The most frequent COPD phenotype in this unselected population in the rural setting of Malaysia was NON-AE, followed by the AE-CB, ACO and AE NON-CB. Patients with AE were significantly older and smoked more cigarettes, while patients with ACO were predominantly female. Regardless of the COPD phenotypes, biomass fuel exposure was a common risk factor of COPD among them. Close to two-thirds of the patients were exposed to biomass fuel, mainly due to the seasonal open burning in agriculture activities.
The HRQoL of patients with AE and ACO was markedly impaired compared to normal individuals. Meanwhile, the HRQoL of patients with NON-AE was reduced when measured by SGRQ-c but not by CAT. The worst HRQoL was reported in patients with AE followed by those with ACO. The HRQoL of patients with AE was significantly worse than that of ACO and NON-AE while the HRQoL of ACO patients was significantly worse than the HRQoL of NON-AE patients. A similar pattern was also observed in each item of CAT and each component of SGRQ-c, except that the differences were not significant in cough, sputum, and sleep for AE versus ACO, as well as cough and daily activity limitation for ACO versus NON-AE. This lack of significance could be due to the smaller sample size of ACO, or the diurnal variation in symptomatology of bronchial asthma which is commonly associated with cough and sputum production.
The distribution of COPD phenotypes in the present study was almost similar to that of western populations,(24 – 27) except that AE NON-CB is less commonly reported than ACO.(28) So far, only two other studies have reported AE CB is the commonest COPD phenotypes followed by NON-AE, AE NON-CB and ACO. The first study was conducted in primary care centres of the Russia Federation,(29) while the second study involved selected COPD patients in the respiratory clinic of a tertiary hospital.(30) Our findings of patients with AE being older and smoked more cigarettes,(25, 27, 28, 31) as well as more female patients with the ACO phenotype are in agreement with other studies.(24, 26 – 28) The finding that the HRQoL of COPD patients was more impaired in the phenotype sequence of NON-AE, ACO and AE is consistent with the findings of previous studies.(24, 26 – 28, 32) Patients with AE are consistently highlighted as having the worst HRQoL,(24, 26 – 28, 31, 32) while those with NON-AE have the best HRQoL.(25, 29) Of patients with AE, Miravitlles et al,(28) Cosio et al,(31) Kania et al,(27) and Chai et al,(30) reported those with AE-CB have significantly worse HRQoL compared to other COPD phenotypes (all p < 0.001); while Corlatenau et al reported the worst HRQoL in patients with AE NON-CB.(32) The CAT was uniformly used to asses HRQoL in these studies, with the latter two studies also using the SGRQ-c questionnaire. Only this study and that by Miravitlles et al,(28) show patients with ACO have significantly worse HRQoL than those with NON-AE.
Exacerbation is the prognostic hallmark of COPD. Frequent exacerbation is associated with poor HRQoL,(33) decline in lung function,(34) recurrence of exacerbations,(33) recurrent hospitalisations,(35) and increased mortality.(36) Seemungal et al and Mackay et al, respectively reported COPD patients with ≥ three exacerbations (SGRQ-c, p < 0.001) and ≥ two exacerbations (CAT, p = 0.025) per year have significantly worse HRQoL.(33, 37) Cheng et al also reported COPD frequent exacerbators have significantly worse HRQoL (mMRC, p < 0.001; CAT, p < 0.001) compared to non-frequent exacerbators.(38) Therefore, this explains the significantly worse HRQoL among our patients with AE. Despite similar exacerbation frequency, our patients with ACO had significantly worse HRQoL than those with NON-AE which highlights that COPD subtypes can also affect the patients’ HRQoL. Miravitlles et al and Hardin et al, respectively reported COPD patients with BA have significantly worse HRQoL than those without [(mMRC, p = 0.008; SGRQ-c, p < 0.001), and (SGRQ-c, p = 0.008), respectively.(39, 40) Such a finding is not surprising in view of the presence of two different inflammatory processes in ACO.
The findings of our study support the recommendation of GesEPOC to phenotype every COPD patients based on their exacerbation frequency and COPD subtypes.(15) Besides, this study also highlights that exacerbation frequency supersedes COPD subtypes in determining the patients’ HRQoL. Therefore, clinicians should manage COPD patients with frequent exacerbations more aggressively, and consider prescribing pharmacotherapies such as long-acting muscarinic antagonist (LAMA), LAMA and long-acting ß2-agonist in combination, inhaled corticosteroids (ICS), roflumilast, macrolide, or N-acetylcysteine according to the COPD phenotype.(1) COPD treatment should also be personalised according to COPD subtypes, such as ICS for ACO, roflumilast for CB, and medical or surgical lung volume reduction for emphysema.(1)
The present study is among the few in Asia that compares the HRQoL of COPD patients based on different clinical phenotypes. All the patients in this study were from the rural area. Their characteristics are very different from previous studies, such as having a high incidence of significant exposure to biomass fuel, required good physical fitness for agriculture activities, and had limited access to more expensive or newer COPD medications. Besides, we evaluated the HRQoL by using different HRQoL assessment tools and compared each of the subitem or component. By doing so we aimed to assess the patients’ HRQoL in more dimensions and to minimise biases.
There were several limitations in this study. Firstly, the number of AE NON-CB patients was disproportionally small and therefore we were unable to analyse it independently. We added AE NON-CB to AE-CB, and analysed in the line of AE for HRQoL analysis. Secondly, the direct comparison of CB versus emphysema subtypes was not possible because of the first limitation. Thirdly, the AE NON-CB phenotype was based on the finding of air-trapping on physical examination and on chest X-ray. Static lung volume measurement of functional residual capacity, residual volume and total lung capacity as well as non-contrast-enhanced thoracic computed tomography scan acquired at full inspiration and expiration that is able to differentiate emphysematous from non-emphysematous air-trapping were not performed.(41) Fourthly, spirometry used to identify COPD patients in this study utilised FVC6 instead of force vital capacity, potentially excluding a proportion of patients with mild COPD. Fifthly, body plethysmography and diffusion capacity for carbon monoxide (DLCO) were not performed. Studies have shown that body plethysmography and DLCO are more sensitive than spirometry in detecting early emphysema, evidenced by increase in residual volume and reduced DLCO.(42) Besides, COPD severity graded by using compression-free FEV1 measured by body plethysmograph is more accurate than FEV1 measured by spirometry.(43) Six, ACO in this study was defined based on a history of BA and very reversible airflow obstruction on spirometry testing. Blood eosinophil count was not routinely performed in the rural areas in Malaysia. The term ACO remains controversial without an agreed-upon definition.(44) Lastly, the exacerbation frequency was subjected to the recall error of the patients. We tried to minimize this error by confirmation from the patients’ medical records and with the patients’ family members.