Reproducibility of the 6-minute walk test in lung transplant recipients

This study was conducted at the outpatient department of Physical Medicine, Rehabilitation and Occupational Medicine (PMROM) of the Vienna Medical University Hospital. All LuTXr who participated in this study were tested three times, briefly before discharge from the acute hospital stay (baseline), retested 1–2 workdays later (retest reliability), and after completion of 2 months of postacute rehabilitation (sensitivity to changes). The measurements of the 6MWT were taken on a straight 50 m indoor corridor located within the PMROM department. Three assessors who had received standardized training by a senior clinical specialist, and who have extensive experience in conducting function testing in LuTXr shortly after surgery, performed all the experiments. The training included standardized verbal instructions, standard operating procedures and supervised practice. Assessors performed all tests under guidance by the senior clinical specialist.

Over the course of a 2.5-year period, a cohort of 30 consecutive non-Austrian LuTXr who had undergone first time LuTX at the department of thoracic surgery, Vienna Medial University were referred for subacute outpatient rehabilitation to the PMROM department and asked to participate in this study. All of these patients agreed. Another 20 sex-matched and age-matched Austrian LuTXr who were awaiting discharge from acute hospital stay at the department of thoracic surgery were also invited and agreed to participate. Thus, a total of 50 (28 females) LuTXr who had undergone first time transplantation were included in the study. This recruitment strategy was applied as foreign citizens who underwent LuTX in Austria were not eligible for inpatient rehabilitation in Austria. By contrast, comprehensive postacute inpatient rehabilitation is strongly recommended and provided for Austrian citizen LuTXr. All participants provided written informed consent and were encouraged to ask questions regarding the study. The ethics committee of the Medical University of Vienna approved the study.

Patients included in the study had to have undergone transplantation of one or two lungs, to be able to stand without support for a minimum of 5 min, and to walk with or without an assisting device for a minimum of 50 m. Exclusion criteria were psychiatric disorders, peripheral neurologic deficits in the lower extremities (except peroneal compression neuropathy), and severe neurologic diseases. Physicians specialized in Physical Medicine and Rehabilitation examined eligible patients. If patients were unable to understand German or English, a translator was made available.

During the entire acute hospital stay, all LuTXr underwent daily rehabilitation. Immediately after discharge from acute hospital care, Austrian citizens (n = 20) received 4–6 weeks of inpatient rehabilitation approximately 80 km away from Vienna, whereas foreign LuTXr (n = 30) participated in comprehensive outpatient rehabilitation at the PMROM department, which had similar content as inpatient rehabilitation but less supervision. Both the subacute inpatient and outpatient rehabilitation interventions were tailored according to the individual needs of the LuTXr. Relevant components of the rehabilitation program comprised 1) regular cardiopulmonary endurance training, 2) therapeutic exercises to improve muscle and body balance functions, and 3) regular respiratory therapy to improve respiratory muscle function including the breathing pattern and optimizing clearance of sputum. Nutritional and psychological counselling was mandatory for inpatients and on demand for outpatients.

The distance walked was assessed in meters (6MWD) and its predicted walking distance (6MWD%_pred) was calculated for males and females separately according to published reference equations [19]. From the 100 m split distances, walking speed intervals were estimated for each minute of the 6MWT and the respective dynamics determined by calculating the percentage change in speed for each consecutive minute covered compared to the first minute, starting from the second minute [(speed min_n − speed min₁/speedmin₁) × 100%].

With a minimal expected intraclass correlation coefficient (ICC) of 0.85 and a hypothesis that the present findings would be consistent with a minimum ICC of 0.9, a minimum sample size of 38 individuals was required to achieve a level of significance of 0.5 and a power of 0.8 (β = 0.2). Because 20% of participants usually refuse a second testing, we sought a sample size of 50.

All statistical analyses were performed using the R environment for statistical computing [20]. Distribution of the 6MWT components were assessed and appropriate reliability indices were compiled using previously suggested [21, 22] data inspection procedures. The following aspects were explored for the variables obtained from the first 2 test days:

Paired t‑tests calculated the 6MWT score changes between baseline, the first re-test and completion of rehabilitation. Receiver operating analyses (ROC) served to explore the ability of the 6MWT scores to detect changes that occur with rehabilitation in an accurate way. Responders and non-responders were classified as follows:

Measurement error among all the participants amounted to 34 m for the 6MWD and 3.9% for the 6MWD%_pred. Both the SRD of the 6MWD and the 6MWD%_pred calculated for the entire group were 79 m (mean: 79 m) and 11%, respectively (Table 3). The ICCs achieved excellent between test days consistency of the individual 6MWD scores ranking (Table 3). Subsequent age-specific sub-analyses revealed numerically higher SEM and corresponding SRD values for both 6MWT scores in younger as compared to older LuTXr (Table 3).

