Prevalence of HIV infection in Ukraine is 1.6% overall, with antenatal prevalence of 0.52%, the highest in Europe and the country is facing one of the fastest growing epidemics in the world[1]. According to UNAIDS estimates, 1.5 million people will be living with HIV/AIDS in Ukraine by 2010, two-thirds of whom will be of reproductive age [2]. The number of HIV infected pregnant women delivering in Ukraine has been increasing annually, reaching 5000 in 2007[3]. In Ukraine, the national Prevention of Mother to Child Transmission (PMTCT) Programme was implemented in 2001[2]. This programme has been successful with MTCT rates declining from above 25% in 2001 to 7% in 2006[1, 4]. In 2007 the government implemented more comprehensive HIV/AIDS programmes and policies to scale up access to prevention, diagnosis and treatment services nationally, but antiretroviral therapy (ART) coverage for adult population remains low, with 35% of individuals with advanced HIV infection receiving ART in 2007 [5, 6].

In children, recent studies have demonstrated that early diagnosis of HIV infection followed by early initiation of ART significantly reduces HIV related morbidity and mortality and has resulted in substantially improved survival [7–10]. This, together with the higher risk of clinical progression for vertically-infected infants compared with older children and adults, due to high HIV-1 RNA viral load (VL) and immature immune systems[7–9, 11–16], reflects the essential need for scaling up access to paediatric HIV care and treatment.

Few data are available on the natural and treated history of vertically infected children in Eastern Europe. This information, including treatment patterns, clinical, immunological and virological impact of treatment and long-term outcomes of HIV infection in children, is important to achieve a better understanding of disease progression in this setting and for health care planning and decision-making related to therapeutic strategies. Our objective was to describe the clinical, immunological and virological characteristics, treatment and response in vertically-infected children living in Ukraine and followed from birth in an ongoing cohort study.

The European Collaborative Study (ECS) is an ongoing cohort study, in which HIV-1 infected pregnant women are enrolled and followed in pregnancy, and their children prospectively followed from birth. The ECS was established in 1985 in Western Europe and centres from Ukraine first joined in 2000. ECS enrolment in Ukraine started with three sites (Odesa, Mykolaiv and Simferopol), increasing to six sites in 2006 (addition of Kyiv, Donetsk and Mariupol) and with the addition of Kryvyj Rih in 2009. All pregnant women are screened for HIV-1 infection at first antenatal visit during their pregnancy (routine, universal opt-out system) with repeat testing in the third trimester for women testing negative. A policy for rapid HIV testing during labour for women with undocumented HIV status started in 2003 for those not having antenatal care. Pregnant women diagnosed with HIV infection before pregnancy, during pregnancy or intrapartum are offered to participate in ECS study, with informed consent. Standard questionnaires which are coded anonymised are used for collection of pregnancy, delivery and follow-up data. Infants are followed up at HIV/AIDS Centres until a diagnosis is made. HIV-1 infected children are then followed up ideally at least twice a year. Infected infants were diagnosed based on persistence of HIV antibody beyond age 18 months up until the start of 2006; subsequently, early diagnosis of HIV-exposed infants with DNA PCR testing was introduced nationwide, with facilities in three inter-regional laboratories. Data collected included maternal clinical and sociodemographic status, PMTCT intervention, delivery data, infant demographic and clinical characteristics, ART, other prophylaxis and laboratory markers of disease progression. Flow cytometry became available in Odessa and Simferopol in 2004; Donetsk and Kiev had this capacity from their first enrolments in autumn 2006.

Definition of HIV-1 infection was based on the development of AIDS and HIV-associated morbidity and mortality, persistence of HIV antibody beyond 18 months of age or detectable virus in two or more blood samples. CD4 counts were expressed as absolute counts, which vary with age. Children were therefore assigned to a Centres of Disease Control and Prevention (CDC) immunological category on the basis of their age and CD4 count at the time of each measurement. CDC immunological categories are as follows: normal (CDC immunological category 1), moderate immune suppression (category 2) and severe immune suppression (category 3) [17, 18]. HIV RNA levels were measured with a lower limit of quantification of 400 copies/ml and thus undetectable VL was defined as VL less than 400 copies/ml[10, 19].

We have investigated the therapeutic management, disease progression and survival rate in a birth cohort of vertically-infected children in Ukraine. Nearly half had started HAART by their most recent follow-up visit, with most of these children starting HAART recently (since 2007). In addition to HAART, 31% of children received prophylaxis and 4% received TB treatment. Substantial immunological improvement among children on HAART was evident and, among those with VL measurements available post-initiation, nearly three-quarters had achieved undetectable VL. Overall, almost a quarter of this cohort of infected children had developed AIDS, most in the first year of life, and around one in ten had died, at a median age of 6 months. Overall, survival increased significantly over the study period.

