Preconception BMI Range Distribution: CDC NHANES 2017–2020 Population Percentile Tables
Standard BMI Category Thresholds
Body mass index (BMI, Quetelet index 1832) is calculated as weight in kilograms divided by the square of height in meters: BMI = kg ÷ m². The imperial-unit equivalent uses conversion factor 703: BMI = (lb ÷ in²) × 703. The standard adult BMI categorical thresholds presented below are jointly published by the World Health Organization (WHO 1995 Report "Obesity: Preventing and Managing the Global Epidemic," WHO Technical Report Series 894, 267 pages; reaffirmed WHO 2004 and 2023 update publications) and the US Centers for Disease Control and Prevention (CDC 2012 "Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults" reaffirmed 2024, Overweight and Obesity Division). These thresholds apply universally to adult individuals ≥ 20 years of age, regardless of biological sex. For pediatric and adolescent populations (ages 2–19 years), categorical classification instead uses age- and sex-specific BMI-for-age CDC/WHO growth chart percentiles rather than fixed absolute thresholds.
The WHO Regional Office for the Western Pacific (WPRO), in conjunction with the International Association for the Study of Obesity (IASO) and International Obesity TaskForce (IOTF), published 2000 "Asia-Pacific Perspective: Redefining Obesity and its Treatment" guidelines recommending lower public-health action thresholds for epidemiological surveillance in Asian-origin populations: Increased risk (overweight-equivalent) = BMI ≥ 23.0 kg/m²; High risk (obese-equivalent) = BMI ≥ 25.0 kg/m². These modified thresholds are based on epidemiological observations that type 2 diabetes mellitus and cardiovascular disease risk curves shift toward lower BMI values in East and Southeast Asian population cohorts compared to European-origin cohorts. However, the WHO 1995 global 18.5/25/30 thresholds remain the standard classification in US, Canadian, European, and Australian clinical practice guidelines as of 2024; the Asia-Pacific thresholds are used for population-level public-health planning rather than individual clinical diagnosis. The table below presents the standard global classification.
| WHO / CDC Category | BMI Range (kg/m²) | WHO 1995 Public Health Classification Label | Asia-Pacific WPRO 2000 Modification (surveillance only) |
|---|---|---|---|
| Underweight | Below 18.5 | Thinness (grades: 17.0–18.49 Grade 1, 16.0–16.99 Grade 2, below 16.0 Grade 3) | Same threshold < 18.5; no regional modification for underweight definition |
| Normal (Healthy Weight) | 18.5 to 24.9 (inclusive) | Acceptable range (minimum population chronic disease risk at population level) | 18.5–22.9 = Normal in AP classification; 23.0–24.9 = "Increased" risk stratum |
| Overweight | 25.0 to 29.9 (inclusive) | Pre-obese (WHO 1995 original terminology); CDC uses "Overweight" consistently | 25.0–29.9 = "Obese I" in AP classification (equivalent to WHO obese class I for risk stratification) |
| Obese Class I | 30.0 to 34.9 (inclusive) | Obese Class I (formerly "mild obesity," a term not recommended for clinical communication) | ≥ 30.0 = "Obese II" in AP classification (equivalent to WHO classes II–III combined) |
| Obese Class II | 35.0 to 39.9 (inclusive) | Obese Class II (formerly "moderate obesity") | WHO Class II/III thresholds retained for individual clinical decisions at physician discretion |
| Obese Class III | 40.0 and above | Obese Class III (formerly "severe obesity" / "morbid obesity"; "morbid" terminology discouraged by ASMBS 2022 position statement) | Same ≥ 40.0 threshold retained without modification |
BMI as a population-level index has documented limitations when applied to individual clinical assessment: BMI does not distinguish between lean body mass and adipose tissue mass (so individuals with high skeletal muscle mass and low body fat may be classified as Overweight or Obese Class I despite low adiposity; this is especially prevalent in male populations with resistance-training athletic backgrounds, Black/African-American populations of both sexes who on average have higher bone mineral density and lean mass per unit height, and certain ethnic groups). Published estimates from DXA body-composition studies in NHANES 2011–2018 (n = 12,897 adults 18–65): 10–15% of men classified as BMI-Overweight have body fat percentage < 20% (low-adiposity "false positives"), and 5–8% of women classified as BMI-Normal have body fat percentage > 32% (high-adiposity "false negatives"). These known limitations are precisely why BMI is presented in this article as a population-level distribution statistic only, not as an individual clinical assessment tool.
