned the websites of national and international organizations like National AIDS Control Organization (NACO), World Health Organization (WHO), United States Agency for International Development (USAID), and United Nations Program on HIV/AIDS(UNAIDS). We also searched the reference lists of articles included in the systematic review.

We conducted the search using three sets of key words. 1) Prevention of Parent to Child Transmission OR PPTCT, Parent to Child Transmission OR MTCT, Antenatal prophylaxis, Nevirapine prophylaxis, Option B regimen, Type of delivery, Delivery, Caesarean, Infant feeding, Mixed Feeding, Replacement feeding, Sero-dis- cordance, AND India; 2) Survival, Early Infant Diagnosis, Access to care, CD 4 count, Adherence, AND India AND HIV and Children OR Pediatric; 3) Anti-retro Viral Therapy OR ART, Adherence, Effectiveness, Efficacy, Survival, AND India AND Children OR Pediatric. The search was restricted to studies published between 1992 and 2015. All studies published in English or with a detailed summary in English were included in the review. The search was conducted in January 2015.

We screened the titles of all the search outputs, and excluded studies that were not found to be relevant. Abstracts of all the titles found to be relevant to the present review were screened, and those that met in the inclusion criteria, as given in Box 1, were included. We exported the search results to EndNote reference manager software. A structured data extraction sheet was used, based on the Cochrane data extraction sheet and the New Castle Ottawa scale. Quality assessment for randomized control trials was done using a five-point scale, namely the process of randomization, treatment allocation, baseline comparability, blinding and drop-out rate. However, we did not screen or exclude studies based on methodological quality as the purview of this review was exploration rather than estimation. Data extraction was done under the following heads: domain and sub-domain of the study, study identifier i.e. last name of author, year of study and publication, PubMed ID, study characteristics (study design, study setting, objective, sample size, and inclusion and exclusion criteria), participant characteristics for studies related to survival and treatment (sex wise, mean age, and age group), objective, exposure, outcome and important findings, drop-out rate and adjustment for confounders.

We retrieved a total of 842 titles from the search term alone or in various combinations. Of these, 686 were not found to be relevant to the present review or were duplicate searches―i.e. the same title retrieved from different

Box 1. Study selection criteria.

sets of keywords. The abstract and full texts of the remaining 156 studies were screened, of which 33 met the inclusion criteria. These 33 selected studies were thematically grouped into three categories: i.e. studies related to parent to child transmission of HIV (19 studies), treatment of children living with HIV (11 studies), and survival of children living with HIV (3 studies). The results of the literature search and thematic categorization of studies are presented in Figure 1.

All the studies included in the review were conducted after 2000, and were health facility-based studies. Most (88%) were conducted in urban public health facilities. While around half were conducted in the high HIV prevalence states of Andhra Pradesh, Tamil Nadu and Maharashtra, 20% were conducted in West Bengal and 10% in Gujarat. Only three studies were conducted in the states of north India. In all, 28 studies were interventional or observational follow-up studies, 2 were cross-sectional studies and 3 were retrospective cohort studies (Table 1).

Most studies explored the effectiveness of single dose nevirapine prophylaxis, and were conducted in a public health facility in southern India. All except two studies had a sample size of between 30 and 1890 HIV infected pregnant women.

The HIV sero-positivity rate among antenatal women ranged from 0.17% to 0.56% [13] [15] [22] . HIV transmission from HIV-infected parent to child who received nevirapine prophylaxis ranged from 2.1% to

27.3% [7] [10] [11] [20] . In a study conducted in a secondary care setting, the parent to child transmission rate was 5.2% among 154 parent-baby pairs who were fully compliant with the then existing protocol of nevirapine prohylaxis, as compared to 17.5% among non-compliant pairs [10] . In a study conducted in 33 PPTCT centers in 18 districts of Gujarat, infants who received ART for prevention were at lower risk of HIV infection than those who did not (5.2% vs 41.1%, p value < 0.0001) [12] . A study in West Bengal comparing the efficacy of a single dose versus an extended dose regimen found that the HIV transmission rate in the former group was 8.9% as compared to 10.4% in the latter group [17] . However, the difference was not statistically significant. Cumulative mortality in the extended dose regimen group was also significantly less (risk ratio = 0.53, 95% CI 0.32, 0.85) than in the single dose regimen group [17] . In a study in Maharashtra, the relative risk of acquiring HIV in infants was 0.80 (RR = 0.80, 95% C.I. 0.58 - 1.10) when given extended dose regimen as compared to a single dose regimen [21] .

