Year: 2018
Type: letter
Abstract: To the Editor—Giménez and colleagues present a retrospective study of 196 allogeneic hematopoietic cell transplant (HCT) recipients in which they analyzed the association of plasma cytomegalovirus (CMV) load area under the curve (AUC) with overall mortality by 180 and 365 days after HCT. They report no evidence of a dose–response relationship between the CMV AUC and mortality, contrary to our recent publication [1]. The authors note differences between our studies that may contribute to the disparate findings and specifically inquire whether a difference in the AUC calculation played a role. Both studies used the curve-trapezoid rule method to calculate the AUC. Whereas Giménez and colleagues calculated the viral load AUC prior to log10-transformation, we log10-transformed the data prior to calculating the AUC and standardized for time at risk by dividing the AUC by the number of days followed to the time of the AUC [2]. Our results remained significant without standardizing by days followed (data not shown). The log10 value of the AUC using crude viral loads emphasizes peak viral loads, whereas the AUC of log10-transformed viral loads captures cumulative exposure to clinically relevant levels of viral replication. Both approaches are valid metrics that provide alternative biomarkers for study. We recalculated the CMV AUC from our study using the approach described by Giménez and colleagues. In a Cox model, a CMV AUC >5 log10 copies × days × mL-1 was associated with an increased hazard for overall mortality by 100 and 365 days after HCT (Table 1), suggesting a threshold effect. The data presented by Giménez and colleagues also suggest a threshold effect and possibly a dose–response relationship, given increasing hazards for overall mortality with higher CMV AUC. However, the width of their confidence intervals implies additional data are needed for a more precise inference. Other important differences in their study include less frequent testing (affecting AUC precision) and a higher preemptive therapy threshold (resulting in higher median viral loads and fewer patients receiving treatment at lower AUCs [Table 1 footnote]). Cox Models of the Association of the Cumulative Cytomegalovirus Viral Load Area Under the Curve (AUC) With Overall Mortality by 100 and 365 Days After Allogeneic Hematopoietic Cell Transplantation Using an Alternative Approach to the AUC Calculation Our cohort included 404 patients. The median maximum CMV DNA log10 AUC among patients with CMV reactivation was 3.95 log10 copies × days × mL-1 (range, 1.59–6.12) in the entire cohort from day 0 to 100 (n = 260) and the same among day 100 survivors (n = 240). Cytomegalovirus results after day 100 were not included. There were 125 deaths (32%) by day 365 after HCT. Thirty-six patients (9%) died in days 0–100, and 89 patients (23%) died in days 101–365. Preemptive antiviral therapy was initiated once the viral load reached 500 copies/mL (125 IU/mL) for standard-risk patients. Patients considered high-risk, such as those receiving ≥1 mg/kg body weight of prednisone or cord blood transplant recipients, were started at any positive viral load. Patients were screened for CMV weekly through day 100, except cord blood HCT recipients who were screened for CMV twice weekly. As of each patient’s final AUC calculation, the fraction of patients receiving antiviral therapy in each of the 5 viral load categories (from lowest to highest) was 16 of 147 (11%), 21 of 34 (62%), 85 of 100 (85%), 100 of 103 (97%), and 19 of 20 (95%). Abbreviations: aHR, adjusted hazard ratio; AUC, area under the curve; CI, confidence interval; CMV, cytomegalovirus; HCT, hematopoietic cell transplantation; HR, hazard ratio. aThis model was adjusted for age, HCT comorbidity index score, underlying disease risk, type of conditioning regimen (myeloablative vs not), time-dependent acute graft-versus-host disease grades 3–4, and time-dependent cumulative steroid dose AUC during the first 100 days after HCT (steroid use was categorized as 0, <1 mg/kg, 1 to <2 mg/kg, and ≥2 mg/kg prednisone equivalent per day). bThe plasma CMV DNA log10 viral load AUC (as copies × days × mL-1) was calculated using the curve-trapezoid rule. The AUC was calculated first and then log10-transformed. Patients with no CMV were assigned an AUC of 0. The CMV DNA log10 viral load AUC was incorporated as a time-dependent variable through day 100 and a time-invariant variable through day 365, using the AUC value at day 100 as the new baseline value (because routine testing ended at day 100). Cox Models of the Association