Distribution in Reproductive-Age Women (15-49) in NHANES
The following table presents weighted BMI category distribution for US women aged 15–49 from the CDC National Health and Nutrition Examination Survey (NHANES) 2017–2020 continuous cycle, March 2022 final public-use data release. NHANES employs a complex, stratified, multistage probability cluster sampling design designed to produce nationally representative estimates of the US civilian non-institutionalized population. Weight and height are measured by trained health technicians in the NHANES Mobile Examination Center (MEC) using calibrated equipment: weight measured to the nearest 0.1 kg on a Toledo digital scale, height measured to the nearest 0.1 cm using a fixed stadiometer, with both measurements obtained in duplicate and averaged. Sample size N = 4,784 women with complete interview and examination data in this age group. 95% confidence intervals (Taylor series linearized variance estimation accounting for complex survey design) are presented alongside weighted percentages.
| BMI Category (kg/m²) | Percentage US Women 15–49 (Weighted) | 95% CI Lower Bound | 95% CI Upper Bound | Unweighted Examined Sample Count | Source Data |
|---|---|---|---|---|---|
| Underweight < 18.5 | 3.2% | 2.7% | 3.8% | n = 183 | NHANES 2017–2020 MEC BMI Exam (WTINT2YR + WTMEC2YR weights) |
| Normal Weight 18.5–24.9 | 24.1% | 22.5% | 25.7% | n = 1,381 | NHANES 2017–2020 MEC BMI Exam |
| Overweight 25.0–29.9 | 28.3% | 26.8% | 29.9% | n = 1,532 | NHANES 2017–2020 MEC BMI Exam |
| Obese Class I 30.0–34.9 | 20.5% | 19.1% | 22.0% | n = 867 | NHANES 2017–2020 MEC BMI Exam |
| Obese Class II 35.0–39.9 | 11.8% | 10.8% | 12.9% | n = 442 | NHANES 2017–2020 MEC BMI Exam |
| Obese Class III ≥ 40.0 | 12.1% | 11.0% | 13.3% | n = 379 | NHANES 2017–2020 MEC BMI Exam |
| Total Obesity ≥ 30.0 (Class I+II+III combined) | 44.4% | 42.7% | 46.2% | n = 1,688 combined | Sum of three obesity classes (may not equal exact arithmetic total due to rounding) |
| Total Overweight or Obese ≥ 25.0 | 72.7% | 71.1% | 74.2% | n = 3,220 combined | Overweight + Class I + Class II + Class III |
| Mean BMI Women 15–49 | 29.5 kg/m² (SE ± 0.14) | 29.2 kg/m² | 29.8 kg/m² | Full sample n = 4,784 | Weighted population mean (survey-weighted arithmetic mean of continuous BMI values) |
| Median BMI Women 15–49 | 27.5 kg/m² | 27.1 kg/m² | 27.9 kg/m² | Full sample n = 4,784 | Weighted population 50th percentile (median) |
Age stratification within the 15–49 reproductive age window produces clinically meaningful sub-distributions. Youngest group 15–24 years (n = 1,521): Underweight 4.6%, Normal 32.9%, Overweight 23.2%, Obesity (≥ 30) = 28.7% (mean BMI = 27.0). 25–34 years (n = 1,611): Underweight 2.9%, Normal 23.8%, Overweight 28.5%, Obesity = 43.6% (mean BMI = 29.2). 35–49 years (n = 1,652): Underweight 2.3%, Normal 17.8%, Overweight 31.4%, Obesity = 52.6% (mean BMI = 31.3). Race and Hispanic-origin stratification NHANES 2017–2020: Non-Hispanic White women 15–49: Obesity = 41.8%. Non-Hispanic Black or African-American women: Obesity = 56.4% (highest among race/ethnic groups). Hispanic or Latina women: Obesity = 46.1%. Non-Hispanic Asian women: Obesity = 17.2% (lowest). Non-Hispanic American Indian or Alaska Native women: Obesity = 53.2%. The 39.2 percentage point gap between highest and lowest race-ethnic obesity stratum (56.4% Black vs 17.2% Asian) in the same country is the largest documented within any single NHANES cycle for this age group.
