NHANES 1999-2006 calf circumference

Full name
Calf circumference percentiles for US adults derived from NHANES 1999-2006 public microdata
Year
2026
Sample size
n = 17.8K
Population
US adults aged 18 and over in NHANES 1999-2006, with complete anthropometric and DXA primary fields; known and unknown pregnancy status excluded
Study type
Cross-sectional nationally representative health survey; percentiles derived by FitnessNorms from CDC public microdata and compared with Gonzalez et al. (2021)
Link
DOI: 10.1093/ajcn/nqab029

About this reference

FitnessNorms derived weighted P5, P25, P50, P75, and P95 calf-circumference estimates from four public NHANES releases (1999-2000 through 2005-2006). Body-measures, demographics, and DXA files were merged by participant ID; the five multiply imputed DXA records were reduced to one participant record using _MULT_ = 1. The complete-case rules reproduce the 17,856-person eligible cohort reported by Gonzalez et al. exactly. The paper says that 67 records were removed using an IQR method with a threshold of 3, but it does not specify the grouping or implementation. A conventional global Tukey fence (below Q1 - 3 × IQR or above Q3 + 3 × IQR) removes 17 records, and tested sex, age, race/ethnicity, and BMI groupings did not reproduce both the paper's removal count and sex split. FitnessNorms therefore publishes a clearly separate 17,839-person derivation using the conventional rule. This avoids attributing the paper's n=17,789 to results that cannot be exactly replicated. Eight-year MEC weights follow the CDC combination rule: the 1999-2002 four-year weights are multiplied by 0.5, and each 2003-2006 two-year weight is multiplied by 0.25. Weighted empirical quantiles use midpoint cumulative-weight interpolation. The weighted sex means round to the paper's reported 39.3 cm for men and 38.2 cm for women. P25 and P75 are newly derived; survey-design standard errors are not estimated.

Known limitations

Used for