Population-level measles RT–PCR cycle threshold values and their relationships with disease severity, transmission, and incidence metrics: a time series analysis of the 2025–2026 outbreak in Mexico

Objectives To determine whether the weekly median real-time reverse-transcription polymerase chain reaction (RT–PCR) quantification cycle (Cq) value of notified measles cases precedes the incidence, transmission, and severity metrics of an outbreak. We focused particularly on the effective reproduction number ( R t ). We also assessed whether any association persisted after adjustment for a linear time trend, specimen composition, and the outcome's own past. Methods We conducted a retrospective secondary analysis of 10,910 Cq determinations and 11,735 confirmed cases recorded in a single Mexican national case-level surveillance extract for Chihuahua and Jalisco. Twelve reported weekly metrics of incidence, transmission, and severity were examined by cross-correlation, four nested lagged regression models (M1–M4), formal Toda–Yamamoto Granger tests, and a secondary calibrated-lag autoregressive model. Moving-block bootstrap intervals were the primary method for regression inference. Autocorrelation-robust and classical methods were used in sensitivity analyses. Individual Cq determinations were analyzed separately against collection timing, severity, and vaccination history. Results Selected-lag associations were negative across all reported metric families, at leads of 1–9 weeks. The strongest negative correlation between Cq and DGE-derived R t occurred at +8 weeks in the second Chihuahua wave ( r = −0.745). In Jalisco, it occurred at +7 weeks ( r = −0.464). Only Jalisco supported the fully adjusted model 4 (M4). There, bootstrap intervals excluded zero for both the laboratory-positive-case and final surveillance-confirmed-case R t estimates but not for the two severity proportions. Excluding up to four endpoint weeks preserved the direction of both R t estimates. At the individual level, Cq increased with zero, one, and two documented vaccine doses (0.761 cycles per dose, 95% CI 0.589 to 0.932) and was 1.02 cycles lower in hospitalized cases. Conclusions Aggregate Cq showed the most consistent retrospective leading association with R t . The retrospective associations occurred at data-calibrated lags of 7–8 weeks during the second Chihuahua wave and in Jalisco. Conversely, leak-free forecasting supported incremental gain in one analytic unit at a median 2-week horizon. The operational predictive horizon of aggregate Cq therefore remains unvalidated. If prospectively validated, weekly aggregate Cq could complement conventional measles surveillance using laboratory information that is already routinely collected.