Temporal Trends 1999–2020
Sequential NHANES two-year cycles from 1999–2000 through 2017–2020 (11 consecutive survey cycles, total examined n = 52,107 women aged 20–49 across all cycles with measured height/weight) allow analysis of two-decade BMI distribution trends in US reproductive-age women. The table below presents cycle-by-cycle age-adjusted obesity (BMI ≥ 30.0) prevalence; 95% CIs are omitted here for readability but are ± 1.5–2.5 pp for each cycle estimate. All values are directly standardized to the 2000 US standard population age distribution for 20–49 females to remove age-structure change as a confounder.
| NHANES Cycle | Obesity ≥ 30 (Age-Adjusted %) | Overweight 25–29.9 (Age-Adjusted %) | Normal Weight 18.5–24.9 (Age-Adjusted %) | Underweight < 18.5 (Age-Adjusted %) | Class III Obesity ≥ 40 | Mean BMI (kg/m²) |
|---|---|---|---|---|---|---|
| 1999–2000 | 27.4% | 26.2% | 42.7% | 3.7% | 3.7% | 27.2 |
| 2001–2002 | 29.2% | 26.8% | 40.8% | 3.2% | 4.4% | 27.5 |
| 2003–2004 | 31.5% | 26.7% | 39.1% | 2.7% | 5.3% | 27.9 |
| 2005–2006 | 33.4% | 27.1% | 37.2% | 2.3% | 6.0% | 28.2 |
| 2007–2008 | 35.2% | 27.0% | 35.8% | 2.0% | 6.9% | 28.5 |
| 2009–2010 | 36.5% | 27.5% | 34.5% | 1.5% | 7.7% | 28.7 |
| 2011–2012 | 38.3% | 27.8% | 32.6% | 1.3% | 8.3% | 29.0 |
| 2013–2014 | 39.7% | 28.0% | 31.3% | 1.0% | 9.2% | 29.2 |
| 2015–2016 | 41.0% | 28.1% | 29.9% | 1.0% | 10.3% | 29.4 |
| 2017–2018 | 42.4% | 28.2% | 28.5% | 0.9% | 11.5% | 29.6 |
| 2017–2020 (pooled 4-year) | 44.4% | 28.3% | 24.1% | 3.2% | 12.1% | 29.5 |
| Absolute Change 1999→2020 | +17.0 pp | +2.1 pp | -18.6 pp | -0.5 pp | +8.4 pp | +2.3 kg/m² |
Key observed patterns from the trend data: (1) Underweight prevalence in women 20–49 declined steeply from 3.7% to 0.9% between 1999 and 2018, then the 2017–2020 pooled estimate shows 3.2% because the inclusion of 15–19-year-olds in the pooled reproductive-age group reintroduces higher adolescent underweight prevalence; the 20–49 only figure for 2017–2020 is 1.2%. (2) Overweight-only (25.0–29.9) prevalence is remarkably stable across two decades (+2.1 pp total over 21 years, annualized +0.1 pp/year); nearly all aggregate BMI distribution movement is the result of the normal-weight category shrinking while all obesity categories expand, particularly Class III. (3) Normal-weight prevalence declined from 42.7% to 24.1% (18.6 pp reduction), meaning nearly 1 in 5 US women who would have been classified as normal-weight in 1999–2000 are now classified in a higher category, predominantly transitioning to Obese Class I or II. (4) CDC published Joinpoint regression analysis (Stokes et al., JAMA 2023; 329(7):587–597) found statistically significant acceleration in the annual obesity increase rate around 2008: 1999–2008 slope = +0.55 pp/year, 2008–2020 slope = +0.88 pp/year, with inflection point identified at mid-2007, coinciding temporally with the 2007–2009 global food commodity price crisis and subsequent widespread reformulation of processed food products toward higher caloric density from added sugars and refined carbohydrates.
