A simple assessment of dyspnoea as a prognostic indicator in idiopathic pulmonary fibrosis

RESULTS

93 patients were included in the study, consisting of 81 males and 12 females. The mean±sd age at diagnosis was 66.3±8.1 yrs. Their clinical and physiological data are summarised in table 1. The mean observation period was 40.6±25.8 months. In this observation period, of 49 (52.7%) patients, 18 (19.4%) died due to respiratory failure, 13 (14.0%) due to acute exacerbation of IPF, 5 (5.4%) due to pneumonia, 5 (5.4%) due to a cancer, 4 (4.3%) due to other known causes and 4 (4.3%) due to unknown causes. 45 patients died in our hospital, so the exact dates of their death could be obtained. We could also obtain the dates of death for four patients who died in other hospitals, although the reasons for their death were not known. Distribution of baseline-modified MRC dyspnoea scores and baseline results of the 6-min walk test are shown in table 1.

The results of univariate Cox proportional hazard model are shown in table 2. FVC % predicted (hazard ratio (HR) 0.965, 95% CI 0.948–0.982; p<0.0001), D_L,CO % predicted (HR 0.978, 95% CI 0.963–0.993; p = 0.0041), baseline arterial oxygen tension (P_a,O2) (HR 0.963, 95% CI 0.938–0.989; p = 0.0060), and modified MRC score (HR 2.402, 95% CI 1.495–3.858; p = 0.0003) were revealed to be significantly predictive of survival. All variables of the 6-min walk test including the walk distance (HR 0.995, 95% CI 0.992–0.998; p = 0.0020), the lowest arterial oxygen saturation measured by pulse oximetry (S_p,O2) (HR 0.944, 95% CI 0.918–0.972; p<0.0001) and the Borg scale (HR 1.285, 95% CI 1.091–1.514; p = 0.0027) were also significant. There were no highly correlated (correlation coefficient ≥0.6) variables among them. In stepwise multivariate Cox proportional analysis, modified MRC score (HR 2.181, 95% CI 1.333–3.568; p = 0.0019) and the lowest S_p,O2 (HR 0.952, 95% CI 0.924–0.981; p = 0.0014) were selected as most significant (table 3). In normal (not stepwise) multivariate Cox proportional analysis, including modified MRC score, the lowest S_p,O2 and walk distance at the 6-min walk test, and D_L,CO % predicted, as independent variables, the results were as follows: modified MRC score (HR 1.903, 95% CI 1.117–3.239; p = 0.0178) lowest S_p,O2 (HR 0.951, 95% CI 0.913–0.990; p = 0.0146) walk distance (HR 0.999, 95% CI 0.995–1.002; p = 0.4105) and D_L,CO % predicted (HR 1.000, 95% CI 0.980–1.020; p = 0.9839).

Kaplan–Meier survival curves based on the modified MRC score and presence of desaturation (lowest S_p,O2 <88%) during the 6-min walk test are shown in figure 1. The median survival estimates for patients with a modified MRC level of 0–1, 2 and 3 were 66.7, 30.9 and 10.2 months, respectively. The median survival estimates for subjects with and without desaturation were 41.0 and 62.9 months, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Kaplan–Meier survival curves according to a) the modified Medical Research Council (MMRC) score (—: MMRC 0–1; ---: MMRC 2; ······: MMRC 3) (p<0.0001) and b) the lowest arterial oxygen saturation measured by pulse oximetry (S_p,O2) during the 6-min walk test (—: S_p,O2 ≥88%; ---: S_p,O2 <88%) (p = 0.023). Survival curves were compared with log-rank statistic. +: censored.

DISCUSSION

We confirmed that dyspnoea in daily living, assessed using the modified MRC scale, as well as the lowest S_p,O2 at 6-min walk test, were the strongest predictors of survival in patients with IPF. Dyspnoea is one of the most fundamental and debilitating symptoms of patients with IPF. The results in the present study indicate that evaluation of the dyspnoea level when diagnosing patients with IPF is important to predict their survival. Moreover, it was revealed that, no matter how well-preserved or bad the pulmonary physiology, the survival of a more dyspnoeic patient is shorter.

Many previous studies have shown that several physiological parameters predict survival in patients with IPF 4, 7. FVC, total lung capacity (TLC), FEV₁, D_L,CO, and P_a,O2 at diagnosis, as well as longitudinal changes in these physiological parameters, were shown to predict survival 10–13. Furthermore, parameters obtained from the 6-min walk test, such as walk distance and presence of desaturation during the test were recently shown to be prognostic in IPF 8, 9. A recent study showed that low oxygen uptake during cardiopulmonary exercise testing also predicts worse survival 21. In the current study, FVC, D_L,CO, P_a,O2, and variables from the 6-min walk test including walk distance, the lowest S_p,O2, and the Borg scale, were revealed to be significant predictors of survival in the univariate model. These results are consistent with previous reports 4, 7–13. In addition to predicting parameters demonstrated in earlier studies, it is an important finding of the present study that dyspnoea in daily living was a stronger prognostic parameter than most physiological markers at diagnosis. It indicates that dyspnoea in daily living should be carefully evaluated and included in the assessment of patients with IPF, even though such assessments have not focused on like physiological parameters to date.

