Failure of Cytomegalovirus-Specific CD8⁺ T Cell Levels at Viral DNAemia Onset to Predict the Eventual Need for Preemptive Antiviral Therapy in Allogeneic Hematopoietic Stem Cell Transplant Recipients

To the Editor—A growing body of evidence suggests that monitoring of cytomegalovirus (CMV)-specific T-cell responses may be useful for refining current preemptive antiviral therapy (PAT) strategies in patients undergoing allogeneic hematopoietic stem cell transplantation (allo-HSCT) [1, 2]. To our knowledge, only 2 intervention studies supporting this assumption have been published [3, 4]. Avetisyan et al [4] showed that PAT could be safely spared in patients without severe graft-vs-host disease (with documented CMV-specific immunity). Supporting the feasibility of this approach, El Haddad et al [5] reported in 2018 that the enzyme-linked immunospot CMV assay (T-SPOT.CMV assay; Oxford Diagnostic Laboratories) may predict (in 94% of patients) progression from a low-level CMV DNAemia state to a high-level state requiring initiation of PAT; nevertheless, the precise timing at which immunological monitoring was first conducted relative to the onset of CMV DNAemia (first positive polymerase chain reaction result) was not clearly described.

Our experience, reported herein, casts some doubt on the ability of CMV-specific immunological surveillance at the time of viral DNAemia detection to reliably anticipate the need for PAT. In this prospective study, we included 30 episodes of CMV DNAemia in 24 nonconsecutive adult patients who underwent T-cell–replete allo-HSCT at our center (Table 1). CMV pp65- and IE1-specific interferon (IFN) γ–producing and CD69⁺-expressing CD8⁺ T cells, which confer protection against active CMV infection in this setting [1, 5, 6], were enumerated by flow cytometry for intracellular cytokine staining (ICS) [6] within 48 hours of CMV DNAemia detection, using the RealTime CMV polymerase chain reaction assay (Abbott Molecular) [6, 7]. PAT was initiated when the plasma CMV DNA load reached levels of ≥1500 IU/mL or featured a doubling time of ≤2 days [7].

Initial (n = 9) and recurrent (n = 21) episodes of CMV DNAemia occurred a median of 34 (range, 2 to 357) and 172 (32 to 1251) days after transplantation, respectively. Five initial and 11 recurrent episodes required PAT. In 18 of 30 episodes of CMV DNAemia, corticosteroid treatment of acute or chronic graft-vs-host disease had been started at the time of immunological analyses. Only 2 of the 18 patients were receiving corticosteroids at high doses. No patient in this series had CMV end-organ disease. Median CMV-specific CD8⁺ T cell counts were comparable, irrespective of whether or not PAT was subsequently initiated according to protocol: for initial episodes, these counts were 1.42/µL (range, 0.69–2.42/µL) in treated episodes (n = 5) and 0.29/µL (0–1.13/µL) in self-resolving episodes (n = 4) (P = .06; Mann-Whitney U test); for recurrences they were 5.70/µL (range, 0–112/µL) and 3.30/µL (range, 0.16–93/µL) for treated and untreated episodes, respectively (P = .40; Mann-Whitney U test). Overall, there was no correlation (ρ = 0.24; P = .21) between CMV-specific CD8⁺ T cell counts at CMV DNAemia onset and CMV DNA peak loads reached within episodes. No reliable CMV-specific CD8⁺ T-cell count cutoff was found to discriminate between patients who did or did not subsequently undergo PAT, irrespective of whether or not patients were receiving corticosteroid therapy.

In addition to the presumed variations in timing of immunological monitoring relative to the onset of CMV DNAemia, there are a number of dissimilarities between the studies that may account for their apparent discrepant conclusions. First, patients undergoing allo-HSCT with high-risk modalities [1] were excluded from the series of El Haddad et al [5], but not from ours (n = 5). Second, only initial episodes were considered in the aforementioned study [5], whereas in ours two-thirds of episodes were recurrences. Unfortunately, the small sample size in our series precluded the performance of subanalyses stratified by allo-HSCT modality and CMV DNAemia episode type. Third, the T-SPOT.CMV assay quantifies pp65 and IE-1 IFN-γ–secreting CD4⁺ and CD8⁺ T cells individually, whereas our ICS method simultaneously enumerates pp65- and IE-1–activated IFN-γ–producing CD8⁺ T cells. To our knowledge, no straight comparison between results produced by the T-SPOT.CMV and ICS assays has been performed in the allo-HSCT setting; nevertheless, values provided by both assays seem to correlate reasonably well in samples from deceased organ donors [8]. The study by El Haddad et al [5] unequivocally proved that the lack of expansion of CMV-specific T cells, commonly seen in patients undergoing corticosteroid therapy, favors progression to high-level CMV DNAemia, which confirms data previously published by our group [3, 9, 10].

In summary, immunological monitoring at CMV DNAemia onset was not useful in predicting the eventual need for PAT, regardless of corticosteroid use. The possibility that the usefulness of immunological surveillance in guiding PAT is dependent on different patient characteristics and the assay used cannot be ruled out.

Notes

Financial support. This work was supported by CARLOS III Health Institute (grants FIS 12/1992 and 15/0060). E.G. holds A RÍO Hortega Research Contract (CM16/00200) from CARLOS III Health Institute. E.M is grateful to Ministry of Economy and Competitiveness (MINECO, Spanish Government) for a postdoctoral contract “Juan de la Cierva” (Ref. FJCI-2015-25992).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References