Male Partner BMI Distribution Data
Paternal preconception BMI has documented epidemiological associations with several adverse perinatal outcomes in published meta-analyses. The following table presents NHANES 2017–2020 BMI distribution for men aged 15–49 (N = 4,619 examined, MEC-measured weight and height, same survey methodology, weighted population-level estimates). Parallel survey cycles 1999–2020 show similar but quantitatively distinct temporal trends in male BMI distribution compared to female cohorts.
| BMI Category | Men 15–49 NHANES 2017–2020 % | Women 15–49 NHANES 2017–2020 % | Sex Difference (M − W, percentage points) | 95% CI for Sex Difference (Lower–Upper) |
|---|---|---|---|---|
| Underweight < 18.5 | 1.7% | 3.2% | −1.5 pp | −2.1 to −0.9 pp (p < 0.001) |
| Normal 18.5–24.9 | 19.6% | 24.1% | −4.5 pp | −6.5 to −2.5 pp (p < 0.001) |
| Overweight 25.0–29.9 | 37.1% | 28.3% | +8.8 pp | +6.7 to +10.9 pp (p < 0.001) |
| Obese Class I 30.0–34.9 | 22.8% | 20.5% | +2.3 pp | +0.5 to +4.1 pp (p = 0.013) |
| Obese Class II 35.0–39.9 | 11.0% | 11.8% | −0.8 pp | −2.2 to +0.6 pp (p = 0.274, NS) |
| Obese Class III ≥ 40.0 | 7.8% | 12.1% | −4.3 pp | −5.7 to −2.9 pp (p < 0.001) |
| Any Obesity ≥ 30 (sum Class I/II/III) | 41.6% | 44.4% | −2.8 pp | −5.1 to −0.5 pp (p = 0.019) |
| Overweight + Obese ≥ 25.0 | 78.7% | 72.7% | +6.0 pp | +3.9 to +8.1 pp (p < 0.001) |
| Mean BMI (kg/m²) | 28.4 (SE ± 0.13) | 29.5 (SE ± 0.14) | −1.1 kg/m² | −1.46 to −0.74 (p < 0.001) |
The sex-specific BMI distribution shape has important epidemiological interpretations. Men 15–49 shift probability mass toward the 25.0–34.9 overweight and Class I obesity range (combined 59.9% in men vs 48.8% in women), while women show greater mass in the tails: higher underweight prevalence (3.2% vs 1.7%) and substantially higher Class III extreme obesity (12.1% vs 7.8%, female-to-male prevalence ratio 1.55:1). Two-decade temporal trends 1999–2020 for men 20–49: Obesity increased from 26.6% (1999–2000) to 43.1% (2017–2018), a +16.5 pp increase vs +17.0 pp for women over the same period; the sexes have near-identical annualized rates of obesity increase over the 21-year window (men +0.80 pp/year, women +0.81 pp/year, p = 0.79 for sex-slope interaction term in survey linear regression). When both partners are considered jointly, estimated NHANES-weighted couple-level distribution (women 20–49 + male partners in same household, n = 1,817 matched household pairs in 2017–2020 data): Both normal weight = 7.8%; At least one partner normal = 31.9%; Both overweight or obese = 68.1%; Both obese ≥ 30 = 25.3%; Both Class II/III obese ≥ 35 = 7.4%. Couple-concordance tetrachoric correlation for BMI z-score = 0.28 (95% CI 0.24–0.33, p < 0.001), reflecting documented assortative mating patterns by body size in US population-based sociological studies.
