Authors’ Reply

Authors’ Reply

We thank Saynor and colleagues for their comments [1] to our Statement on Exercise Testing in Cystic Fibrosis [2]. Saynor et al. [1] suggest an alternative to the Godfrey protocol for cycle ergometry as best practice. The Godfrey protocol uses 1-min step increments to maximal exertion, with the steps adjusted to complete the test in 8–12 min. Saynor et al. [1] suggest an exercise test protocol instead which consists of two parts: a ramp increase in exercise intensity by 10–25 W/min to volitional fatigue and – after 5 min of active recovery and 10 min of rest – a constant work rate test at 110% of peak power to verify oxygen uptake at peak exercise [3]. They further state that our proposed indicators of a peak patient effort may not allow determining peak oxygen uptake (VO_{2 peak}).

It is important to remember why exercise testing is being recommended. Exercise testing is performed to determine functional capacity and the sources of exercise limitation, to provide guidance and counselling for physical activity, physical interventions and/ or to prognosticate. To accomplish this, the results from a Godfrey protocol symptom-limited exercise test provide us with the information to answer these important questions. We fully agree with Saynor et al. [1] that any cardiopulmonary exercise test protocol should produce valid and reliable results. However, to be recommended for clinical practice, the test protocol also needs to be feasible. Adding a second verification test puts more burden on the patient and the exercise testing facility, which has to be well justified.

The protocol suggested by Saynor et al. [1] has been used in a relatively small (n = 14) group of children and adolescents with cystic fibrosis and mild to moderate lung disease in a research setting so far [3]. Data on adults or patients with advanced pulmonary disease are missing. In contrast, the Godfrey protocol has been used both for clinical and research purposes in large groups of patients all over the world. For our statement, the following criteria for selecting an exercise testing protocol have been applied: ‘The exercise testing protocol should be as universally applicable as possible to allow follow-up of patients over time, facilitate comparisons between centers and enable analyses of large databases in the future, e.g. patient registries. Thus, the protocol should be valid for testing young and older as well as severely impaired and very fit patients, and suitable to address the clinical indications for an exercise test in CF’ [2] (page 5).

The main argument of Saynor et al. [1] is that their complex protocol [3, 4] elicits, only in three patients, a higher VO₂ than a ramp test alone in some patients. However, in their publications, the authors failed to report individual peak heart rate, peak respiratory exchange ratio, etc. during their ramp test, so that it remains unclear how many tests would have been rejected based on the criteria of a maximum effort proposed in our statement [2]. In other words: it might well be that the incremental tests reported by Saynor et al. [3], which resulted in much lower peak VO₂ values than the following constant workload tests in a few patients, would have been invalid (‘no maximal effort’) according to the criteria defined in our Statement. In the case of a non-maximal effort, we recommend to consider repeating the test rather than using the data [2, see flow chart there, page 13]. If this approach had been taken by Saynor et al. [4] to select only valid ramp tests, possibly the reproducibility of VO_{2 peak} determined during such tasks would have improved. Indeed, data from the Children’s Health and Exercise Research Centre in Exeter – the same institution as Saynor et al. – show high reproducibility of VO_{2 peak} measurements obtained during an incremental cycle test in children without a supramaximal validation [5]. The coefficient of variation of the typical error (4.1%) was even lower in this latter study [5] compared to the value (9.3%) reported in the study by Saynor et al. [4], which included a supramaximal validation. Perhaps a secondary analysis of the data collected by Saynor et al. [4] with exclusion of ramp tests without a maximal effort could provide more insight.

The concept of using secondary criteria to validate a maximal effort during an incremental cycling task as recommended in our Statement does not seem to be fully appreciated by Saynor et al. [1]. As clearly described in our Statement, criteria are used following an incremental test to volitional fatigue to substantiate a maximal effort. Since there is large interindividual variation in peak values of heart rate, respiratory exchange ratio (RER), etc., the approach taken by Saynor et al. [1] to look at VO₂ at a given heart rate or RER during an incremental test to find out that the VO₂ is lower than at peak exercise is not surprising and no valid argument against our approach.

Validity of a cardiopulmonary exercise test not only requires a valid measure of VO_{2 peak} but is further substantiated by its association with important clinical outcomes, such as prognosis, quality of life or exacerbations. Interestingly, the respective relationships have been shown in all respective publications cited by Saynor et al. [1] using an incremental exercise test without a supramaximal constant load verification test.

With respect to total exercise time when using the Godfrey protocol with the selection of work rate increments based on patients’ height and FEV₁ as suggested in our Statement, we did never observe exercise times less than 6 min if the effort was maximal. Even when using the equation developed by Hulzebos et al. [6] to calculate work rate increments for an incremental cycle test in children and adolescents, exercise time varied between 7 and 16 min. Nevertheless, the publication has been acknowledged in our Statement [2].

Our Statement on Exercise Testing in Cystic Fibrosis will be reviewed and likely revised in 2020. In that revision, the existing evidence for using an additional supramaximal verification test in the assessment of VO_{2 peak} will be assessed and the information integrated in the document.