In Ukraine, ART became available for use in adults in 2001, with a somewhat slower roll-out of paediatric ART [15]. Treatment of HIV-infected children has been a priority of the national HIV programme, and the large-scale provision of ART for infected individuals started in 2004. Subsequently, the proportion of children with advanced HIV disease receiving ART increased, although it was estimated that at least a quarter of diagnosed children with advanced HIV disease nationally were not receiving treatment in 2007 [6]. Our findings indicate that more than half of the infected children had not received HAART and of those who received it, 86% had initiated treatment since 2007 onward. In addition, more than a third of moderately or severely immunosuppressed children had not yet started HAART, highlighting the remaining unmet need for treatment.

Overall median age at HAART initiation (18 months) was higher than reported in Western Europe [20], but has declined substantially, to around 9 months in recent years. This partly reflects the fact that in the early years of the study, diagnosis of infection in vertically-exposed infants depended on antibody testing at 15-18 months of age, with diagnosis through PCR testing only introduced in 2006. Another factor behind this trend has been the changes in paediatric treatment protocols over time: up to 2005, treatment of infected children was delayed until eligibility criteria were met. Current WHO guidance recommends treating all HIV-infected infants aged < 12 months with HAART, on the basis of results from the CHER trial and other studies on the clinical and immunological benefits of early initiation of HAART [7–9, 21]. However, this approach requires prompt diagnosis of infected infants and rapid communication of the result to the family and caring physicians, which can be challenging in Ukraine where PCR testing is carried out in just three regional laboratories. Our findings highlight the importance of prompt early diagnosis in this population, as 15.5% of infants had rapid disease progression (developed AIDS or died) in their first year of life. Although PCR testing in Ukraine was available, only a quarter of infected children in our study born in 2007/08 had a PCR test before four months of age. Despite the remaining challenges with respect to early diagnosis of infected infants in Ukraine, the trend towards earlier diagnosis of HIV infection we have documented here, together with greater availability of HAART, are likely to explain the improved estimated survival rate among children in our cohort born in 2004/08 compared with those born earlier. Our finding of a substantially higher rate of AIDS-free survival among children who had never received HAART in comparison with those treated is most likely the result of confounding by indication, whereby children with immunological/clinical indications and thus with poorer survival prognosis were more likely to be prescribed HAART, but also more likely to have or to progress to AIDS or to die. Children on HAART here showed good virological response, and the 75% who achieved undetectable viral load was similar to that reported in other studies - for example 77.5% in Italy and 78% in the UK [20, 22]. However, our results are limited by the fact that many children had only recently started HAART and thus had no post-initiation VL measurement available.

Cotrimoxazole is a widely available antibiotic, recommended by WHO and UNAIDS for prevention of opportunistic infections in individuals living with HIV/AIDS since April 2000[23]. Several guidelines recommend that cotrimoxazole should be started in all infants born to HIV-infected mothers at 6 weeks of age, continuing until confirmation of negative HIV status[24, 25]. Results of one trial showed a 43% reduction in mortality and a 23% reduction in hospitalizations among HIV-infected children receiving cotrimoxazole compared to the control group[26]. Consequently, large-scale provision of cotrimoxazole to all infants born to HIV-infected mothers regardless of their age, CD4 count and resistance to the drug in low resource countries is recommended to prevent mortality and morbidity caused by PCP infection[26]. In our study PCP was the most common AIDS-defining disease followed by recurrent bacterial infection, consistent with other findings in Côte d'Ivoire, Thailand, the U.K., the U.S. and some European countries [11, 22, 27–29]. However, cotrimoxazole was provided to only 29% of children in our cohort. In addition, the majority (80%) of children who died with reported AIDS-defining diseases had infectious diseases including PCP but only 10% had received cotrimoxazole prophylaxis. Despite governmental policy for free cotrimoxazole coverage for HIV-infected children in Ukraine, our findings indicate that infected children do not have adequate access to prophylaxis; furthermore, anecdotal reports suggest that parents will be charged for the drug's purchase. The reason for low provision of cotrimoxazole to prenatally HIV-exposed infants in Ukraine remains unclear. However, anecdotal reports suggest contributing factors may include non-availability of cotrimoxazole in some AIDS Centers, a lack of paediatric formulations and problems with reimbursement to parents. In order to improve morbidity and mortality among vertically HIV infected children, there is an urgent need for scaling-up access to cotrimoxazole prophylaxis.