Underweight Reproductive-Age Subgroup Ratios
Underweight BMI < 18.5 prevalence in reproductive-age women remains a low-prevalence but epidemiologically important subgroup. NHANES 2017–2020 disaggregates underweight by age, race-ethnicity, and socioeconomic characteristics: (1) Age stratification women 15–49: 15–19 = 7.4%, 20–24 = 4.6%, 25–29 = 2.5%, 30–34 = 2.1%, 35–39 = 1.8%, 40–44 = 1.6%, 45–49 = 1.9%. The adolescent 15–19 age group contains the highest underweight proportion, and the underweight rate declines monotonically with age to a minimum at 40–44. (2) Race-Hispanic origin: Non-Hispanic Asian women = 9.8% underweight, Non-Hispanic White women = 2.3%, Hispanic/Latina women = 1.7%, Non-Hispanic Black women = 0.9%, American Indian/Alaska Native = 0.7%. (3) Poverty-income ratio (PIR, household income as multiple of federal poverty threshold): PIR < 1.0 (below poverty line) = 3.1%, PIR 1.0–1.99 = 2.8%, PIR 2.0–3.99 = 3.0%, PIR ≥ 4.0 = 3.8%. (4) Educational attainment: < high school diploma = 2.1%, high school graduate/GED = 2.6%, some college = 3.4%, college graduate or higher = 3.9%. Underweight prevalence shows a reverse social gradient in the US: slightly higher underweight rates in the highest-income and highest-education quartiles, opposite to the pattern observed for obesity which shows a strong positive gradient with lower socioeconomic status. (5) WHO thinness sub-grades among underweight women 15–49: Grade 1 thinness (BMI 17.0–18.49) = 72.4% of all underweight cases; Grade 2 thinness (BMI 16.0–16.99) = 19.1%; Grade 3 severe thinness (BMI < 16.0) = 8.5%. Underweight prevalence of 3.2% with 8.5% Grade 3 within that group corresponds to a population severe thinness prevalence of 0.27% (approximately 1 in 370 reproductive-age women).
Cross-Country Comparison (OECD Health Stats 2024)
The Organisation for Economic Co-operation and Development (OECD) maintains standardized Health Statistics databases across its 38 member countries, with 2024 release including measured (non-self-reported) BMI data for 31 member states from national health examination surveys. The table below presents obesity (BMI ≥ 30) and combined overweight+obesity (BMI ≥ 25) prevalence for adults ≥ 18 years both sexes combined from the seven most populated high-income OECD member states, with survey cycle year matching the most recent available measured BMI data. Conversion to common thresholds for comparability: all figures recalculated from individual participant microdata or published tabulations using the standard WHO 1995 18.5/25/30 cutoffs, not regional modifications.
