On the Assessment of Trend in CD4⁺ Cell Count and Viral Suppression After Transition to Adult Care

Xia et al’s publication on the effects of transition from paediatric to adult care on retention in HIV care, viral suppression, CD4⁺ cell count, and mortality among individuals with perinatal HIV infection in New York City offers insight into an interesting and increasingly important topic in HIV clinical care and research.¹ Their primary conclusions were that rates of viral suppression and the number of individuals with a CD4⁺ cell count ≥500 cells/μl significantly increased after transition to adult care. However, their other primary conclusion was that one-year mortality was significantly higher post-transition. These conclusions from the same study are not logically consistent. If viral suppression and CD4⁺ cell counts are increasing, mortality should be decreasing. These discordant findings could be a result of selection bias due to the authors’ follow-up definition.

For the assessment of the retention in care, viral suppression, and CD4⁺ cell count outcomes, a trend analysis using Generalized Estimating Equations (GEE) was utilized. Individuals were included if they transitioned from paediatric to adult care during 2006–2015 and had at least 1 year of pre-transition follow-up. The post-transition follow-up continued for 3 years after transition, until death, or until the end of the study period on December 31, 2016, whichever occurred first. The trend analysis limited assessment of CD4⁺ cell count and viral suppression to those that completed 3 years of follow-up and did not die during the study. Therefore, the analysis excluded anyone who died during the study or transitioned after December 31, 2013, as those would be unable to contribute 3 years of follow-up. Of the 735 individuals who transitioned during 2006–2015 and had at least one year of pre-transition follow-up, 76 individuals died overall and 41 died within the first year after transition. Among the 41 who died in the first year after transition, 34 (82.9%) died from HIV-related diseases.

There are several concerns with the design of the study to assess the retention in care, CD4⁺ count, and viral suppression post-transition outcomes. First, if the 76 individuals died during the study had lower rates of viral suppression and lower CD4⁺ counts compared to those who survived three years of post-transition follow-up, the exclusion of these individuals from the post-transition analyses could lead to positively biased inferences, a result of the healthy survivor effect. This bias is likely, as those who died in the first-year post transition had lower rates of pre-transition viral suppression (12.2 vs. 47.8%) and lower proportion of higher CD4⁺counts (2.4 vs. 37.2%) compared to those who survived. Similarly, those that transitioned later in calendar time but were excluded from the analysis may have had higher rates of viral suppression and higher CD4⁺ cell counts as ART prescription and viral suppression have been shown to improve over time in the United States.^2,3 Exclusion of these individuals could lead to negatively biased inferences.

Additionally, the authors do not describe how missing laboratory results were handled in the study. Retention in care was defined by the presence of ≥1 lab value during the year of interest. Those who were not retained in care were still included in the trend analysis for viral suppression and CD4⁺, but if lab values were missing, it is unclear how viral suppression and CD4⁺ outcomes were determined. One option would be to impute the missing values. A second option would be to carry the last observation forward, although this strategy would likely overestimate viral suppression and CD4⁺ counts, as those who are lost to care are more likely to have unsuppressed viral loads and lower CD4⁺ counts. The third option would be to assume that those with missing values had unsuppressed viral loads and low CD4⁺ counts; however, this strategy may underestimate viral suppression and CD4⁺ counts if these individuals had transferred and were not actually lost to care.

The transition from paediatric to adult HIV care is increasingly important, and these findings suggest that transition to adult care increases viral suppression and CD4⁺ cell counts, which would imply that this transition leads to improved HIV outcomes. However, if its inferences are biased, this study could lead to an underestimation of the risks inherent in this important transition. Studies addressing this topic should take care in defining their exclusion criteria (which may induce selection biases) and explaining their handling of missing data, in order to accurately characterize this critical transition.