Health Status and the Spiral of Decline

Abstract

COPD is complex and there are multiple determinants of poor health, so most studies of Chronic Obstructive Pulmonary Disease (COPD) now include health status, or health related quality of life measurement, along with FEV₁ as a primary endpoint. It is important to make a distinction between quality of life—which is unique to the individual, and health status measurement—which is standardized quantification of the impact of disease. Health status scores correlate weakly with FEV₁ but they are predictors of mortality independently of age and FEV₁. The determinants of impaired health status may change over time and the clinical consequences associated with a given change in score may change with disease progression. The largest study to date suggests that health status decline is linear over time, but longer-term studies are needed to confirm these findings. Inhaled corticosteroids have been shown to reduce the rate of decline of health status, possibly due to exacerbation prevention.

Key words: :

• Chronic obstructive pulmonary disease
• Questionnaire
• St. George's Hospital Questionnaire
• Forced expiratory volume
• FEV
• Health status

INTRODUCTION

The symptoms and outcomes associated with chronic obstructive pulmonary disease (COPD) are highly variable, and a function of the complex range of pulmonary and systemic disease effects. The multiple consequences of COPD, including breathlessness, exercise limitation, muscle wasting, fatigue, often have no immediate relationship to expiratory airflow limitation (Fig. 1; ([1])). For this reason, it is now recognized that no single measure can adequately reflect the nature or severity of COPD ([1]). Thus, most COPD studies now include health status (or “health related quality of life measurement”) along with FEV₁ as an endpoint.

Figure 1 A model of COPD progression and consequences. The multiple consequences of COPD, including breathlessness, exercise limitation, muscle wasting, fatigue, and exacerbations.

It is important to make a distinction between quality of life and measurements of health status. Quality of life impairment results from the unique impact of disease on an individual. By contrast, health status questionnaires provide a standardized method for quantifying the impact of disease on health, wellbeing, and daily life for use in populations of patients. The items that constitute the questionnaire should not contain biases due to age, race, gender and educational level, or individual disease effects that may not apply to every patient. Most questionnaires provide a profile of domain or sub-scores, but the most useful information usually comes from the overall score ([1]). Examples of COPD specific health status questionnaires include the St George's Hospital Questionnaire (SGRQ) which is for both asthma and COPD, ([2]) the Chronic Respiratory Questionnaire (CRQ) ([3]), the Breathing Problems Questionnaire (BPQ), ([4], [5]) and the Quality of Life for Respiratory Illness Questionnaire (QOLRIQ) ([6]). Using these instruments, it is possible to measure long-term rate of decline in health status in COPD ([7]). This article focuses on the SGRQ for this purpose.

Health status, FEV₁, exacerbations and mortality

In clinical trials of the effect of treatment in COPD, the correlation between changes in FEV₁ and health status is weak ([8]). There is a significant correlation between FEV₁ and SGRQ score, but also much scatter around the regression line (Fig. 2). Some patients experience a measurable improvement in FEV₁, but no improvement in SGRQ, whereas others demonstrate large improvements in health status, but no detectable change in FEV₁. The Inhaled Steroids in Obstructive Lung Disease (ISOLDE) trial ([7]) of 751 patients randomized to receive fluticasone propionate or placebo showed that placebo-treated patients reached a clinically significant level of deterioration of 4 units in SGRQ Total score every 15 months. A further analysis of those data showed that the deterioration in SGRQ score was only weakly correlated with the decline in FEV₁ (Fig. 3); exacerbation rate was more closely associated with the worsening in SGRQ ([9]). Exercise capacity, measured using walking tests, is a better correlate of impaired health status than FEV₁ ([2], [10], [11]), although peak VO₂ max during ergometer testing does not correlate well with health status scores ([12]).

Figure 2 Health status and FEV₁. In clinical trials of the effect of treatment in COPD, the correlation between changes in FEV₁ and SGRQ score is significant, but there is much scatter around the regression line. Source: ([8]).

Figure 3 SGRQ and FEV₁ Slopes in the Obstructive Lung Disease (ISOLDE) trial ([9]). Deterioration in SGRQ score was only weakly correlated with the decline in FEV₁. Source for figure data: Spencer S, Calverley PM, Burge PS, et al. Impact of preventing exacerbations on deterioration of health status in copd. Eur Respir J. 2004;23:698–702 ([9]).

Several prospective studies have reported a correlation between baseline SGRQ score and mortality, including the QuESS study ([13], [14]), and those by Domingo-Salvany et al. ([15]) and Oga et al. ([16]). Patients with the worst baseline SGRQ score were more likely to die at one year than patients with lower scores. In a predominantly male population, a multivariate Cox model showed that age was the strongest predictor of death, followed by FEV₁ and SGRQ ([16]). By contrast, in the Domingo-Salvany et al. study, performed entirely in men, the SGRQ was a stronger predictor of mortality than FEV₁ ([15]). When VO_{2 max} was added to a multivariate analysis in the study by Oga et al., this measure then emerged as the most powerful predictor of mortality, and SGRQ and FEV₁ were no longer statistically significant predictors (Table 1) ([16]). This finding supports the hypothesis that the correlation between health status and mortality is probably due to an association between health status and exercise capacity ([16]).

Table 1 Stepwise multivariate cox proportional hazards analysis to investigate the most significant predictor of mortality in male patients with chronic obstructive pulmonary disease

	Relative Risk	95% Confidence Interval	p Value
Peak VO2, ml/mn	0.994	0.992–0.996	< 0.0001
Age, yr	1.077	1.010–1.149	0.024
Postbronchodilator FEV1, % predicted			0.056
DLO2/VA ml/min/L/mm Hg			0.089
Body Mass index, kg/m^2			0.21
Cumulative smoking, pack-years			0.32
SGRQ total			0.42

[Source: (16)].