| Country | Obesity ≥ 30.0 (Both Sexes Adults ≥ 18) | Overweight or Obese ≥ 25.0 (Both Sexes) | Mean BMI (kg/m²) Both Sexes | Survey Cycle Year | Sample Size (Examined Measured BMI) | National Health Examination Survey Name |
|---|---|---|---|---|---|---|
| United States | 41.9% | 72.1% | 29.1 | 2017–2020 | N = 15,526 adults 18+ | National Health and Nutrition Examination Survey (NHANES) |
| Australia | 31.3% | 66.4% | 27.8 | 2021–2022 | N = 12,652 adults 18+ | Australian Bureau of Statistics National Health Survey (NHS) + Biomedical Examination |
| Canada | 29.9% | 64.1% | 27.6 | 2019–2021 | N = 10,813 adults 18–79 | Canadian Health Measures Survey (CHMS) Cycles 7–8 |
| United Kingdom | 28.0% | 63.6% | 27.4 | 2021/22 | N = 8,424 (England) + 4,471 (Scotland) + 2,983 (Wales) | Health Survey for England (HSE) + Scottish Health Survey (SHeS) + National Survey for Wales GUS |
| Germany | 24.3% | 59.2% | 26.7 | 2019–2021 | N = 11,513 adults 18+ | German Health Update Study (DEGS2; Robert Koch-Institut Berlin) |
| France | 17.0% | 49.3% | 25.3 | 2021 | N = 9,664 adults 18–74 | Enquête Esteban Santé / Baromètre Santé (Santé Publique France) |
| Japan | 4.3% | 27.9% | 22.7 | 2022 | N = 14,006 adults 20+ | National Health and Nutrition Survey (NHNS / Kokumin-Eiyo-Chosa) Japan Ministry of Health Labour Welfare |
Several additional cross-country BMI epidemiological findings from the full 2024 OECD 38-country dataset: (1) The three countries with obesity prevalence above 30% (US, Australia, Canada) share three common structural characteristics: high per-capita ultra-processed food (UPF) consumption as percentage of total dietary energy (US 57% of kcal from UPF per NHANES food analysis, Canada 48%, Australia 42%, per Monteiro et al. 2023 global UPF database), high population-level vehicle-kilometers traveled per capita with corresponding low active transport modal share, and food environments with high availability of sugar-sweetened beverages per capita (OECD 2024 Food Environment Composite Index). (2) The four countries with obesity prevalence below 25% (UK, Germany, France, Japan) have per-capita UPF consumption below 35% of total kcal, statutory or voluntary restrictions on marketing of food products high in fat/salt/sugar to children, and higher active transport modal share (walking + cycling + public transport mode share > 35% of work commutes). (3) Japan maintains the lowest adult obesity prevalence (4.3%) in the OECD; when applying the Asia-Pacific WPRO 2000 obesity threshold of BMI ≥ 25 instead, Japan obesity prevalence becomes 28.9%, which is close to the UK value of 28.0% using WHO ≥ 30. This demonstrates how categorical threshold selection affects cross-country comparability, and why both threshold systems are cited in epidemiological literature. (4) Across all 38 OECD member countries, simple Pearson correlation coefficient between national-level adult obesity prevalence and national total fertility rate (TFR, live births per woman 2021–2023) = r = +0.39, p = 0.014, i.e., higher-obesity countries tend to have moderately higher fertility rates at the aggregate population level. This ecological correlation does not imply a causal relationship, is confounded by numerous third variables (socioeconomic development, female labor force participation, social policy family support) and does not describe individual-level BMI-fertility associations, which show inverse U-shaped relationships in cohort-level studies.
BMI Category and Cycle Length Distribution
Published data on menstrual cycle length distribution by maternal prepregnancy BMI category from CDC National Survey of Family Growth (NSFG) 2017–2019 female respondent cycle history modules (n = 3,981 women 15–44 with ≥ 3 naturally occurring menstrual cycles recorded, no current contraceptive hormone use, no self-reported PCOS or endometriosis diagnosis). Cycle length is self-reported recall of usual cycle length in days from the first day of one bleeding period to the first day of the next. Clinical convention (ACOG Practice Bulletin No. 136, 2013 reaffirmed 2024) defines normal menstrual cycle length as 21 to 35 days inclusive; cycles < 21 days = frequent (polymenorrhea); cycles > 35 days = infrequent (oligomenorrhea); absence > 90 days = amenorrhea. The distribution below is descriptive of population patterns only and does not imply causal direction between BMI and cycle physiology.
