The novel work by Herne et al1 reported very high seroprevalence of cytomegalovirus (CMV) (97.4%) in patients with mycosis fungoides (MF) when compared with matched bone marrow donors (57.3% seropositivity) in a teaching hospital (University of Texas, M. D. Anderson Cancer Center, Houston, TX) in the United States.
MF and the closely related Sézary syndrome (SS) comprise the most common variant of primary cutaneous T-cell lymphomas (CTCLs) and a common form of extranodal non-Hodgkin lymphoma. MF and SS present with cutaneous involvement characterized by patches, plaques, and tumors, along with erythroderma in SS. Histologically, MF and SS manifest pronounced epidermotropism with clonal CD4+/CD45RO+ tumor cells. The etiology is unknown. Other viruses, such as HTLV-1, can produce clinical and histologic features that mimic MF/SS as part of adult T-cell leukemia and lymphoma (ATLL), but studies have not shown clear evidence for an association between HTLV-1 and MF/SS.2,3
Infection with CMV is by intimate contact and is usually subclinical. Seroprevalence increases with age, and rates of CMV seropositivity vary across adult populations, from around 50% to 60% in developed nations4-6 to 95% in developing nations.4,7 CMV is not known to be oncogenic.
Given this intriguing association between MF and CMV, we sought to reproduce Herne et al's finding by undertaking a retrospective blinded case-control study to compare the CMV seroprevalence rate in MF/SS patients compared with other patients attending a regional center for CTCL. The study had a power of 88% to detect a difference in prevalence of 30% between the 2 groups (60% in controls and 90% in MF/SS patients).
Non-MF/SS patients attending the clinic served as controls. We tested stored serum samples from these patients for CMV IgG antibodies using two commercially available systems: one automated (Abbott Laboratories, Abbott Park, IL), and the other, a manual CMV IgG enzyme-linked immunosorbent assay (ELISA; Trinity Biotech, Bray, Ireland). Statistical analysis was performed using Epi Info version 6.04 (Centers for Disease Control and Prevention, Atlanta, GA).
Age, sex, and ethnic origin were not statistically different between the MF/SS and non-MF/SS groups (Table 1). There was excellent correlation between the 2 different serologic assays. We found that 32 (60.4%) of 53 confirmed MF/SS patients were seropositive for CMV IgG, compared with 16 (61.5%) of 26 controls (P = .92). Seropositivity rates in early-stage (IA, IB, IIA) versus late-stage (IIB-IVB) MF/SS were 19 (54.3%) of 35 and 13 (72.2%) of 18, respectively (P = .24). Non-MF/SS patients had a range of diagnoses, including psoriasis, lichen simplex chronicus, dermatitis, and other cutaneous B-cell lymphomas. Median automated serology values for cases and controls revealed no significant differences.
Our findings failed to support those of Herne et al1 and showed no difference in CMV IgG seropositivity rates between MF/SS patients and other patients who attended our dermatology clinic who were similar in terms of age, sex, and race. We note that the study by Herne et al used bone marrow donors with prospective routine baseline CMV serology as controls. However, the MF/SS group was retrospectively tested for CMV IgG and may have been tested with a different assay. In our study, all samples were tested in parallel and showed almost complete concordance between the 2 assays. These conflicting findings need to be resolved by further studies.