Bland-Altman plots indicated a systematic bias between the two baseline measurements, and the magnitude of agreement appeared to decrease when 6MWD increased. The 95% CI limit of agreement of the 6MWD and the 6MWD%_pred were between 52–106 m and 7–15%, respectively (Fig. 2).

When compared to baseline, the 6MWD significantly increased by 174.2 m (95% CI: 150–198) and its 6MWD%_pred score by 26% (95% CI: 23–29) upon completion of rehabilitation. ROC analyses revealed the 6MWT scores as highly accurate in detecting both a minimum and a true change because of postacute rehabilitation. This was true for changes with rehabilitation as assessed from baseline or baseline retesting. The respective AUC values exceeded 0.89 and are provided in Fig. 3.

This study’s major findings revealed: 1) a systematic change in the mean between test and retest; 2) ICC values indicating excellent relative reliability; and 3) SRD values that allowed detection in 6MWT score changes observed at the end of postacute rehabilitation.

A previous systematic review found the 6MWD to be a reliable measure in CRD with respective ICCs ranging from 0.72 to 0.99 [1]. Data from our study in LuTXr suggest that the 6MWD and 6MWD%_pred scores are as reliable as in CRD. In this study, the mean ICCs of the 6MWD scores and their respective 95% confidence intervals exceeded 0.89 and are proof of a high degree of consistency and agreement in LuTXr shortly after LuTX surgery, if the 6MWT is repeated on a second test day.

Despite the fact that LuTXr are familiar with the 6MWT, we observed an improvement in 6MWD of a mean of 27 m (95% CI: 14.6–39.4) during the baseline retest. This is surprising, as such changes in the mean are typically observed when patients are either unfamiliar with the test or have not performed it for a sufficiently long period of time. By comparison, in novice test subjects an improvement in 6MWD on retest was found to be 19 m in patients with pulmonary hypertension [25], and 26.3 m amongst COPD patients [1]. Derived from the SEM, the smallest clinical important difference (CID) for the 6MWD in LuTXr on an individual level would be 28.5 m. This estimate, which is similar to the CID scores determined for CRD (29–34 m [26], 30 m [1], 29–42 m [27]), is close to the changes in the mean observed on retesting in LuTX. This result strongly points to the value of repeating the test on a second occasion in LuTXr shortly after transplantation, regardless of their previous experience with the 6MWT. This view could be further corroborated by the parallel increase in perceived physiological effort with the improved 6MWD scores in LuTXr at baseline reresting, a result which was not observed in research amongst CRD patients [25]. Thus, LuTXr seem motivated to perform higher on a 6MWT retest occasion despite the standardization of the verbal encouragement provided. Greater confidence in the “new” lung and more appropriate subjective interpretation of shortness of breath because of deconditioning rather than need of oxygen supply, and, potentially, a more positive experience with the testing on the first test day may all serve as explanations. This may be particularly true for older as compared to younger LuTXr, as they were found to have more pronounced learning effects in our secondary, age-specific sub-analysis. Consequently, the second test would better reflect the LuTXr’s true exercise capacity. It is unlikely that a standardized psychoemotional intervention could minimize a potential learning effect in LuTXr, particularly in day-to-day clinical practice, where repeating the 6MWT on a second test day is typically not feasible. It is worth noting that, like in COPD, a learning effect in LuTX patients who covered less than 300 m was absent. Amongst these individuals, physical limitations of exercise performance likely outweighed the psychoemotional ones.