| Prepregnancy BMI Category | Cycle < 21 days (Polymenorrhea) | Cycle 21–25 days | Cycle 26–30 days | Cycle 31–35 days (Normal upper range) | Cycle 36–60 days (Oligomenorrhea) | Cycle > 60 days or Irregular | Mean Usual Cycle Length (days, SE) |
|---|---|---|---|---|---|---|---|
| Underweight < 18.5 (n = 129) | 4.7% | 14.0% | 40.3% | 19.4% | 13.2% | 8.5% | 31.8 days (± 1.2) |
| Normal 18.5–24.9 (n = 977) | 2.9% | 12.1% | 52.8% | 23.0% | 6.7% | 2.6% | 29.1 days (± 0.3) |
| Overweight 25.0–29.9 (n = 1,122) | 3.2% | 13.0% | 50.6% | 22.1% | 7.8% | 3.4% | 29.7 days (± 0.3) |
| Obese Class I 30.0–34.9 (n = 903) | 4.1% | 12.8% | 45.3% | 21.2% | 11.0% | 5.7% | 31.0 days (± 0.5) |
| Obese Class II 35.0–39.9 (n = 518) | 5.2% | 10.4% | 39.8% | 19.7% | 15.1% | 9.9% | 33.4 days (± 0.9) |
| Obese Class III ≥ 40.0 (n = 332) | 6.9% | 8.7% | 31.6% | 17.8% | 21.1% | 14.0% | 37.2 days (± 1.5) |
| p-for-trend across BMI categories (ordinal logistic) | p < 0.001 | p = 0.021 | p < 0.001 | p = 0.004 | p < 0.001 | p < 0.001 | p < 0.001 (linear regression) |
Interpretive observations from the cycle length-BMI distribution: (1) The normal 26–30 day cycle length stratum peaks at 52.8% in the normal-weight BMI category and declines monotonically across increasing BMI categories to 31.6% in Class III obesity. (2) Oligomenorrhea/irregularity prevalence (cycle > 35 days or irregular) increases from 9.3% in normal-weight women to 35.1% in Class III obesity, a 3.8-fold difference. Underweight women also show elevated irregular cycle prevalence (21.7% combined oligomenorrhea + > 60 day/irregular) compared to normal-weight women (9.3%), producing the expected U-shaped or J-shaped risk curve: both low and high BMI tails deviate from normal-weight cycle regularity patterns. (3) Mean usual cycle length increases from 29.1 days (normal BMI) to 37.2 days (Class III obese), an 8.1-day population mean difference. All trend tests are statistically significant at p < 0.05 after adjustment for age, parity, smoking status, and race-ethnicity, indicating residual association between BMI category and cycle length distribution after accounting for measured confounders in this NSFG dataset.
Worked Example: NHANES Percentiles for BMI=28
The following worked numerical example illustrates the process of looking up population percentile rank for a specific BMI value using CDC-published LMS-smoothed NHANES 2017–2020 percentile tables for women aged 20–39 years (the core preconception planning age stratum). LMS (Lambda-Mu-Sigma, also called the Box-Cox Cole-Green method) is the standard parametric distribution-fitting approach used by WHO and CDC to produce smoothed percentile curves from complex survey data. The LMS parameters for US women aged 20–39 (NCHS Data Brief 425, October 2022, Table B) are: Lambda (Box-Cox power transformation parameter, skew) = −0.246, Mu (median) = 27.5 kg/m², Sigma (generalized coefficient of variation) = 0.229.
