Meta-analysis shows strong association of HHV-6B reactivation and subsequent aGVHD

The significance of the association between post-HCT HHV-6B reactivation and grade 2-4 acute graft-versus-host disease (aGVHD) has been hotly debated for years due to seemingly conflicting results from many studies employing differential methodologies. However, a recently published meta-analysis has confirmed the statistical significance of this association; moreover, it showed that patients who reactivate HHV-6 after alloHCT are 2-3 times more likely to develop subsequent grade 2-4 aGVHD. For comparison, patients who reactivate EBV and CMV are only 1.24 times and 1.09 times more likely to develop aGVHD, respectively (Styczynski 2016 and Green 2015).

Danielle Zerr, MD, MPH, Children’s Hospital, Seattle, was the senior author of the meta-analysis.

A team, led by Dr. Danielle Zerr, sought to systematically summarize and analyze evidence regarding HHV-6B reactivation and aGVHD. One of the biggest issues concerning past investigations of this topic has been heterogeneous methodologies. For example, some studies would use antibody titers to diagnose HHV-6, whereas others would use PCR; some would test all patients for HHV-6 regardless of symptoms, whereas others would only test symptomatic patients. Different diagnostic methods will have variable sensitivity/specificity and not systematically testing all patients in a cohort introduces clinical bias; both of these situations will confound aggregate analyses. As such, a strength of this meta-analysis is that all included studies used fairly homogenous methods: every patient, regardless of symptoms, was regularly tested for HHV-6 post-HCT by amplifying DNA isolated from blood samples using PCR.

Another strength of this meta-analysis is that the included studies were stratified into 3 different groups of studies based on the temporality of their analysis: HHV-6 reactivation was analyzed as a risk factor for subsequent aGVHD (group 1), aGVHD was analyzed as a risk factor for subsequent HHV-6 reactivation (group 2), and analysis was performed without temporal specification (group 3). Ten of the 14 studies in group 1, which included 1,773 total patients, reported that HHV-6B significantly predicted the incidence of grade 2-4 aGVHD and meta-analysis of the 11 group 1 studies, that provided adequate data, demonstrated a 2.65-fold increased risk in aGVHD when patients reactivated HHV-6B (95% CI 1.89-3.72, p<0.001). The risk (OR 2.41, [95% CI 1.70-3.43], p<0.001) was similar when meta-analysis was performed using the 12 studies (out of 18 total studies) in group 3 that provided adequate data. It should be noted that studies in group 1 were typically the most recently published, included higher number of patients, and serve as the most rigorous analysis of the 3 groups due to the additional temporal component in the statistical analysis. However, despite the lack of temporal analysis in the group 3 studies, it is telling that the association between HHV-6B and aGVHD is still statistically significant. That the association persists regardless of variations in statistical analysis indicates that there is likely something going on that we have yet to uncover, though some researchers have proposed possible models of pathogenesis (Phan 2018b).

Although this meta-analysis is the most complete analysis of HHV-6 reactivation and aGVHD to date, there are still some limitations that should be considered. First, the authors reported some evidence of publication bias, based on their funnel plot results. Secondly, the methods and outcome definitions of the included studies were only homogenous, albeit not identical. Finally, this paper merely establishes a significant correlation between HHV-6B reactivation and grade 2-4 aGVHD and does not serve as definitive proof, per se, of a causal role for HHV-6B. Further studies characterizing mechanisms of pathogenesis as well as randomized, interventional studies are needed to confirm causality.

Read the paper here: Phan 2018