Virologic Suppression in Nevirapine-Exposed HIV-Infected Infants Initiating Antiretroviral Therapy in Rural Uganda

Abstract

We measured virologic suppression among 34 nevirapine (NVP)-exposed HIV-infected children with median age of 8.6 months (range: 3.2–19.9) initiating NVP-based antiretroviral therapy (ART) in rural Uganda. In Kaplan–Meier analysis, the cumulative probability of virologic suppression, defined as having two consecutive HIV-1 RNA <400 copies ml⁻¹ by 18 months was 56%. In multivariate Cox proportional hazard modeling, the following pre-ART measurements were independently associated with an increased probability of viral suppression: increasing age [hazard ratio (HR) =1.28 per 1 month increase in age, p = 0.002], lower viral load (HR = 3.54 for HIV RNA > 7 50 000 copies ml⁻¹, p = 0.03) and high CD4% (HR = 6.0 for CD4% > 25, p = 0.003). These results lend additional support to the 2010 World Health Organization recommendations that protease inhibitors be used to treat NVP-exposed children, but that NVP-based ART should be initiated before the decline of CD4% to optimize outcomes in NVP-exposed children when protease inhibitors are not available.

Background

Clinical trials from urban African settings demonstrate that antiretroviral therapy (ART) should be initiated early to prevent mortality in HIV-infected infants [1] and that HIV-protease inhibitor (PI) regimens are superior to nevirapine (NVP) regimens for the treatment of infants who have had perinatal NVP exposure [2–5]. As a result, in 2010 the World Health Organization (WHO) recommended universal ART for all HIV-infected children <24 months old [6]. WHO guidelines also specify that PI-based regimens should be used when either infant or mother had perinatal exposure to NVP, but with the caveat that ‘where PIs are not available, affordable or feasible, NVP-based therapy is recommended’ [6]. Recognizing that access to PIs remains limited in many southern African countries, we sought to identify predictors of virologic suppression among a cohort of HIV-infected, NVP-exposed infants who received NVP-based ART regimens in rural Uganda before the change in guidelines and the availability of PIs.

Methods

Between August 2007 and April 2008, 351 children were enrolled in the Tororo Child Cohort (TCC) Study, an ongoing longitudinal study investigating interactions between HIV and malaria in children living in Tororo, Uganda [7]. Children were enrolled between the ages of 6 weeks and 12 months and followed until 5 years of age. Of these, 57 were determined to be HIV-infected using DNA PCR (Amplicor HIV-1 DNA PCR Test, version 1.5; Roche, Branchburg) on dried blood spots (DBSs), and confirmed by plasma HIV RNA level (dynamic range 40–10 000 000 copies ml⁻¹, COBAS TaqMan HIV-1). DBSs were tested within 1 week of collection. For children with indeterminate results, additional DBSs were obtained and tested within 1 week. When in transit or storage, DBSs were kept in paper bags at room temperature. For this analysis, we included only the 34 HIV-infected children who had been exposed to NVP in the perinatal period, had initiated ART and had at least one HIV RNA level measured after ART initiation. NVP exposure was defined as the mother, infant or both receiving single dose NVP around the time of delivery. At the time of this study, extended dosing of NVP was not utilized. Clinical assessments, including WHO stage and weight-for-age Z-score, were performed at enrollment and monthly thereafter [8]. The parent or guardian of each child provided consent for enrollment into this study.