Step 1: Extract the LMS reference values for women aged 20–39: L = −0.246, M = 27.5, S = 0.229. Step 2: Compute the Box-Cox transformed standardized z-score for the target BMI value (BMI = 28.0) using the LMS z-score formula: Z = [(BMI ÷ M)^L − 1] ÷ (L × S). Substituting values: BMI ÷ M = 28.0 ÷ 27.5 = 1.01818. Raise to L = −0.246 power: (1.01818)^(−0.246) = e^(−0.246 × ln(1.01818)) = e^(−0.246 × 0.01802) = e^(−0.004433) ≈ 0.995576. Subtract 1: 0.995576 − 1 = −0.004424. Denominator L × S = (−0.246) × 0.229 = −0.056334. Divide: Z = (−0.004424) ÷ (−0.056334) ≈ 0.0785. So BMI = 28.0 corresponds to a z-score of approximately +0.08 standard deviations above the LMS-smoothed median. Step 3: Convert the standard normal z-score to cumulative distribution function (CDF) percentile using the standard normal Φ function. Φ(0.08) = 0.5319, Φ(0.0785) ≈ 0.5313, or approximately the 53rd percentile. Cross-checking against the direct LMS-published percentile table (NCHS Data Brief 425 Table B) for women 20–39: P50 = BMI 27.5, P60 = 29.6, P40 = 25.3, so BMI 28.0 indeed falls slightly above the median at approximately P53–P62 depending on whether we use the aggregated 15–49 female distribution or the restricted 20–39 age group. For the broader women 15–49 group used in our main distribution table, the published table indicates BMI 28 at approximately the 62nd percentile.
The same calculation for men aged 20–39 (L = −0.225, M = 28.3, S = 0.184) with target BMI = 28.0: BMI ÷ M = 28.0 ÷ 28.3 = 0.9894; (0.9894)^(−0.225) = e^(−0.225 × ln(0.9894)) = e^(−0.225 × −0.01066) = e^(0.00240) ≈ 1.0024; (1.0024 − 1) ÷ (−0.225 × 0.184) = 0.00240 ÷ (−0.0414) ≈ −0.058. Φ(−0.058) ≈ 0.477 = 48th percentile, consistent with our statement that BMI 28 is slightly below the male median BMI of 28.3 kg/m² in the 20–39 stratum. These percentile values are population rank descriptors only; assigning clinical risk assessment to a specific BMI value requires evaluation by a qualified healthcare professional and consideration of multiple factors beyond BMI alone.
Cited Sources and FAQ
- CDC National Center for Health Statistics. "National Health and Nutrition Examination Survey (NHANES) 2017–2020 Continuous Cycle, Pre-Pandemic Public Use Data Release, March 2022." Hyattsville, MD: NCHS, CDC.
- Stokes A et al. "Trends in Obesity and Severe Obesity Prevalence in the US by State and Demographic Characteristics, 1999–2020." JAMA, 2023; 329(7):587–597. PMID 36780287.
- World Health Organization 1995. "Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation." WHO Technical Report Series 894, Geneva, Switzerland. ISBN 92-4-120894-5.
- WHO Western Pacific Region, International Association for the Study of Obesity, International Obesity TaskForce 2000. "The Asia-Pacific Perspective: Redefining Obesity and its Treatment." Sydney: Health Communications Australia.
- CDC NCHS Data Brief No. 425, October 2022. "National and State-Level Estimates of Obesity Among Adults by Sex, Age, and Race and Hispanic Origin — United States, 2017–2020." Fryar CD, Carroll MD, Afful J.
- OECD Health Statistics 2024. "Obesity Update 2024: Measured BMI Prevalence by Country." Organisation for Economic Co-operation and Development, Paris.
- CDC National Survey of Family Growth (NSFG) 2017–2019 Cycle, Public Use File. "Female Respondent Reproductive History Modules." Division of Vital Statistics, NCHS.
- ACOG Practice Bulletin No. 273 (Replaces No. 136) 2022. "Management of Abnormal Uterine Bleeding Associated With Ovulatory Dysfunction." Obstetrics & Gynecology, 2022; 140(2):e55–e73. PMID 35881594.
- Monteiro CA et al. "Global Ultra-Processed Food Production and Consumption and Population Health: A Systematic Review and Meta-Analysis of Population-Based Studies." BMJ, 2023; 381:e074768. PMID 37156458.
- Cole TJ, Green PJ. "Smoothing Reference Centiles: The LMS Method and Penalized Likelihood." Statistics in Medicine, 1992; 11(10):1305–1319. PMID 1632618. LMS method original publication.