Combination therapies with thiazolidinediones are associated with a lower risk of acute exacerbations in new-onset COPD patients with advanced diabetic mellitus: a cohort-based case–control study

Multiple comorbidities of type 2 diabetes (T2DM) are common, and only 14% of patients may have no other comorbidities [1]. When considering the impact of different classes of drugs on comorbidities, the complexity of choosing therapeutic drugs for glycaemic control increases. Chronic obstructive pulmonary disease (COPD) is a common comorbidity in patients with T2DM, and approximately 10% of T2DM patients have COPD [2, 3]. Recent studies have shown that pre-existing diabetes or incident diabetes may worsen the risk of death in patients with COPD or acute exacerbations of COPD (AECOPD) [4, 5]. Therefore, optimizing DM care in the COPD population is imperative.

The recently updated guideline from the American Diabetes Association (ADA) recommends metformin, if not contraindicated and if tolerated, as the preferred initial oral antihyperglycaemic drug (OAD) for the treatment of T2DM [6]. As the progressive natural course of T2DM, when metformin monotherapy is no longer effective, the majority of advanced T2DM patients require a combination of different 2nd OADs or insulin therapy to achieve and maintain optimal glycaemic control. The ADA does not prioritize specific 2nd drugs based on their efficacy, side effects and impact on comorbidities except for cardiovascular and renal effects [6].

Whether strict blood glucose control in COPD-T2DM patients can improve the outcomes of COPD is unclear, but poor sugar control worsens the severity and clinical course of COPD based on previous studies [7, 8]. Although previous studies have demonstrated the impact of OADs on the clinical outcomes of COPD, such as metformin (MET) and thiazolidinediones (TZDs) [9‐11], relatively few studies have focused on the effect of glucose-lowering agents on COPD outcomes in T2DM patients, particularly in patients with poor glycaemic control requiring add-on therapy to MET. At present, a knowledge gap remains in choosing the best drugs that are conducive to glycaemic control and can improve the clinical efficacy of new-onset COPD in patients with advanced T2DM. Therefore, the aim of this study was to examine the impact of add-on OADs on AECOPD risk in new-onset COPD patients with advanced T2DM who required combinational therapy. We conducted a cohort-based case–control study using data from the National Health Insurance Research Database (NHIRD) of Taiwan.

In the present study, we showed that COPD-T2DM patients with OAD use were treated only for glycaemic sugar control, while a double combination of SU and TZDs and a triple combination of MET, SU, and TZDs were correlated with a decreased risk of AECOPD. The results were fairly consistent in patients with moderate or high complexity COPD.

In the COPD-T2DM cohorts, our analysis results revealed that add-on TZDs could decrease the risk of AECOPD in patients receiving double and triple OAD combinations. The results are consistent with previous findings that TZDs are associated with a reduced risk of AECOPD after adjusting for the severity of DM itself, which may have a significant effect on AECOPD [11]. Here, we further considered the effect of COPD severity on AECOPD using COPD complexity classification. To rule out the effect of COPD medications on AECOPD, we found that TZD has a similar effect on reducing the frequency of AECOPD after adjusting for confounding factors.

COPD-T2DM is considered a syndrome that can share risk factors (such as smoking) [17], genes (such as β₂-adrenergic receptor gene, ADRB2) [18], proteins (such as Nod-like receptor containing a pyrin domain 3, NLRP3) [19, 20] and pathways (such as systemic inflammation and oxidative stress) [21‐23]. Although the underlying mechanism of these shared components is complex and has not been fully elucidated, the important common approach for concurrently treating COPD and T2DM to target systemic inflammation would be a reasonable therapeutic strategy [24].

In addition to the function of lowering glycaemic sugar, some OADs may also have anti-inflammatory activity due to their pleiotropic effects [25, 26]. TZD is an OAD with anti-inflammatory activity. Since the late 1990s, TZD has been studied and has been used in combination with MET to treat T2DM [27]. The anti-inflammatory effects of TZD occur through cellular mechanisms that activate the nuclear transcription factor peroxisome proliferator-activated receptor gamma (PPAR-γ) and, at least in part, glucocorticoid nuclear translocation [28, 29]. AECOPD patients with frequent exacerbations have more inflamed existing airways and systemic inflammation and poor inflammation resolution [30‐34]. Therefore, TZDs may exert anti-inflammatory effects that prevent the pro-inflammatory status of AECOPD. Additionally, the major comorbidities of COPD-T2DM, such as CVD, may induce or worsen AECOPD [35]. TZDs also exert important functions in regulating vascular inflammation through PPAR-γ activation and inhibit vascular smooth muscle proliferation, thereby having an effect against atherosclerosis [36‐38]. The protective role of TZDs on cardiovascular outcomes may contribute to reducing CVD-related AECOPD, particularly pioglitazone [39, 40].