The CD4 cell count and percentage were measured every 12 weeks by fluorescence-activated cell sorting (Becton-Dickinson, San Jose, CA, USA). ART was initiated in all eligible children. At the beginning of the study, Ugandan guidelines matched the 2006 WHO guidelines, recommending ART for children with any clinical stage IV event, most clinical stage III events or CD4 percentages below age-specific cutoffs (<25% for children <12 months of age or <20% for children aged 12–35 months). In 2008, Uganda adopted interim WHO recommendations that all HIV-infected children under age 12 months receive ART, regardless of CD4 status; three children who had not previously met immunologic or clinical criteria were then initiated on ART [9]. The 2008 WHO interim guidelines also suggested that PIs be used to treat NVP-exposed infants when possible, but PIs were not available. Most ART-eligible HIV-infected children received zidovudine (AZT), lamivudine (3TC) and NVP. If a child had anemia at baseline, stavudine (d4T) was used in place of AZT. Nevirapine was dosed using body surface area at 160–200 mg m−² (max 200 mg) per dose twice daily or by WHO weight band guidelines; a 14 day induction of once daily dosing was utilized [8]. Infants initially received syrup formulations and transitioned to tablet based on tolerance and parental preference. Generic or brand name formulations were utilized based on availability. Fixed dose combination tablets including Trioimmune (3TC + d4T + NVP) and Combivir (3TC + AZT) were utilized in some older children. Patients were seen 2 and 4 weeks after ART initiation to assess adherence and tolerability to study medications, and monthly at a devoted study clinic thereafter. Adherence was assessed by parental report of a number of missed doses; pill counts were performed when providers had specific concerns. Six patients subsequently changed from AZT to d4T during the study because of anemia, per Ugandan guidelines. In accordance with the WHO 2006 guidelines, clinical and immunologic treatment failure was defined by new CD4 decline to below threshold used for treatment initiation (CD4% 25 for children <12 months of age or <20 for children aged 12–35 months) or new clinical stage progressing event. Plasma HIV RNA levels were measured every 12 weeks at the CDC laboratory in Entebbe, Uganda, using either the COBAS TaqMan HIV-1 (range of detection 40–10 000 000 copies ml⁻¹) or the Cobas Amplicor HIV-1 RNA Monitor Test (range of detection 400–1 000 000, version 1.5; Roche, Branchburg). As children were followed in the parent study prior to ART initiation, the timing of 12 week interval laboratory assessments was not related to the date of ART initiation.

The primary end point for this analysis was virologic suppression, defined as the first of two consecutive measurements of plasma HIV RNA levels under 400 copies ml⁻¹. Survival analysis was used to estimate the cumulative probability of achieving viral suppression using Kaplan–Meier (K–M) product limit formula within 18 months of follow-up. Rates of suppression by 18 months were compared by log rank test. Uni- and multivariate Cox proportional hazards modeling was used to evaluate the age at ART initiation, baseline HIV RNA above and below the median, baseline CD4%, breastfeeding at ART initiation, gender, WHO stage at initiation and being underweight (weight-for-age Z ≤ −2) as predictors of viral suppression. We then performed a multivariable analysis that included all variables. Our final multivariable model included only predictors that were statistically significant after backward selection, using a p-value of 0.05. CD4 recovery was assessed by calculating the median change in CD4% from baseline to at least 18 months after ART initiation for each patient. We then compared the overall median change among children who achieved virologic suppression with the median change among children who did not, using the Wilcoxon Test to evaluate the significance of the difference between groups. All analyses were performed using STATA version 10.0 (STATACorp, 2003, Stata Statistical Software, Release 8, College Station, TX, USA; StataCorp LP).

Results

Among 57 HIV-infected children in the cohort, 52 initiated ART (Fig. 1). Three of the 52 children were excluded because they did not have documented RNA measurements following ART initiation: two had died and one had moved away from the study location. Of the remaining 49 children, 34 had perinatal exposure to NVP and were included in this analysis (Table 1). Of those, 31 had two or more HIV RNA levels following ART initiation. A total of 19 children (56%) achieved virologic suppression within 18 months of starting ART. In K–M analysis, the probability of virologic suppression at 6, 12 and 18 months was 45, 58 and 58%, respectively. The median (range) time to virologic suppression was 3.7 (0.9–8.0) months.

Fig. 1.

Enrollment of participants.

Table 1.

