Abstract
Introduction: Detrended fluctuation analysis (DFA) quantifies the rate of decay in self-correlation in a time series with an exponent “alpha” that is related to COPD exacerbation frequency. The relationship between α of different spirometric parameters in the same individual on the same days is not known.
Methods: We examined data from the London COPD cohort on 28 COPD patients who had recorded both FEV1 and FVC and PEF on daily diary cards for 300 days. Measurements were made after medication in the morning. At recruitment, these patients had a mean age (SD) 65.3 (9.3) years; FEV1 1.08 (0.36) l, FEV1 % predicted 37.3 (14.1), FEV1/FVC ratio 0.43 (12.8). DFA has been described (Frey et al Nature. 438: 667-70, 2005). The analysis was also repeated with data collected during exacerbations removed
Results: The patients had an α of 0.97 (SD 0.22) for PEF, 0.93 (0.22) for FEV1 and 0.95 (0.23) for FVC. No differences was seen in any of the estimates (p>0.24). Figure 1 illustrates the relationship between alpha for FEV1 and PEF (r=0.69; p<0.001). There was no significant difference if data collected during exacerbation data were excluded, PEF α was 0.95 (SD 0.19); FEV1 α = 0.95 (0.19) and FVC α = 0.97 (SD 0.20).
Conclusion: In COPD patients, long-term correlations (self-similarity) exist in daily FEV1, FVC and PEF. The estimates of α are similar and linearly related to each other.
- © 2012 ERS