3.2 Treatment and medical costs in different groups
Referring to the treatment, the proportions of antibiotic and joint antibiotic use were higher in non-eosinophilic AECOPD patients, but the use of short-term steroids showed no difference between these two groups. Interestingly, lengths of stay (LOS) in eosinophilic AECOPD were shorter than those in non-eosinophilic AECOPD among elderly patients, but it showed no difference in middle-aged patients. Besides, the overall hospital costs (OHC) and daily hospital costs (DHC) was lower in elderly eosinophilic AECOPD patients. (Table 2).
3.3 Association between eosinophilic exacerbation and clinical remission
Compared to patients with non-eosinophilic exacerbation, patients with eosinophilic exacerbation had an higher probability of clinical remission at the time point of 7 (p =0.020), 10(p =0.002) and, 14 (p =0.003), 21 (p< 0.001), and 28 days (p =0.003) (Table 3). Figure 2 shows and the changes of cumulative remission rates along with hospital days, with the horizontal axis (x-axis) representing time in days and the vertical axis (y-axis) the cumulative proportion of clinical remission at the different time points. The higher rate of clinical remission was observed in eosinophilic AECOPD patients at most of the time point when comparing with non-eosinophilic patients (Figure 2a).
In elderly patients, the Cox proportional hazards model showed that eosinophilic exacerbation was significantly associated with in-hospital clinical stability and eosinophilic exacerbation tended to have a higher rate of clinical remission compared to non-eosinophilic exacerbation at 7 (p =0.003), 10 (p =0.040), 14 (p =0.003), 21 (p= 0.020), and 28 (p =0.010) day time points. In contrast, no statistically significant differences were observed at all time points (all p >0.05) in the middle-aged patients (Table 3 and Figure 2b, 2c).
4. Discussion
Earlier and accurate individual risk assessment of patients is important for successful disease management[27]. Despite recent technical progress in diagnosis and treatment, the management of AECOPD remains difficult; thus, it is necessary to identify clinically meaningful predictors for successful management[28]. This retrospective study investigated the relationship between eosinophils and clinical remission of AECOPD and demonstrated several new findings. First, we observed a lower medical cost in eosinophilic AECOPD patients. Second, subgroup analysis stratified by age groups revealed a higher rate of clinical remission observed among elderly but not in middle-aged eosinophilic patients. In the baseline analysis, we found that neutrophils, hs-CRP level, and OHC in eosinophilic cases were lower and hospital time was shorter than those in non-eosinophilic cases. These findings are consistent with those of previous studies despite different clinical settings, enrollment criteria, and definitions of disease severity and exacerbation[17, 26, 29]. Our findings of statistically significant differences in LOS and OHC between groups was consistent with those of a previous study [30]. Furthermore, the Cox proportional hazards model used to assess and explore the risk of eosinophilic exacerbation to in-hospital remission of AECOPD patients, showed that eosinophilic group tended to show higher rates of clinical remission at all time points (7, 10, 14, 21, and 28 day time points). In the subgroup analysis assessing the association between eosinophilic exacerbation and remission, eosinophilic exacerbation group in elderly patients showed a higher clinical remission rate at all time points. In contrast to these findings in elderly patients, no significant differences in middle-aged patients were observed.
Eosinophilic exacerbation is associated with eosinophilic airway inflammation which was associated with T-helper cell type 2 (Th2) inflammation. Increasing evidence in recent years has shown that eosinophilic airway inflammation might play an important role in COPD[31]. Subsequent infiltration of eosinophils in the airways of COPD patients regulates the immunoglobulin E (IgE)-dependent mechanism, which is considered as an important part of Th2 cell cytokine production process[32], with an irregular Th1/Th2 ratio and interleukin-17/IgE ratio that may be caused by imbalanced Th2 cell production[33]. Previous clinical trials have explored the relationship between blood eosinophil counts over time and in-hospital management outcomes[16]. Eosinophilia has been observed both in COPD and AECOPD patients and represents a novel pathological mechanism of a special clinical phenotype in COPD. Moreover, eosinophilic exacerbation also affects COPD outcomes. Previous studies reported an association between COPD patients with persistently increased blood eosinophil counts and better outcomes; furthermore, evidence from randomized clinical trials has shown that patient blood eosinophil counts are crucial for predicting a good response to corticosteroid administration[34, 35]. Corticosteroid treatment at the time of an exacerbation may yield better interventional outcomes with shorter LOS and better treatment responses[36]. In the current study, with the similar rate of steroids use, elderly patients with eosinophilic AECOPD gained higher clinical remission rate than non-eosinophilic elderly patients, which implied that steroids acted better in elderly eosinophilic patients, highlighting the value of steroids in AECOPD patients.
Age is always listed as a risk factor for COPD; elderly patients are often accompanied by other typical comorbidities that negatively impact prognosis and health status, and the number of comorbidities is associated with the risk of COPD exacerbation[11, 37]. However, although COPD is common in older people, whether there is difference between elderly patients and middle-aged patients in terms of eosinophilic phenotype-related COPD outcomes remains unclear. Previous study reported the value of blood eosinophilia for predicting clinical remission of AECOPD by setting 10 days as the observation time point; due to the design and information collected by the study, observing the association at more set time points or adjusting for age were not possible. In this study, we were able to further demonstrate this phenomenon by setting more observation time points (at 7, 10, 14, 21, and 28 days in hospital), and patients were divided into different age groups, which would make the difference between middle-aged and elderly patients more visible. We found a significant difference in clinical remission between elderly patients and middle-aged adults over all the observation time points.
The current study has some limitations. First, although 703 patients were enrolled, this study was a single-center retrospective study. Further multi-center prospective studies evaluating the value of blood esinophils are needed. Second, patient follow-up after discharge was not performed in this study, as the data were collected from electronic medical records, and readmissions and other follow-up-related information of these patients could not be identified. To better understand the relationship between eosinophils and elderly ages, additional eosinophilic phenotype-related clinical and basic studies focusing on AECOPD at different ages might be performed
5. Conclusion:
In this study, the eosinophilic phenotype was associated with better clinical remission at 7, 10, 14, 21, and 28 days in the hospital among elderly but not middle-aged patients with AECOPD. Early identification of this special phenotype in elderly patients may be valuable in better personalized management for AECOPD.