HIV-infected infant characteristics at the time of ART initiation (n = 34)

Characteristic	Value
Age in months, median (range)	8.6 (3.2–19.9)
Age ≥12 months, n (%)	4 (12)
Female, n (%)	16 (47)
Breastfeeding, n (%)	29 (85)
Underweight, n (%)a	11 (32)
Pre-ART CD4 percentage, median (range)	19 (9–36)
Pre-ART CD4 <25%, n (%)	28 (82)
Pre-ART log10 HIV-RNA level, median (range)b	5.8 (3.5–7.0)
Pre-ART HIV-RNA level ≤750,000 copies/ml, n (%)b	16 (50)
Pre-ART WHO Stage, n (%)
I	27 (79)
II	3 (9)
III	4 (12)
IV	0 (0)
Nevirapine exposurec, n (%)
Both mother and infant	24 (71)
Mother only	6 (18)
Infant only	4 (12)

^aWeight-for-age Z < −2.

^bTwo study participants did not have a pre-ART viral load.

^cReport of single dose Nevirapine.

Higher age, HIV-RNA level <750 000 copies ml⁻¹, and CD4% ≥25 at the time of ART initiation were significant predictors of virologic suppression in multivariate Cox proportional hazards modeling (p = 0.002, p = 0.03, p = 0.003, respectively) (Table 2). Gender, WHO stage at ART initiation, not breastfeeding at time of ART initiation and being underweight were not statistically significant predictors. All children who initiated ART ≥12 months of age (4 of 4), age <12 months with CD4% ≥25 (5 of 5) or age <12 months with CD4% <25 or with HIV RNA level <7 50 000 copies ml⁻¹ (4 of 4) achieved virologic suppression versus only 31% of those with age <12 months, CD4% <25 and HIV RNA>7 50 000 copies ml⁻¹ (6 of 18). Of the two children who initiated D4T/AZT/NVP, one achieved viral suppression and the other completed 18 months of follow-up without suppression. Among the four children who changed from AZT to D4T because of toxicity, two achieved virologic suppression, one died and one completed 18 months of follow-up without suppression.

Table 2.

Predictors of time to virologic suppression by 18 months of follow-up

Predictor variable	Bivariate analysis		Multivariable analysis
	HR (95% CI)	p-value	HR (95% CI)	p-value
Age at ART initiation (per 1 month increase)	1.29 (1.14–1.47)	<0.001	1.28 (1.10–1.49)	0.002
Pre-ART HIV RNA <7 50 000 copies ml−1	4.95 (1.74–14.1)	0.003	3.54 (1.13–11.1)	0.03
Pre-ART CD4 ≥25%	5.00 (1.75–14.3)	0.003	6.00 (1.83–19.6)	0.003
Not breastfeeding at time of ART initiation	3.23 (1.29–8.09)	0.01	–	–
Female gender	1.55 (0.62–3.86)	0.35	–	–
Pre-ART WHO stage I or II	1.58 (0.46–5.45)	0.47	–	–
Not wasted at time of ART initiation	2.64 (0.87–8.00)	0.09	–	–

Among the 29 children with CD4% information at baseline and after 18 months of ART, the median CD4% recovery for those with virologic suppression (n = 17) was significantly higher than that for those without virologic suppression (n = 12) [22 (range 12–38) vs. 18 (range −3 to +33), p = 0.02]. Even among children without virologic suppression, the majority (11 of 12) demonstrated increases in CD4%.

During the study period, none of the children met immunologic or clinical criteria per WHO criteria to switch to a second-line treatment, which are decline of CD4 measures to treatment thresholds after initial recovery or development of new WHO stage III or IV condition.

