Association between statin use and tuberculosis risk in patients with bronchiectasis: a retrospective population-based cohort study in Taiwan

Data sources

Data were obtained from Taiwan’s National Health Insurance Research Database (NHIRD). The NHIRD is an administrative claims database from Taiwan’s National Health Insurance programme, which covers all inpatient and outpatient health services of 99% of Taiwan’s population. It includes diagnostic codes for all diseases based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) for records before 2015 and the ICD-10-CM for records after 2015. In addition, the NHIRD includes data on patient demographics, healthcare utilisation and prescribed medications. For research purposes, Taiwan’s Health and Welfare Data Science Centre integrated the population database, which was linked to different health-related datasets, and managed the application to avoid violations of personal information.

Study subjects

Patients over 20 years of age who were diagnosed with new-onset bronchiectasis between 2008 and 2017 and were followed up for at least 3 months were included in this study. Patients were divided into two groups, cases and controls, based on whether or not they used statins. Bronchiectasis was defined according to the ICD-9-CM before 2015 or ICD-10-CM codes after 2015. Statins were defined based on the Anatomical Therapeutic Chemical classification codes used in the medical records of patients who used statins (online supplemental table S1).

In Taiwan, statins are commonly used for the management of high cholesterol levels and the prevention of cardiovascular diseases. Commonly used statin drugs include simvastatin, lovastatin, pravastatin, fluvastatin, atorvastatin, cerivastatin, rosuvastatin and pitavastatin. To assess the rational drug use status, we also considered the defined daily dose (DDD), which is a standardised measurement used to compare the average daily dose of different medications within a specific drug class. The DDD is defined by the WHO as the assumed average maintenance dose per day for a drug used for its main indication in adults. In this study, the DDD of statin drugs was divided into two categories: DDD≤180 and DDD>180 per year.

Furthermore, to understand the potential impact of statin use, we presented additional analyses, including statin exposure time and the group of current or past statin users. Statin exposure time was identified into fixed periods before and after the diagnosis date of bronchiectasis within 1 year. The cut points for this categorisation were determined by the first and third quartile of the exposure time, allowing for a standardised assessment of the duration of statin use. Additionally, patients taking statins were classified into three groups: non-users, current users (before and after bronchiectasis diagnosis) and past users (only before bronchiectasis diagnosis). The statin utilisation patterns and the potential association with the risk of TB could be explored in these classifications. These additional analyses contributed to estimate the dynamics of statin use for TB risk within the bronchiectasis patient cohort.

Outcome and measurements

The primary outcome of this study was the incidence of new-onset TB. New-onset TB was defined based on the relevant diagnostic codes from inpatient and outpatient claims (ICD-9-CM, 010-012 and 013-018; ICD-10-CM, A150, A155 and A154-A159), used at least two times within 6 months for outpatients or one admission for inpatients, or based on prescriptions of anti-TB drugs. Anti-TB drugs, which included isoniazid, rifampicin, pyrazinamide and ethambutol, were defined by the Anatomical Therapeutic Chemical classification codes.

For estimating the risk of TB, patients with a diagnosis code of TB or using anti-TB drugs more than 14 days before the diagnosis date of bronchiectasis were excluded. All study subjects were followed up until new-onset TB, death, or the end date of the study, 31 December 2018. The flowchart of patient selection is illustrated in figure 1.

Request permission

In addition, the confounding factors of age, sex, income, Charlson Comorbidity Index (CCI) and comorbidities, including heart diseases, cerebrovascular disease, end-stage renal disease (ESRD) and diabetes, were measured in this study. Based on age, patients were divided into two age subgroups: <65 and ≥65 years. Based on the CCI, they were divided into three disease severity groups: CCI=0, CCI=1–2 and CCI≥3. Comorbidities were defined according to the ICD-9-CM and ICD-10-CM diagnostic codes in patients diagnosed with bronchiectasis with at least three outpatient visits or one inpatient admission within 1 year (online supplemental table S1).

Statistical analysis

Descriptive statistics were employed to summarise categorical variables, presented as frequencies with percentages, and continuous variables, presented as means with SD. Differences between patients taking statins and those not taking statins (cases and controls) were analysed using Pearson’s χ² test or Student’s t-test, as appropriate.

The trend of TB incidence among patients taking statins compared with those not taking statins was illustrated using Kaplan-Meier curves, and the differences between the two groups were assessed using the log-rank test. Considering of potential immortal time bias due to stain exposure time, Cox regression models with time-dependent covariates were used to estimate the HR with 95% CIs for TB incidence among patients with bronchiectasis.11 In this model, subjects treated with statin were defined as the exposure group during a specified period of time, and those would be switched to the non-exposure group when they stopped receiving statin during another specified period of time. Stratified analyses by sex and age groups were conducted, adjusting for potential confounding factors, including of gender, age groups, CCI group, comorbidities (myocardial infarction, congestive heart failure, cardiovascular disease, diabetes mellitus and ESRD) and income.

Considering of potential co-prevalent disease, where TB may have been present at the time of bronchiectasis diagnosis and discovered later through subsequent investigations, we performed the sensitivity analysis. Specifically, we focused on patients diagnosed with bronchiectasis more than 1 year after the initial diagnosis, and we excluded individuals at higher risk, such as those with cancer or ESRD. These measures were implemented to understand the impact of potential biases related to co-prevalent disease.

