[Estimation of the menarcheal age in a longitudinal study: comparison of methods]

Menarcheal age was assessed in 164 girls followed longitudinally for six years, belonging to Caracas upper social strata. The retrospective, prospective and "Status Quo" methods were used; age was recorded as the recalled date, the mid-interval interpolated date and the age at examination; and called "real" age (RA), "interval" age (IA) and "visit" age (VA). Means were calculated and medians estimated using the Logit method. Means of differences between RA and VA and between IA and VA were similar: 0.4 years approximately and significantly greater than the corresponding mean between RA and IA (0.2 years), thus mid-interval and RA mean menarcheal ages were similar. Values ranged from 12,866 to 12,341 with a standard deviation of one year. Mean menarcheal "visit" age was significantly later than both IA and RA means. Median menarcheal VA using the "Status Quo" and the Logit methodologies, was 12,337 +/- 0.028, similar to "real" age, and has been chosen for further analysis and correlations. The results are consistent with the low menarcheal age values reported for Venezuela.