From the authors:

Based on results of their important studies and assumptions concerning the Study of Health in Pomerania (SHIP) database, J.A. Neder questions the relevance of our report on reference values for cardiopulmonary exercise testing (CPET) 1. The establishment of close-to-reality reference values for the clinical interpretation of CPET is indeed an important issue since a consensus has not yet been reached on the definition of normalcy. Therefore, we consider our study to be an additional contribution to present an “ideal” set of normative values for CPET. In this context, the major strength of the current study first lies in the use of data from a large population-based sample of adults. Secondly, the consideration of echocardiographical and lung functional data besides comprehensive information on past and current medical history, as well as the performance of all kinds of clinical examination methods contributes exceedingly to the establishment of a disease-free reference sample.

J.A. Neder mainly criticises the issue of voluntary participation and the aspect of too little consideration of physical activity in the current study sample on CPET. At this point, we would like to put emphasis on the fact that the SHIP study sample itself, including the participants in CPET, was randomly selected from the general population via registration offices 2. Within a democratically ruled country, eventual participation in the study is dependent on voluntary participation. Therefore, one can argue that every epidemiological survey, to some extent, is biased, as stated in the published article; a selection bias towards younger and healthier individuals was evident. However, by means of adjusting for age and applying stringent exclusion criteria towards healthy participants, this should only marginally affect the normative limits for CPET.

J.A. Neder impeaches the applicability of the present reference values in the specific sub-population of sedentary elderly subjects since the described predicted values for peak oxygen uptake (V′_O2) were systematically higher than those previously described. Besides a selection bias towards younger and slimmer volunteers our data may be biased by analysing CPET results of physically more active participants aged ≥50 yrs (<2 h·week⁻¹ versus ≥2 h·week⁻¹; p<0.01). Within the group of participants in CPET (n = 534 versus n = 1,174) no significant difference in the levels of physical activity was found (p = 0.241). So far, we agree with J.A. Neder. However, in our view the effect of excluding participants with coexisting and as yet unknown diseases has been shown to have an even more important impact in this subpopulation. Due to results derived within the examination process, 74% of the subjects aged >50 yrs had to be excluded. This clearly shows the importance of a wide spectrum of examination methods beside patients self report to detect and consequently exclude coexisting pathologies. Neder et al. 3 tried hard to establish a disease free study sample, but alike comparable studies to some extent all cardiorespiratory disorders might not have been detected in advance. Furthermore, the majority of studies concerning reference values might be criticised for not being population based and, thus, of limited comparability to our study design. Besides less accurately assessed exclusion criteria, another major shortcoming of other previous studies seems to be the inclusion of individuals who smoke even though an impact of cigarette smoking on exercise capacity assessed by peak V′_O2 and V′_O2 at anaerobic threshold has been shown 4, 5.

One can easily argue that every existing study on reference values for CPET shows limitations. Nevertheless, every piece of work on this issue seems to contribute to close-to-reality normative values. It is incontrovertible that the present findings are the first considering such an amount of different medical examination methods for the establishment of a healthy study sample across a wide age range. However, everybody working on reference values for CPET should be encouraged to do so, since, and at this point we completely agree with J.A. Neder, the “ideal” set of reference values for CPET might still to be generated.

Statement of interest

None declared.