High prevalence of TB co-infection in HIV-infected people is a well recognised problem in resource-limited settings [30]. The incidence of TB in Ukraine was estimated at 82/100,000 in 2007 [31]. Although TB has been identified as the leading cause of death among HIV-infected population in Ukraine[6], only 15% of HIV-infected individuals with TB received TB treatment in 2006[6], reflecting old-fashioned approaches to TB control and weak linkages between HIV and TB services [32–34]. Children with HIV-infected parents have an increased likelihood of TB exposure [30], mainly due to the increased risk of acquiring TB from within their household; HIV-infected children may also be exposed to TB if staying on the same inpatient wards as HIV-infected adults with active TB [34]. Infants born to HIV-infected mothers do not receive BCG immunization in Ukraine; infants who serorevert are eligible for immunization but not infected children. Furthermore, TB-HIV coinfected children in Eastern Europe often have severe presentation of TB disease [32], In our study co-infection with extra-pulmonary M. tuberculosis infection was reported for 13% of children with AIDS; however, we may have underestimated the true prevalence of TB in the cohort. Optimizing TB prevention, screening and treatment strategies through integration of HIV-TB services are essential to reduce TB related morbidity and mortality among HIV-infected children.

We estimated that overall, 17% of children will have progressed to AIDS or death by age 12 months, increasing by only 4% over the next 6 months. This is consistent with early natural history studies, which demonstrated that 15-20% infants had an early severe disease (opportunistic infections, especially PCP, and encephalopathy), progressing rapidly to AIDS/death within their first year of life, while the remainder had more gradual disease progression, with around 3-5% progressing to AIDS or death annually [11, 35]. Our finding of a significant improvement in AIDS-free survival over time reflects the roll-out of ART for infected children in Ukraine and the trend towards earlier diagnosis of infection in infants born to HIV-positive mothers.

Nearly a fifth of the infected children here were living in an institutional care setting. According to one national report, one-third of HIV-infected children with established status in Ukraine were under institutional care by the end of July 2007[6]. High rates of infant abandonment have also been reported from the Russian Federation [36, 37], and in Western Europe earlier in the HIV epidemic, usually associated with maternal IDU and marginalised social status of many HIV-infected women [2, 38, 39]. Implementation of comprehensive PMTCT programmes may reduce risk of infant abandonment [2, 4, 40], and declining rates of abandonment have been documented concurrent with scale-up of PMTCT [40]. Since institutional care is associated with detrimental affect on all aspects of infant's development and HIV-infected infants are vulnerable to stigmatised care [37, 40] strategies are needed for improving access to antenatal and other medical and social care for hard-to-reach pregnant women including IDUs. Furthermore, in our young (median age of 23 months) cohort of infected children, at least 4% had already experienced maternal death, highlighting a need for alternative social care for orphans and respite care for children with sick parents.

Unsurprisingly, PMTCT prophylaxis coverage among the mother-child pairs in this sub-cohort of infected children was lower than that seen for the whole cohort, which had 93% coverage in 2006 [4]; however, many of the children here were infected despite receipt of PMTCT prophylaxis, largely sdNVP or short-course ZDV, which are less effective than HAART. The current PMTCT policy in Ukraine recommends use of HAART as PMTCT prophylaxis as well as for maternal treatment, which is likely to result in a lower national MTCT rate, although this should be considered in the context of annually increasing numbers of HIV-infected women delivering in Ukraine[3]. MTCT rates remain elevated in some groups of women, particularly IDUs. One-third of mothers here reported an IDU history, higher than that seen in the ECS overall (around a fifth), reflecting the increased risk of intrapartum diagnosis and thus of MTCT among IDUs [4]. In order to achieve further and sustainable declines in MTCT and to optimise paediatric HIV treatment and care, it will be essential to implement non-stigmatised multidisciplinary services and continue to improve the coverage and quality of PMTCT interventions.

There are several limitations to this observational study, which has potential for both measured and unmeasured confounding. As an observational study in a real-life setting, there were some missing variables for some of the children, including CD4 and VL measurements. Information on adherence to therapy and viral resistance mutations was not collected, and thus we were unable to adjust for these potential confounders. The short median duration of HAART by the time of this analysis (6.6 months) is a further limitation; future analyses of data from this cohort will allow the longer-term impact of HAART to be addressed in addition to the impact of HAART started at younger ages and before progression to symptomatic disease. Follow-up of infected children from birth prevents under-estimation of mortality that can be a problem in non-birth cohorts and we expect our results to be broadly generalisable to the HIV-infected paediatric population in Ukraine.

In conclusion, in this large birth cohort of vertically HIV-infected children in Ukraine, we have documented in a "real-life", operational setting that these children have a good response to HAART, which is being used earlier and more widely, reflecting the ongoing scale-up of early infant diagnosis and paediatric treatment nationally. The improvements in survival in more recent years of the study reflect these trends. Further research is needed to explore the issues of adherence, long term exposure to HAART and subsequent adverse effects in this population.

The ECS was a coordination action of the European Commission (PENTA/ECS 018865). Claire Thorne is supported by a Wellcome Trust Research Career Development Fellowship. Some of this work was undertaken at GOSH/UCL Institute of Child Health which received a proportion of funding from the UK Department of Health's NIHR Biomedical Research Centres funding scheme. The Centre for Paediatric Epidemiology and Biostatistics also benefits from funding support from the Medical Research Council in its capacity as the MRC Centre of Epidemiology for Child Health.