In previous COPD-related studies, dyspnoea in daily living has already been recognised as an important prognostic factor. In a 5-yr, prospective, multicentre trial of 227 patients with COPD, survival was not significantly linked to COPD staging based on FEV₁, whereas dyspnoea measured by the MRC scale predicted their survival 22. Celli et al. 16 developed a multidimensional grading system to predict survival of patients with COPD, called the BODE (body mass index, degree of airflow obstruction, dyspnoea, exercise capacity) index. The BODE score was reportedly better than pulmonary function data at predicting the mortality risk in COPD. Dyspnoea in daily living is one of the important components of the BODE index. Thus, dyspnoea has already been recognised as an important factor in predicting survival and has been included in COPD evaluations. IPF is a disease that leads to a number of morbidities or sequelae, such as dyspnoea and exercise limitation, like COPD. Multidimensional evaluations that include dyspnoea are also becoming important in IPF. Actually, dyspnoea impairs the life of patients with IPF; it measurably worsens over time 23 and appears to be one driver of health-related quality of life 24. Therefore, evaluating dyspnoea in daily living should be recognised as important in general practice as well.

There is some confusion in the evaluation of dyspnoea. In terms of the modified MRC score, two types of scale have been used. One has five grades for rating the level of dyspnoea 14, 15 and the other has six grades 25. A scoring system grading 21 levels of dyspnoea was used in multistep scores (CRP score; clinical, radiological and physiological score) in terms of IPF 4. The baseline dyspnoea scale has also often been used in previous studies related to COPD; it uses five grades (0–4) for each of the categories, functional impairment, magnitude of task, and magnitude of effort (0–12 for total score) 26. The University of California San Diego Shortness of Breath Questionnaire has also sometimes been used 27. This questionnaire comprises 24 items that assess dyspnoea over the preceding week. However, these scores are too cumbersome to apply in clinical practice. Although the Borg scale, which lists phrases increasing subjective intensity of breathlessness from 0 to 10, is often used, it is for assessing dyspnoea on exertion 28, not in daily living. We used the modified MRC scale with four grades of dyspnoea because it is easier to rate and is also used in the BODE index developed for COPD. A complicated method for grading is not advisable for practical use. Our data suggest that a simple evaluation of the dyspnoea level can predict mortality in patients with IPF, without the need for complex calculations.

The association of dyspnoea with survival in IPF has been reported in some previous studies. Turner-Warwick et al. 29 showed a significant relationship between dyspnoea assessed with a five-point (0–4) scale and survival in cryptogenic fibrosing alveolitis, although the details of the assessing method were not mentioned. However, in that study, the diagnosis was not based on the current guidelines, and it is highly possible that patients with nonspecific interstitial pneumonia (NSIP), whose survival must have been better than that of patients with IPF, were included at that time. Moreover, ∼30% of patients had connective tissue disease. In contrast, the strength of our study is that patients were limited to those with IPF. Patients with connective tissue disease were excluded, and patients with NSIP were also excluded because the diagnosis was made according to the current consensus statement. King et al. 4 reported that a dyspnoea scale which consisted of 21 points (0–20) was significantly related to survival of patients with IPF. However, a 21-point scale is quite inconvenient to rate and not suitable for clinical practice. Only one report has demonstrated the efficacy of the MRC dyspnoea score for predicting survival in IPF; it consists of a relatively small series of patients with IPF 30. Moreover, the MRC score used in that study had six grades for dyspnoea. In contrast, we chose the modified MRC score, which consists of five grades of dyspnoea because it is quite popular and has often been used in COPD. The present study is the first report, to our knowledge, to demonstrate that the five-point modified MRC scale predicts survival in IPF. Another novel point about the current study is that we included variables from a 6-min walk test, which has been proven to be important in predicting survival of patients with IPF. Then, the modified MRC score and the lowest S_p,O2 at a 6-min walk test were then selected as significant. We confirmed the importance of dyspnoea in daily living in 93 patients. Interestingly, it was more predictive of survival than physiological measurements. Thus, the dyspnoea level should also be assessed and included in further research on IPF. In both research and clinical practice, the five-point modified MRC is quite useful because of its power to predict survival and its simplicity.

The lowest S_p,O2 during the 6-min walk test was found to be another significant predictor for survival in the present study. It is already evident that the presence of desaturation is related to a severe prognosis in some previous studies 10. We adopted the level of desaturation as a continuous variable, not as a categorised variable divided by the presence of desaturation. Thus, the present study demonstrated that severe desaturation leads to a worse prognosis.

The strengths of this study include the relatively large sample size and the relatively high mortality; more than half of all the patients died, making the results of analysis reliable. The limitations of this study are as follows. First, there is a the lack of serial changes in physiological parameters in the analysis. Longitudinal changes in physiology reportedly predict survival 10–13. Another result could be obtained if serial changes in physiological parameters were included in the analysis. However, serial changes in physiology can not be measured at diagnosis. Predicting survival from parameters at diagnosis is indeed crucial. Secondly, our results might have been different if baseline haemodynamics had been included in the analysis, but unfortunately these data were not available, because it is reported that pulmonary hypertension significantly impacts survival in IPF 31, 32. Thirdly, the results are not necessarily applicable to patients with supplemental oxygen, because patients with supplemental oxygen were excluded from the study. Finally, the retrospective study design is a limitation. Collecting additional prospective data is warranted to confirm our findings.

In all probability, survival of patients with IPF will be prolonged if the dyspnoea level in daily activity improves. No pharmacological therapy has been shown to improve dyspnoea or to prolong life in IPF. However, pulmonary rehabilitation that consists of exercise training reportedly improves dyspnoea 33, but this claim is not indisputable 34. Further research must determine whether pulmonary rehabilitation improves dyspnoea and, consequently, prolongs survival in IPF. The effect of oxygen therapy should also be examined. If oxygen ameliorates the level of dyspnoea in daily living and prevents desaturation in patients with IPF, long-term oxygen therapy might prolong survival. To test this hypothesis, further prospective study is needed.

In summary, the present study demonstrates that dyspnoea in daily living, assessed with the modified MRC score and the lowest S_p,O2 during the 6-min walk test, provides stronger prognostic prediction than other physiological parameters in patients with IPF. This assessment tool for dyspnoea is beneficial because of its power to predict survival and its simplicity.