Background: Many of the indices utilized in the assessment of nutritional status are dependent upon stature. Knee height has been used to predict stature in western countries when standing height cannot be measured, however, an adequate stature-knee height equation has not been established for the population in Taiwan.
Methods: A total of 1179 healthy men (603) and women (576) were recruited from those who had undergone comprehensive check-ups at their own expense from March through August 1998 at Lin-Kou Chang-Gung Memorial Hospital, Taiwan. The subjects were freely ambulatory without kyphosis or scoliosis noted on their chest X-ray or KUB films. The automatic supersonic system was used to measure stature and body weight. A Ross caliper was used to measure recumbent left knee height. Multiple linear regression was used to obtain equations to estimate stature using knee height and age.
Results: Stature of adults highly correlated with their knee height (r = 0.73 with 95%CI between 0.69 and 0.77 for men, and r = 0.68 with 95%CI between 0.63 and 0.72 for women) but declined with age (men: r = -0.17 with 95%CI between -0.09 and -0.24; women: r = -0.22 with 95%CI between -0.14 and -0.29). Knee height changed very little with age in both genders. To predict stature, a linear regression model should include knee height and age as predictors to reach r2 as 0.61 and 0.58 for men and women, respectively. The stature-knee height equations were "stature = 85.10 + 1.73 x knee height -0.11 x age" for men, "stature = 91.45 + 1.53 x knee height -0.16 x age" for women.
Conclusion: Our stature-knee height equations provided good predictions on stature of adult population in Taiwan, and was superior when compared with equations developed from other populations.