Previous studies have shown that MET use can reduce the utilization of health care in COPD-T2DM patients and reduce its adverse prognostic effects [9, 10]; however, no significant beneficial effect of combination therapy with MET use was found in our research. The major causes of the different results can be attributed to the enrolled and analysed patients. The favourable effect of MET on reducing COPD-specific health care utilization was only presented in COPD patients with lower complexity but not moderate to high complexities, and the study did not show the effect of mono- or combinational therapy with MET. In our study, we clearly defined patients treated with combination therapy including, MET and the data demonstrated no effect on reducing the risk of AECOPD in all COPD complexities. COPD-T2DM patients may have better survival outcomes with MET treatment; however, the consequence may be confounded by COPD severity and the medication regimen for COPD.

DPP-4 is another target of OAD that can drive the T helper type 1 (Th1) immune response and is also considered to be involved in COPD pathogenesis [41]. Additionally, the active protease DDP-4/CD26 may act on CXCL12, which is associated with exacerbating tissue damage in COPD [42]. Therefore, COPD may be better controlled using an inhibitor of DDP-4 (DDP-4i); however, our results did not show significant differences in the risk of AECOPD when therapy was combined with DDP-4i.

Our study suggests that TZDs are a better choice for combinational therapy when glycaemic control deteriorates from initial control in COPD-T2DM patients. This recommendation was based on a more strictly defined patient population, controlled for important clinical confounders and considering the impact of OADs on AECOPD. Our study does have some limitations. The use of ICS in COPD plays an important role in glycaemic level control. No statistically significant change in the HbA1c level was found in a small prospective randomized, double-blind placebo-control, 42-day short-term study [43]; however, a large retrospective study showed that long-term use of ICS for the treatment of COPD or asthma was correlated with the progression of diabetes [44]. In addition, the effect of ICS on the deterioration of glycaemic control may be related to the dose [45]. The efficacy and safety between the two more frequently used ICSs (fluticasone and budesonide) have been reported; but, only fluticasone showed a dose-related increase in the risk of pneumonia in COPD and may induce stress hyperglycaemia [46]. In general, ICS is usually prescribed to patients with more severe and frequent exacerbations of COPD, and systemic inflammation might also be related to elevated glycaemic levels [21]. The role of ICS and specifically the different types of ICS used on AECOPD in the different classes of OADs is important to evaluate. Unfortunately, in the current study, we only analysed ICS use as a confounding factor when assessing the association between TZD and AECOPD. Thus, the assessment of ICS remains to be further investigated. Second, data on drug exposure were obtained from prescription records, which may not reflect actual usage. According to the 2017 GOLD guidelines, a fixed combination of LAMA/LABA is recommended as the first-line treatment for COPD, and ICS is recommended as an additional treatment under specific conditions [47]. Moreover, ICS is considered to be the first-line treatment for COPD patients with severe airflow limitation [48]. In this study, there was a low use of bronchodilators and a high use of ICS alone in COPD from 2003 to 2014, which may be because the dual bronchodilator was first approved in Taiwan in 2014. ICS alone may not truly be used alone because different combinational inhaler therapies were difficult to recognize from our database. Moreover, the trend was consistent with the higher use of ICS/LABA treatment for COPD in real-world data before the change in the GOLD guidelines in 2017 [49]. Third, the NHIRD database lacks other important clinical information, such as smoking, certain vaccinations, and medication compliance, which may lead to AECOPD occurrence. Fourth, the administrative claims database from which the NHIRD sample was derived did not consider certain clinical characteristics, such as the severity of COPD. Thus, we chose new-onset COPD patients to decrease the bias due to COPD severity. In addition, we applied the cross-sectional analysis developed by Mapel et al. [15] to adjust the potential effect of the level of COPD severity on AECOPD risk. We recognized that some uncontrolled influences remain that may affect the results of our study. Finally, this study included a cohort of Taiwanese patients and therefore may not be generalizable to other populations due to variations in genetics and treatment guidelines for both diseases in other areas. Future prospective studies on the effects of TZDs are warranted to confirm our findings.

These results showed that combination therapy with TZDs is associated with a reduced risk of AECOPD regardless of double or triple combinational regimens in COPD-T2DM patients, particularly in moderate to severe complexity COPD populations. The number of T2DM patients with co-existing COPD may increase in the future. TZDs play different protective roles for both diseases and are suggested to be used in these patients, but prospective randomized controlled trials are needed to verify our results.