of the Cumulative Cytomegalovirus Viral Load Area Under the Curve (AUC) With Overall Mortality by 100 and 365 Days After Allogeneic Hematopoietic Cell Transplantation Using an Alternative Approach to the AUC Calculation Our cohort included 404 patients. The median maximum CMV DNA log10 AUC among patients with CMV reactivation was 3.95 log10 copies × days × mL-1 (range, 1.59–6.12) in the entire cohort from day 0 to 100 (n = 260) and the same among day 100 survivors (n = 240). Cytomegalovirus results after day 100 were not included. There were 125 deaths (32%) by day 365 after HCT. Thirty-six patients (9%) died in days 0–100, and 89 patients (23%) died in days 101–365. Preemptive antiviral therapy was initiated once the viral load reached 500 copies/mL (125 IU/mL) for standard-risk patients. Patients considered high-risk, such as those receiving ≥1 mg/kg body weight of prednisone or cord blood transplant recipients, were started at any positive viral load. Patients were screened for CMV weekly through day 100, except cord blood HCT recipients who were screened for CMV twice weekly. As of each patient’s final AUC calculation, the fraction of patients receiving antiviral therapy in each of the 5 viral load categories (from lowest to highest) was 16 of 147 (11%), 21 of 34 (62%), 85 of 100 (85%), 100 of 103 (97%), and 19 of 20 (95%). Abbreviations: aHR, adjusted hazard ratio; AUC, area under the curve; CI, confidence interval; CMV, cytomegalovirus; HCT, hematopoietic cell transplantation; HR, hazard ratio. aThis model was adjusted for age, HCT comorbidity index score, underlying disease risk, type of conditioning regimen (myeloablative vs not), time-dependent acute graft-versus-host disease grades 3–4, and time-dependent cumulative steroid dose AUC during the first 100 days after HCT (steroid use was categorized as 0, <1 mg/kg, 1 to <2 mg/kg, and ≥2 mg/kg prednisone equivalent per day). bThe plasma CMV DNA log10 viral load AUC (as copies × days × mL-1) was calculated using the curve-trapezoid rule. The AUC was calculated first and then log10-transformed. Patients with no CMV were assigned an AUC of 0. The CMV DNA log10 viral load AUC was incorporated as a time-dependent variable through day 100 and a time-invariant variable through day 365, using the AUC value at day 100 as the new baseline value (because routine testing ended at day 100). Neither analysis argues against the routine use of preemptive therapy for CMV at prespecified thresholds given that lower CMV AUC is not associated with mortality, provided that preemptive treatment can quickly mitigate viremia at the selected threshold. Our publication underscores the importance of preventing persistent viral episodes given increased risk for end-organ disease and higher AUCs. A recent study identified that peak CMV viremia ≥150IU/mL was associated with persistent viremia and nonrelapse mortality [3]. An important meta-analysis solidified the rationale for using CMV load as a surrogate endpoint for outcomes after solid organ transplantation, in part based on the association of higher viral load with CMV disease [4]. A phase 3 trial of letermovir for CMV prophylaxis after allogeneic HCT provided additional support for the association of CMV reactivation with mortality given an approximately 5% lower nonrelapse mortality at week 24 (P = .03) and week 48 (P = .12) relative to placebo [5]. Cytomegalovirus prevention through prophylaxis has been a goal in the field for decades, but routine implementation may be limited by cost, and preemptive strategies will remain important. Financial support. This work was supported in part by an investigator-initiated grant from Chimerix Inc. This work was also supported by the National Institutes of Health (5K23AI119133 to J. A. H.). Potential conflicts of interest. J. A. H. has served as a consultant for Chimerix Inc and Nohla Therapeutics Inc and has received research support from Chimerix Inc and Shire. M. B. reports grants and personal fees from Merck and Co; grants and personal fees from Astellas, Shire, Roche/Genentech, Gilead, and Chimerix; and personal fees from Clinigen and Microbiotix, outside the submitted work. J. T. S. and B. T. M. report research support from Chimerix Inc. H. X. and W. M. L declare no competing interests. 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.
Source: Clinical Infectious Diseases
DOI: 10.1093/cid/ciy139
Institutions: University of Washington, Fred Hutch Cancer Center
Language: French
Subfield: Epidemiology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
Awards: G3659335753, G9000254091
OpenAlex ID: W2795340135