Complexity of chronic asthma and chronic obstructive pulmonary disease: implications for risk assessment, and disease progression and control

Abstract

Although assessment of asthma control is important to guide treatment, it is difficult since the temporal pattern and risk of exacerbations are often unpredictable. In this Review, we summarise the classic methods to assess control with unidimensional and multidimensional approaches. Next, we show how ideas from the science of complexity can explain the seemingly unpredictable nature of bronchial asthma and emphysema, with implications for chronic obstructive pulmonary disease. We show that fluctuation analysis, a method used in statistical physics, can be used to gain insight into asthma as a dynamic disease of the respiratory system, viewed as a set of interacting subsystems (eg, inflammatory, immunological, and mechanical). The basis of the fluctuation analysis methods is the quantification of the long-term temporal history of lung function parameters. We summarise how this analysis can be used to assess the risk of future asthma episodes, with implications for asthma severity and control both in children and adults.

Figure 1

Reductionist (A) and complex-systems (B) approaches

Figure 2. Fractal properties of structures (A) and time series (B)

Shown is an idealised fractal structure and time series in which the fluctuations at different spatial and time scales are similar, respectively.

Figure 3. Avalanche-like opening of the lung during slow inflation

This behaviour can be simulated in a three-dimensional model of the airway tree before (left) and after (right) a simultaneously avalanche-like opening of many segments. White and red are closed and open segments, respectively.

Figure 4. Fluctuations in twice-daily peak expiratory flow (PEF)

Representative time series of twice-daily PEF seen during 6 months, showing self-similar fluctuations with similar variability at different time scales. The inset graph shows a shorter time scale in which the statistical properties of the PEF series are similar to those of the entire series. Despite the random-looking appearance, the fluctuations are not random but ordered, which means that any particular value is dependent on previous values.

Figure 5. Quantification of risk prediction of future obstructive events

(A) Results calculated with the method proposed by Frey and colleagues for both short-acting and long-acting bronchodilator treatments. The mean values of the long-range correlation exponent α for the two groups are provided. (B) Computer simulations based on the model proposed by Frey and colleagues show how the conditional probability π varies with the long-range correlation exponent α. The different curves correspond to different values of peak expiratory flow (PEF). For example, the 95% predicted PEF curve is obtained by assuming that the present PEF is 95% of its predicted value, and the probability π that PEF is less than 80% of the predicted PEF in the next 14 days is computed for a range of α values. The curves show that the risk of future obstructive events with low PEF is very dependent on α.

Figure 6. Models of classic and possible future ideas for monitoring exacerbations

(A) Classic monitoring method. (B) Multidimensional approach. (C) Fluctuation analysis approach.