Few studies compared the HIV transmission rate with respect to feeding patterns. A study in Maharashtra showed that infants who were exclusively formula fed had a lower transmission rate than the overall transmission rate (3.1% vs 8.0%) [11] . Similar results were seen in studies in Gujarat where the transmission rate was 6.2% in the replacement feeding group as compared to 25.5% in the breast-feeding or mixed feeding group [12] . A study in West Bengal compared the efficacy of both perinatal nevirapine prophylaxis and feeding patterns in a single arm, and found that whereas the HIV transmission rate without any of these interventions was 57.2%, it reduced to 15% in the presence of intervention with nevirapine prophylaxis and total formula feeding [15] . However, contrary results were obtained in a study in Tamil Nadu, where the HIV transmission rate among breast-fed infants was zero percent, compared to 6.5% for the entire cohort [18] .

Most of the studies which compared the HIV transmission rate with respect to type of delivery showed that cesarean section was better than vaginal delivery in preventing HIV transmission. In a cohort study in Gujarat, the transmission rate in the vaginal delivery group was 4.1% as compared to 2.9% among mothers who underwent cesarean section [23] . However, one study in Maharashtra showed that cesarean section delivery was not better than vaginal delivery in reducing the HIV transmission rate [20] .

The presence of concomitant chronic conditions was also found have an effect on transmission of HIV. The parent to child transmission rate among pregnant women co-infected with HIV-TB was 30%, which was significantly more than the transmission rate of 12% in the absence of tuberculosis (p value = 0.02) [24] .

Of the 11 studies covering treatment of CLHIV, four explored levels of ART adherence, and factors associated with non-adherence to treatment; 2 examined the magnitude of loss to follow up among children on ART care, and 3 explored the effectiveness of different highly active anti-retroviral therapy (HAART) regimens in the treatment of HIV-infected children.

The attrition rate for ART among children (14 years or less) living with HIV on ART was 12.6% and 15.6% in two studies conducted in Andhra Pradesh [25] [26] . Children from the lower socio-economic group, orphans, males and younger children (aged less than two years) had a higher likelihood of attrition and thus subsequent mortality than others. Children with lower annual risk of AIDS were more likely to be lost to follow up but had less chance of mortality than others. Improvement in nutritional status was noted in terms of reduction in wasting, increase in body weight z score, and improvement in hemoglobin after the initiation of HAART [28] [29] [35] ; Children showing an improvement in these parameters had a lower risk of subsequent mortality than others.

Appropriate adherence to ART (more than 95% of doses) varied from 65.8% to 95.5% across studies [28] [32] [34] . In a study conducted in a pediatric center for excellence for HIV care, which assessed adherence to treatment using the pill count and three-day recall period methods, appropriate adherence was more than 95%. Reasons for non-adherence cited by caregivers were multiple caregivers, work constraints of the caregiver, and death or illness in the family [28] . However, a study in an ART center in Delhi reported low adherence (less than 95% of doses) in 34.4% of the children who had initiated ART. Factors associated with low adherence were low educational status of the caregiver, late initiation of ART after diagnosis, orphan status and female sex [28] . Thus, there was a considerable difference in the adherence rate in the two types of settings. One study in Tamil Nadu showed that mortality and drug resistance among children 2 - 36 months of age were higher in a nevirapine alone regimen as compared to a protease inhibitor boosted lopinavir regimen [33] .

Of the three studies that assessed the survival pattern of CLHIV, one study each reported time to infant diagnosis, time of entry into care and CD 4 count at the time of diagnosis as factors related to survival. CD 4 count at the time of diagnosis was an important predictor of progression of HIV [37] . Children with a baseline CD 4 count of less than or equal to 14% had significantly higher mortality as compared to children with a CD 4 count of more than 20% [38] . Delayed age at diagnosis had an adverse impact on the overall long-term outcome of the disease and subsequent mortality. Attrition of children diagnosed with HIV before being initiated on treatment was an important factor in the progression of the disease and subsequent mortality [36] .

In a cohort study in South India, 13% of the children were lost to follow up in the one year after diagnosis. Most of the children who were not eligible for treatment during the first year were lost to follow up [37] . Children who had a delayed entry into care after ART eligibility were those diagnosed after their parent’s HIV diagnosis, less than 18 months of age, female, orphan, and living more than 90 minutes from the nearest HIV center [38] .