The VO_2max is a valuable summative measure because it quantifies both the cardiovascular and pulmonary responses to exercise and it also reflects leg fatigue. For this reason, it is perhaps not surprising that it is a good predictor of mortality.

Measurement of health status and disease progression

Health status measurement should be made using instruments that have interval scaling properties. This means that the “severity distance” between scaling points is always the same, regardless of where the points lie along the scale—just like a ruler ([17]). This becomes important to understand when making the distinction between changes in health status score and the consequences of that change. Consider the implications of a change in water temperature. Warming from −1°C to + 1°C, ice melts; with the same incremental change warming from 99°C to 101°C, water boils. The implications of a change in temperature on H₂O depend on baseline temperature.

The same may apply to a change in health status score. The change in SGRQ with an exacerbation has been reported to be 9 units ([18]). If that occurs from a baseline of 30 units (a score that indicates mild disease), it is likely that the patient will just go to see their primary care provider for antibiotics. But exactly the same change in SGRQ score from a baseline of 60 units may be associated with a hospital admission. Although the magnitude of the exacerbation may be the same, its impact (or apparent severity) will depend on the patient's baseline state, not necessarily on the size of the deterioration. A distinction must be made between the magnitude of a change in health status and the clinical consequences of that change. This becomes particularly important when considering long-term changes in health and disease progression.

Longitudinal changes

Clinical trial data suggest that health status decline is linear over time ([7]), but longer-term studies may be needed to determine whether this is the case. It is likely that loss of exercise tolerance is an important determinant of loss of health status, however to date this has not been studied alongside health status in long-term studies.

The estimated rate of deterioration in COPD patients treated with short-acting bronchodilators in the ISOLDE study of 752 patients randomized to receive fluticasone propionate or placebo was 3.2 units/year and 2.0 units/year in patients receiving inhaled corticosteroid ([7]). These estimates may be skewed by early withdrawal of patients with worse health or faster decline (Fig. 4) ([19]), which may lead to an underestimate of the true rate of deterioration. This may occur particularly in cohort studies in which each patient remains in the study for only part of the follow-up period ([20]).

Figure 4 Withdrawal from treatment as an outcome in the ISOLDE study of COPD. The estimated rate of deterioration in COPD may be skewed by early withdrawal of patients with worse health or faster decline ([19]).

A longitudinal study of time spent each day outside the home by a cohort of 1476 patients with moderate to very severe COPD revealed that, over an 8 year period, the total SGRQ score rose (i.e., worsened) by 1.94 unit/year, the activity score rose by 2.1 units/year, and the impact score rose by 3.0 units/year. Surprisingly, the symptom score actually fell (improved) by 2.0 units/year ([20]). These longitudinal data must be interpreted with caution, however, since the median period of time spent in the study by any given patient was less than 3 years, so the data were collected in a continuously changing cohort of patients over the study's 8-year timeframe. Under these circumstances, differential dropout of the sicker patients may heavily influence the calculated change in scores. This may explain the reported long-term improvement in symptoms score.

Figure 5

Effects of treatment on health status

One-year studies have shown evidence of improved health status, as assessed by the SGRQ. For example, in a double-blind double-dummy studies of tiotropium compared with ipratropium study of tiotropium in chronic COPD patients, at one year the mean change from baseline in SGRQ total score was −3.7 units for tiotropium and −0.4 for ipratropium ([21], [22]). A significantly greater percentage of patients in the tiotropium group (49%) showed at least a 4-unit improvement in SGRQ total score compared to those in the placebo group (30%) ([22]). The mechanism for the improvement may be muti-factorial and time-dependent. Initial benefits may be due to an effect of treatment on lung volumes and dynamic hyperinflation, but a continued or accumulating benefit could be due to prevention of exacerbations ([21], [22], [23], [24], [25]). In the ISOLDE study, the reduced rate of deterioration in SGRQ scores with fluticasone propionate over three years does appear to be due to prevention of exacerbations ([9]) In that study, however, whilst treatment produced a reduction in rate of deterioration of SGRQ scores, the scores after three years of treatment were still clearly worse than at baseline ([7]). By contrast, in a recent 3-year study of salmeterol combined with fluticasone propionate, health status scores after three years were better than at baseline ([26]).

CONCLUSIONS

Health status impairment is the result of multiple pathways, including inflammation, tissue damage causing disturbed structure and function, and exacerbations that contribute to the symptoms that cause impaired health status. It is important to remember that impaired health status does not cause overall impairment, but only reflects it. Nonetheless, impaired health status may have an effect on mortality through an impact upon patient behavior. For example, it is much easier for the breathless patient to open up a package of convenience food with high salt content, low anti-oxidants and high fat than to take the time and effort to prepare healthy food. As COPD progresses, patients fail to exercise, feel depressed, and experience low self-esteem. It is very likely that not only does impaired health status reflect underlying patho-physiological processes, but also that impaired well-being contributes to those processes.

ACKNOWLEDGMENTS

This article was prepared with the editorial assistance of Genevieve Belfiglio, a medical writer working with Advanced Studies in Medicine. This paper is based on a presentation at a meeting, titled “Influencing the Spiral of Decline in COPD,” which took place in Atlanta, Georgia, on May 2–4, 2007. The author is responsible for the content of the article but gratefully acknowledges assistance from Dick Briggs Jr, MD, University of Alabama at Birmingham, Birmingham, AL. This article was funded by Boehringer-Ingelheim Pharmaceuticals, Inc and Pfizer Inc.

Conflict of interests: Dr Jones has received speakers' fees and/or served as a consultant for Boehringer-Ingelheim Pharmaceuticals, Inc.