Best-fit PEFR nomograms for Hong Kong Chinese

Mark SH Chan陳選豪, Nai-ming Wong黃乃明, Albert YF Kong江炎輝, Yuk-tsan Wun溫煜讚, Tai-pong Lam林大邦, Wilson Tam談維新

Dear Editor

We reported to the Hong Kong Practitioner the best practical nomogram of peak expiratory flow rates (PEFR) for Hong Kong Chinese basing on our recent study.1 We find that the nomogram printed in the journal does not reproduce exactly the original nomogram submitted. The male PEFR line is flattened at the top probably due to a mal-alignment printing error. For those aged 60 years or above, the nomogram gives lower values than the regression equation. We also note that colleagues could be confused on finding a PEFR higher for an age of 70 than for 60. This is because we present a linear regression equation for our study population (age range of 5 -65 years). This equation gives a less complicated nomogram; it is easier to visualize the expected value from the nomogram and easier to compute it. Hence we call it the best practical equation. The R2 of this equation is 0.687 for males or 0.458 for females. (R2 is the coefficient of determination, a measure of how well future outcomes are likely to be predicted by the model.2) The linear equation has limitations. It works very well if the relationship between two variables is consistent throughout the whole range, like the Fahrenheit and the Centigrade scales of the temperature. For the age range of 5 - 65 years, PEFR rises steeply from age 5 to 20, then levels till somewhere between ages 40 and 45, and finally declines gradually. It is called a curvilinear relationship. While a linear regression can also be used to represent this relationship, it overestimates the PEFRs at the extremes of the age range.

Apart from the best practical regression equation, we also develop the best fit regression equation to comply with the curvilinear relationship. These equations are slightly more complicated and give slightly better R2 (0.726 for males and 0.481 for females). For the females, both the linear and curvilinear equations give similar R2.

We present here the best fit regression equations and the corresponding nomograms for our colleagues (height in centimeters).

Best regression for male:

PEFR=251.976 + 10.655*Age - 0.133*Age Square -4.814*Height + 0.031*Height Square

Best regression for female:

PEFR= -324.932 + 5.518*Age - 0.063*Age Square + 4.473*Height - 0.005*Height Square

Sincerely

Mark SH Chan, MPainMed, DFM, FHKCFP, FRACGP
Family Physician in Private Practice

Nai-ming Wong, MBBS, DOM, FHKCFP, FRACGP
Family Physician in Private Practice

Albert YF Kong, FRCP, FHKCFP, FRACGP, FHKAM
Pediatrician in Private Practice

Yuk-tsan Wun, MD, FHKCFP, FRACGP, FHKAM
Research Committee, HKCFP.

Tai-pong Lam, PhD, MD, FRCP, FHKAM
Associate Professor,
Family Medicine Unit, The University of Hong Kong

Wilson Tam, PhD
School of Public Health, University of Hong Kong