Late cytomegalovirus disease after hematopoietic cell transplantation: significance of novel transplantation techniques

Abstract

Preemptive therapy (PET) and letermovir prophylaxis are effective in preventing cytomegalovirus (CMV) disease within the first 100 days after allogeneic hematopoietic cell transplantation (HCT) but are associated with late-onset CMV disease. We retrospectively examined the clinical manifestations, risk factors, prevention algorithm, and outcome of late CMV disease in CMV seropositive day 100 survivors transplanted between 2001-2017 (PET cohort) and 2018-2021 (letermovir cohort). There were 203 episodes of late CMV disease among 2469 day 100 survivors, and the estimated cumulative incidence of first late CMV disease was 7.2% (95% confidence interval [CI], 6.2-8.3) with no difference between the PET (7.4%; 95% CI, 6.4-8.6) and the letermovir group (5.4%; 95% CI, 3.2-8.3). Thirty-seven patients (1.5%) had a second episode of CMV disease. In multivariable Cox regression models, posttransplant cyclophosphamide was associated with an increased risk of gastrointestinal CMV disease. CMV viremia or disease detected before day 100, corticosteroid treatment after day 100 at dose ≥1 mg/kg, acute and chronic graft-versus-host disease, lymphopenia, HLA-mismatched related donor status, were also associated with late CMV disease. HLA-mismatched donor status and late use of corticosteroids (≥1 mg/kg) were risk factors for late CMV disease recurrence. Late CMV disease occurred most frequently in a setting of prolonged low-level untreated viremia and was independently associated with death by 2 years after HCT. In summary, late CMV disease continues to occur in the present era. Improved prevention strategies for late CMV disease are needed.

Graphical abstract

Figure 1.

Cumulative incidence of late CMV disease and anatomical site distribution. (A) Cumulative incidence probability for late CMV disease in the PET (2001-2017) and LET (2018-2021) cohorts. (B-D) Cumulative incidence probability of developing a second episode of CMV disease within 1 year after the first episode of CMV disease in the PET cohort (B) and specific organ involvement in the PET (C) and LET cohorts (D).

Figure 2.

Characteristics of late CMV disease. Anatomical site distribution (A,D) and proportion of first, second, and third episodes (B,E) observed in 3-month intervals, and proportion of GI GVHD in patients with first late GI CMV disease (C,F) in the PET (A-C) and LET (D-F) cohorts. (A-D) Data presented as the number of episodes of CMV disease diagnosed in 3-month intervals; different numerical values on the y-axis were used for panels A (0-70) and D (0-10).

Figure 3.

Risk factors for late CMV disease. Forest plots presenting risk factors for first late CMV disease between day 100 and day 730 after HCT in the entire population (A) and in the PET cohort (B). Results from multivariable Cox proportional hazard models. Low-risk viremia is defined as <10 antigen-positive cells or viral load <1000 IU/mL; high-risk viremia as ≥10 antigen-positive cell or viral load ≥1000 IU/mL. cGVHD is time dependent. ∗Represents continuous variables. aGVHD, acute GVHD; ALC, absolute lymphocyte count; CNI, calcineurin inhibitor; HR, hazard ratio; MAC, myeloablative conditioning regimen; NMA/RIC, nonmyeloablative/reduced-intensity conditioning regimen; PBSC, peripheral stem cells.

Figure 4.

Risk factors for specific organ manifestations of late CMV disease. Forest plots presenting risk factors for first late CMV pneumonia (A,C) or first late GI disease (B,D) between day 100 and day 730 after HCT in day 100 survivors in the entire study population. Shown are results from different multivariable Cox proportional hazard models that used different sets of variables. Low-risk viremia is defined as <10 antigen-positive cells or viral load <1000 IU/mL and high-risk viremia as ≥10 antigen-positive cells or viral load ≥1000 IU/mL. ∗Represents continuous variables. ALC, absolute lymphocyte count; CNI, calcineurin inhibitors; HR, hazard ratio; MMF, mycophenolate mofetil; MTX, methotrexate.

Figure 5.

Main causes of virologic surveillance failures. The proportion of main causes of surveillance failures in PET (A) and LET (B) cohorts. Description of categories: no surveillance: PCR monitoring was not continued or the frequency was not as recommended; on treatment: CMV disease occurred despite antiviral treatment lasting >7 days; Ag based: patients monitored with pp65 antigenemia testing after day 100 instead of PCR testing; breakthrough: CMV diseases with undetectable viremia with adequate monitoring; threshold: detectable viremia below recommended threshold for preemptive treatment before CMV disease diagnosis, resulting in delay in antiviral treatment; positive treatment for viremia <7 days: treatment threshold was reached only within 7 days before disease onset; and no data: no information regarding PCR surveillance or start of antiviral treatment. Ag, antigen.

Figure 6.

Impact of late CMV disease on overall mortality. Overall survival after first late episode of CMV disease by 6 weeks after diagnosis (A) and adjusted risk of death due to CMV disease (B). ∗All factors included in the multivariable model are presented in supplemental Table 4.