FEV1/FVC Ratio — GLI-2012 Equation-Derived Percentiles
The FEV1/FVC ratio (Tiffeneau Index) percentiles on this site are derived directly from the Global Lung Function Initiative 2012 (GLI-2012) LMS equation for the ratio, published by Quanjer et al. (DOI: 10.1183/09031936.00080312). GLI-2012 publishes a dedicated LMS equation for FEV1/FVC, fit directly to the ratio in the healthy reference population, rather than computing it post-hoc from separate FEV1 and FVC predictions. The lookup workbook (lookuptables.xls) ships separate FEV1FVC males and FEV1FVC females sheets with their own Lspline, Mspline, and Sspline columns and their own coefficient block.
Percentiles are computed as follows:
- Equations: Caucasian coefficients used (all ethnicity flags set to zero), representing 81% of male and 86% of female subjects in the GLI dataset.
- L equation: For males,
L = q0 + q1·ln(Age) + Lspline. For females,L = q0 + q1·ln(Age); there is no Lspline term, and the female Lspline column in the workbook is identically zero across the full 3-95 year range. The female equation header on the GLI-2012 sheet states this explicitly, and the derivation script branches on sex to enforce it. - M and S equations:
M = exp(a0 + a1·ln(Height) + a2·ln(Age) + Mspline)andS = exp(p0 + p1·ln(Age) + Sspline), same form as FEV1 but with the ratio-specific coefficients in the table below. - Height assumption: Each age-sex bracket uses the NHANES median standing height for that group (Fryar et al. 2021, DOI: 10.15620/cdc:103734). Height has a smaller effect on the ratio than on FEV1 in isolation, because FEV1 and FVC scale similarly with height and partly cancel;
a1is non-zero in both sexes, so the effect is not eliminated. - Percentile formula:
P = M × (1 + L × S × z)1/L, where z is the standard normal quantile. All five percentiles (P5, P25, P50, P75, P95) are computed natively from the LMS equation; no approximation or substitution is used. Output is rounded to 3 decimal places (the ratio ranges roughly 0.59 to 0.99 across the published age span).
| Coefficient | Males | Females |
|---|---|---|
| a0 (intercept) | 0.7403 | 0.550559 |
| a1 (ln Height) | -0.1595 | -0.107805 |
| a2 (ln Age) | -0.0366 | -0.054419 |
| p0 (S intercept) | -2.9595 | -3.23948 |
| p1 (S × ln Age) | 0.1156 | 0.18503 |
| q0 (L intercept) | 4.7101 | 7.032 |
| q1 (L × ln Age) | -0.6774 | -1.197 |
| L includes Lspline? | yes | no |
| Bracket | Males (cm) | Females (cm) |
|---|---|---|
| 5-9 | 119.5 | 118.5 |
| 10-14 | 149.0 | 151.5 |
| 15-19 | 175.5 | 162.5 |
| 20-29 | 176.4 | 162.1 |
| 30-39 | 176.0 | 162.0 |
| 40-49 | 175.7 | 161.8 |
| 50-59 | 174.5 | 161.0 |
| 60-69 | 172.5 | 159.5 |
| 70-79 | 170.0 | 157.5 |
| 80+ | 168.0 | 155.0 |
The full GLI-2012 derivation pattern (LMS method, ethnicity coefficients, spline interpolation) is documented on the FEV1 methodology page. The shared source paper and dataset details are on the GLI-2012 reference page.
On rating display: the FEV1/FVC ratio is not a higher-is-better or lower-is-better metric, so a standard Poor-to-Excellent rating column does not apply. The age/sex pages render the percentile chart without the rating column and without reference lines or tinted bands on the chart itself, so the chart remains purely a population-distribution view consistent with how every other metric on the site is rendered.