Beyond Body Mass Index: The Impact of Height and Height-Normalised Weight on Overall Survival of Lung Cancer Undergoing Surgery.

Introduction: Unlike most malignancies, increased adiposity, as expressed by a higher body mass index (BMI), is associated with improved prognosis after lung cancer surgery at the population level. Height, one of the determinants of BMI, is associated with better survival, independent of other confounders, even though BMI is calculated as weight divided by height squared. The association of weight with survival is difficult to assess because, at the individual level, weight is closely linked to height and does not directly reflect adiposity. In this study, we examined the impact of height and weight on overall survival in a large population of patients undergoing upfront surgery for lung cancer.

Methods: We extracted data on all consecutive patients with stage I-IIIA non-small cell lung cancer included in a surgical nationwide dataset over a 16-year period. For each sex, height was categorised in sex-specific quartiles (sH). Sex-specific height-normalised weight (sHNW) was defined as the ratio of an individual's weight to the mean weight of individuals of the same sex and height, and it was categorised into quartiles. Finally, the sum of the category membership (ranging from 1 to 4 according to quartiles) of sH and sHNW was calculated, and the results were categorised into four groups of sH/sHNW. Overall survival (OS) was assessed by Kaplan-Meier, and differences evaluated by log-rank. Cox models were built.

Results: The study included 50 653 patients. Mean age was 65.61 ± 9.45 and 31.99% were women. sH predicted OS, taller height being protective [crude HRs of second, third, and fourth quartiles vs. first quartile: 0.94 (95% CI 0.91-0.98), 0.89 (0.86-0.92), 0.77 (0.74-0.81); p < 0.0001]. sHNW was also associated with OS, with lower sHNW category being associated with worse outcome and higher sHNW categories being protective [crude HRs of second, third and fourth quartiles vs. first quartile: 0.88 (0.85-0.92), 0.82 (0.79-0.85), 0.85 (0.81-0.88); p < 0.0001]. The four classes of sH/sHNW showed higher differences in prognosis with respective crude HRs of 0.88 (0.84-0.93), 0.76 (0.73-0.80) and 0.70 (0.66-0.74) in the intermediate lower, intermediate higher and higher class as compared with the lower class. Five-year overall survival rates were 58.65% (56.89-60.45), 62.96% (62.15-63.78), 67.71% (67.02-68.41) and 70.12% (68.98-71.26), in the lower, intermediate lower, intermediate higher and higher class, respectively. All Cox models showed that sHNW and sH/sHNW predicted survival independently from common confounders.

Conclusions: Our study demonstrated that sHNW and sH/sHNW are strong prognostic factors of resectable lung cancer. This finding could have both epidemiologic and biological relevance.