FVC, total lung capacity, and the differential association to mortality.

I would like to thank Pedone et al 1 for their recent article in CHEST (April 2012). The article presented data on the association between pulmonary restriction, properly defi ned as having a total lung capacity (TLC) below the fi fth percentile of the reference population (lower limit of normal [LLN]), and all-cause mortality. As the authors correctly pointed out, defi ning restriction as FVC , 80% predicted (or even lower than the LLN) is neither specifi c nor sensitive as an indicator of true restriction. 2 They presented a hazard ratio for death of 2.73 when FVC , LLN and of 6.87 when TLC , LLN. Assumably, much of the increased risk due to FVC , LLN is due to true restriction (the subset of FVC , LLN that also has TLC , LLN), although this was not specifi cally presented in the article. I believe that it would be a very helpful (and probably fairly easy) addition to their analysis to present the hazard ratio for death separately for those with FVC , LLN and TLC , LLN compared with FVC , LLN but TLC . LLN. The hypothesis behind this requested analysis would be that FVC , LLN but without true restriction (TLC . LLN) does not show increased risk of death (or shows a lower risk of death). I realize that with only 17 deaths total, this may not give results that achieve statistical signifi cance, but lacking other similar data, the results may help us to interpret studies of FVC where TLC is not available. An additional comment: In Figure 2, the survival lines for the reference categories, respectively for FVC . LLN and TLC . LLN, are identical. This seems a bit strange because even though the deaths represented in these two lines could be identical, the numerator of the percentage must be different.