In this study we measured intermittent claudication-like leg pain during walking, which was observed in 64% of our LuTXr when tested at baseline. This type of pain typically occurred in the shanks and/or thighs after 2–3 min of walking and its intensity was rated to be moderate to severe at the end of the 6MWT. It was notable that none of the LuTXr suffered from peripheral occlusive disease. Nevertheless, the LuTXr walking speed was well maintained throughout the test, as illustrated in Fig. 1. It therefore seems likely that claudication-like leg pain does not affect the 6MWD scores in a relevant way. Although underlying mechanisms causing this activity-related pain in LuTXr remain widely unknown, relative overuse of the weight-bearing working muscles during walking best explains this type of pain. Therefore, overly accumulating metabolites associated with excessive muscle fatigue would elicit pain by activating type III and IV nerve endings located within muscles and surrounding tissue. This could be due to disuse-related, impaired muscle fiber metabolic capability (of both type I and II fibers) [28] and/or a relative lack of oxygen supply to the working muscle. A lack of oxygen supply would be associated with a presumptive loss in muscle fiber capillary density, resulting in impaired matching between oxygen delivery and oxidative metabolism [29, 30]. In LuTXr, such disuse associated muscle metabolic impairments during exercise could further be aggravated by adverse effects of immunosuppressive medication to mitochondria and/or to the expression of type II muscle fibers [31‐33].

Perceived exhaustion is widely measured at the end of the 6MWT as it provides an estimate for the degree of deconditioning of LuTXr. Instead of 6MWT BORG scale measures of dyspnea in chronic lung disease, we collected perceived exhaustion BORG scale ratings as this variable seems more appropriate in LuTXr early after LuTX. In LuTXr oxygen saturation levels were within normal ranges (> 90%) throughout the test. Walking speed in the 6MWT is self-adjusted and the motivation to walk for 6 min as far as possible is known to vary considerably between patients. Considering all this, assessments of BORG ratings may aid in validating the 6MWT scores obtained, and furthermore appear to be particularly useful if 6MWT scores represent normal or close to normal values. In such cases, the degree of exhaustion would allow estimation of how vigorously the test subject performed the 6MWT. It is important to note that the variability of subjective BORG scale ratings is higher than that observed from objective measures [34]. Thus, these scores may not yet qualify for inclusion in a complex 6MWT score, unless fully validated.

For clinicians it is of utmost importance to know whether the 6MWT is sensitive to interventions. This study revealed that the changes in distance walked upon completion of rehabilitation clearly exceeded the SRD 79 m. Considering our favorable findings from ROC analyses, the SRD appears small enough to allow identifying improvements in 6MWD at the end of postacute rehabilitation or therapeutic exercise interventions within the first year of rehabilitation [35‐37]. Such favorable AUC values seemed not to be affected even when the 11 participants lost to follow-up upon completion of rehabilitation were considered. This observation is indirectly supported by both a) the non-significant differences in baseline 6MWT scores between those who completed the study and those lost to follow-up, and b) the reports that neither a low functioning and health states nor death was the main reason of not showing up. Conversely, the tight schedule of the regular comprehensive medical assessments at the thoracic surgery department along with the burden of travelling very long distances of several hundred kilometers on one day were primary reasons for not participating in follow-up assessment. However, the 6MWD may not be sufficiently sensitive if the improvements to inpatient rehabilitation need to be identified 5 years after LuTX [38]. Of note, in this study the majority of LuTXr had already suffered from bronchiolitis obliterans. It is worth noticing that our study also included young patients. Because their postoperative 6MWD%_pred (but not absolute 6MWD) scores were significantly lower at baseline, they might improve faster and to a greater extent after surgery because of their stronger physiological reserve capacity [39]. However, considering the smaller SEM and SRD values observed from our older LuTXr group, responses to postacute rehabilitation would likely be equally identified with the 6MWD%_pred scores in younger and older LuTXr.

We studied retest reliability in LuTXr when discharged from the acute hospital after LuTX. Although the mean 6MWD score values after discharge from the acute hospital stay are smaller than in a later phase of LuTX, the dispersion of scores may be expected to be similar or even broader. This would boost the ICCs observed without affecting the 6MWD’s sensitivity to identify relevant changes. Additionally, when reliability is studied in an early phase after LuTX, a learning effect might become more overt despite an individual’s familiarity with the test. This interpretation suggests that our results, which are related to the reliability of the 6MWT in postacute phase after LuTXr, cannot be extended to later phase LuTXr. However, another study that assessed the mean and SD of the 6MWD 1 year [40] and 5 years after LuTX [38] suggests that the reliability of the 6MWD would not change in a relevant way if investigated in a later phase after LuTX. In these studies, the mean 6MWD scores (523 m at 1 year [40] and 490 m at 5 years[38] after LuTX) were clearly higher than in ours, but the dispersion of scores as expressed by their SD was comparable. As learning effects were found to be present and of a similar extent as early as 3 months after a 6MWT in CRD [41], this evidence suggests that the metric properties of the 6MWT could be administered to LuTXr at any phase after LuTX. However, future studies will have to clarify this point.