Delaying ART in childhood can reduce long-term CD4 count in adulthood

Polly Clayden, HIV i-Base

The decision to start ART in children is made with guidance based on age and CD4 percentage or count. Guideline recommendations are based on observed short-term risk of morbidity and mortality. ART can be delayed in children with CD4 values above the recommended thresholds for initiation to avoid toxicities, resistance and some of the practical considerations associated with giving ART to children.

Investigators from ICH and the PENTA group suggest that current guidance assumes such a delay in treatment initiation is without detrimental long-term consequences. In a paper published ahead of print in JID, 28 December they write that evidence suggests differences between children and adults in the level of T-cell repopulation due to children’s greater thymic activity. A number of paediatric studies show poorer recovery of CD4 count on ART is associated with older age and lower CD4 count at initiation. Using longitudinal data from the PENTA 5 study and non-linear mixed-effects models, the group investigated the relationships between age, CD4 count at start of treatment, and CD4 repopulation. As well as confirming the associations previously described, their findings illustrate the importance of the naïve subpopulation for this recovery and they explore the consequences for ART naïve children of different age groups and with different CD4 counts.

The PENTA 5 trial assessed different ARV regimens in perinatally infected, treatment-naïve children. Among the 127 children starting treatment, the median age at initiation was 5.3 (IQR 2.4 to 8.6) years; CD4 count was 620 (IQR 343 to 912) cells/mm3; z-score (indicating the rank of a recorded CD4 count within the expected distribution for HIV-negative children of the same age, born to HIV-positive mothers expressed in terms of the standard, normal distribution) was -2.3 (IQR -4.1 to -1.3) and follow-up was 5.7 (IQR 5.1 to 6.5) years.

In a multivariate model the investigators estimated the children’s pre-treatment z-score to be -0.41+ 0.07 (point estimate ±SE) lower for each year older at initiation and their long term z-score -0.5+ 0.03 lower for each year older at initiation, both p<0.001. In addition to these effects, there was a strong positive association (p<0.001) between pre-treatment and long-term z-score – that is, children with z-scores below (or above) average for their age before treatment still had below (or above)-average scores in the long term.

Naïve and memory CD4 counts were recorded in a substudy of 26 children. This analysis revealed T-cell reconstitution in these children appeared to arise mainly from the naïve compartment with a comparatively small increase memory cell count, although on a faster timescale. However this potential for recovery via the naïve pool is apparently progressively reduced with age and/or duration of infection. The model illustrated suggests that the threshold currently recommended for initiating treatment in younger children results in a higher count in the long term than that for older children. Therefore guidelines for older children may not be optimal for maintaining CD4 counts in adulthood.


Lewis J et al. Age and CD4 count at initiation of antiretroviral therapy in HIV-infected children: effects on long-term T-cell reconstitution. JID. Published ahead of print 28 December 2011.

Links to other websites are current at date of posting but not maintained.