Discussion

Only 56% of 34 NVP-exposed infants initiating NVP-based ART in rural Uganda achieved virologic suppression. Poor virologic response of NVP-exposed African infants treated with NVP-based regimens has been demonstrated in other studies [1–4]. It is likely that infants did not achieve virologic suppression because of NVP-associated resistance mutations. The pattern of HIV RNA levels in 9 of the 15 of children who did not suppress were consistent with the existence of pre-existing drug resistance, with an initial two log decline within the first 6 months of treatment, but immediate rebound to pre-treatment levels before 12 months (data not shown). Several studies have shown that high rates of K103N and other NVP-associated mutations are present among infants that become HIV-infected after perinatal NVP exposure [4, 10, 11]. Infants could have obtained NVP-associated mutations by direct maternal transmission at birth or during breastfeeding [12] or because of non-suppressive exposure to NVP at the time of delivery. Alternatively, treatment failure and resistance mutations may also have developed because of poor drug levels and difficulties in adherence in young infants, who commonly regurgitate their medications [13], during this treatment course.

Age was a powerful predictor of virologic suppression within the cohort. While the number of children older than 12 months was small (n = 4), improved outcomes in NVP-exposed patients initiating NVP-based ART after longer periods of time from exposure have been noted in other studies. Mothers who start NVP-based combination therapy >6 months after perinatal exposure to NVP have high rates of virologic suppression compared with those who start sooner [4, 14]. One Ugandan study of NVP-exposed Ugandan infants found that those who initiated ART at 6–12 months of age had lower virologic suppression rates than those who initiated at >12 months of age [15]. The authors hypothesized that this could reflect overgrowth of wild-type HIV without resistance mutations. However, data from adults suggest that resistance mutations remain 'archived' and typically re-emerge when these agents are reintroduced [16]. There are other possible explanations for the improved outcomes in older children. Younger infants could have been infected at birth and experienced NVP exposure while the older could have become infected by breastfeeding several months after birth and NVP exposure, when the mother's virus may have reverted to wild-type. Young age was also shown to be a predictor of lower rates of virologic suppression in a study of HIV-infected European infants [17]. Notably, we also found a low rate of virologic suppression (6 of 15) among the infants who did not have a history of perinatal exposure to NVP and initiated ART in the same study clinic (data not shown). Young infants have distinct pharmacologic issues that could influence plasma levels of antiretroviral medications, including adherence problems, malabsorbtion and variations in pharmacokinetics resulting from developmental changes in CYP enzyme activity.

Our study had several limitations. First, the study population is vulnerable to ‘survival’ bias, because we enrolled older children who had survived long enough to defer treatment to a later age. Second, because this was an observational study, there may have been unknown confounders for which we did not adjust. Third, given the small sample size, we had limited power to detect the influence of all predictor variables tested. Further study including measurements of resistance mutations, pharmacokinetic studies and more detail about medication adherence could shed further light on why these children fail.

We also found that a CD4% >25% and plasma HIV RNA <7 50 000 c ml⁻¹ at baseline were strong predictors of virologic response. The number of children in these subgroups was small, but these findings suggest that infants who initiate ART early, before significant decline in CD4%, will have improved outcomes. Close follow-up of such children will be important, as it is possible that archived mutations will lead to late treatment failure. We were also encouraged to find that while children without virologic suppression had less recovery of CD4% compared with those with suppression, the majority of children (11 of 12) demonstrated some degree of CD4% recovery.

In summary, the overall poor rate of virologic suppression among NVP-exposed infants in this study underscores the urgent need for increased access to PIs in resource-limited settings for the treatment of NVP-exposed infants and children. However, our findings also reinforce the 2010 WHO guidelines for the use of NVP-based ART when PIs are not available, and show that virologic suppression can occur in many children after perinatal NVP exposure [6]. There is a critical need for increased access to early infant HIV diagnosis, affordable infant formulations of PIs, and further research to improve outcomes in HIV-infected African infants and children.

Funding

This work was also supported by the National Institute of Child Health and Human Development, National Institute of Health (K23-60459-01A2). Support was also provided by the Doris Duke Charitable Foundation (GD is a recipient of the Clinical Scientist Development Award).