All statistical analyses were performed using SAS V.9.4, and a p value of <0.05 was considered statistically significant. Kaplan-Meier curves were plotted using STATA (V.12).

Patient and public involvement

None.

Statins and bronchiectasis

Statins reduce the risk of cardiovascular disease due to their anti-inflammatory action and lower the levels of the non-specific inflammatory marker C reactive protein via immunomodulatory effects.13 They have been shown to reduce proinflammatory cytokine levels, affect monocyte chemotaxis, and interfere with neutrophil-endothelial interactions and migration.14 Accordingly, statins have emerged as a possible therapy for reducing acute exacerbations of bronchiectasis-chronic obstructive pulmonary disease overlap.15 In addition, a randomised controlled study that included 30 patients with bronchiectasis who received statin therapy for 6 months found that patients’ coughs and quality of life improved after treatment.16 Furthermore, statins have pleiotropic effects on the immune system and have been associated with better outcomes in several infectious diseases.17 Our study found that statins may further reduce the risk of TB in patients with bronchiectasis.

Possible mechanisms of TB risk reduction

Several possible mechanisms could explain the decrease in TB risk in patients with bronchiectasis receiving statin treatment. As shown in an in vitro study, statins can activate caspase-1 and enhance the secretion of interleukin 1β, interleukin 18 and interferon-gamma. By enhancing the Th1 response, they exert a protective role against TB infection.18 Additionally, statins can hinder the activation of γδ T cells induced by TB.19 Moreover, these drugs enhance macrophage phagosome maturation and autophagy and inhibit autophagosome maturation, resulting in enhanced protection and reduced TB burden, which has beneficial effects on preventing TB growth.20 Another study also confirmed that statins promote phagosome maturation and have additive effects on treatment with rifampicin in a BCG strain of M. bovis and M. leprae.21

Furthermore, statins could reduce the growth of TB in THP-1 macrophages compared with the control without antibiotics.22 An animal study examining the effects of statins as adjuvant treatment with anti-TB drugs, which included rifampicin, isoniazid and pyrazinamide, showed that treatment with statins reduces the bacterial load, shortens treatment times and decreases the relapse rate, which is consistent with the results of cell strain studies.23

Statins and TB

Our study showed that statins have a beneficial effect on the risk of TB in patients with bronchiectasis. However, in a previous study, statins did not have a protective effect against TB. This retrospective cohort study conducted in South Korea using a nationwide claim database enrolled patients with newly diagnosed type 2 diabetes and found that statin use did not reduce the TB risk among these patients.24 In that study, the high comorbidity prevalence among statin users compared with non-statin users could be a high-risk factor for TB development. Furthermore, the study only enrolled patients with newly diagnosed diabetes, and their results cannot be generalised to all patients with diabetes.

Another retrospective study conducted in Taiwan also enrolled patients with type 2 diabetes older than 65 years. The authors used the Cox proportional hazards regression model to determine the risk of developing TB in these patients. After adjusting for age, sex, medication and comorbidities, they reported that patients with diabetes older than 65 years receiving statin treatment had a lower risk of TB.25

In a nested case–control study that used the national health insurance claim database in Taiwan between 1999 and 2011, statin therapy could reduce the risk of active TB, and the length of treatment was associated with the effect of TB protection.26 Su et al27 also used the national health insurance claim database in Taiwan between 2000 and 2013 and revealed a lower risk of TB among statin users, with an HR of 0.53 (95% CI 0.47 to 0.61; p<0.001). Our results were similar to theirs. Statin treatment had a dose–response relationship with the risk of TB, and statin users with a cDDD>180 per year had the most potent protective effect compared with controls and statin users with a cDDD≤180 per year. Bronchiectasis, characterised by chronic respiratory conditions, is often associated with an increased risk of respiratory infections, including TB. Statin therapy has been suggested to potentially reduce the risk of TB, with the duration of treatment being linked to a protective effect against TB development. Previous studies indicated a reduced risk of TB among statin users in the general population in Taiwan.26 27 In our study, specifically focusing on patients with bronchiectasis, similar results suggested that statin use may contribute to a lowered risk of TB in this population.

One strength of this study is that it was based on a nationwide claim database, which minimised the possibility of selection bias. The NHIRD enrolled 22.96 million people in this programme. Further, to increase the accuracy of the TB diagnosis, we used strict diagnostic criteria based on which patients should both have the diagnostic codes and receive TB treatment.

Limitations

This study has several limitations. First, data on some potential risk factors, such as smoking, alcohol use and nutritional status, were not available. However, some smoking-related health consequences may partly be reflected in some comorbidities, such as coronary heart disease and chronic obstructive pulmonary disease, which were included in the CCI analysis. Second, statin prescription information was based on administrative claim data and may not reflect medication adherence. To address this limitation, the analysis of statin exposure time, current or past statin users and cDDD of statins were used to estimate the TB risk for more information about the potential impact of statin use in our study population. Third, the TB definition was validated in a previous study,27 confirming the accuracy of the diagnosis. Additionally, the determination of the lag time for TB in bronchiectasis was a challenge due to the retrospective study using secondary database. The diagnostic delay associated with undiagnosed or subclinical bronchiectasis before formal diagnosis may introduce variability, impacting the precision of the reported median time to TB. Finally, because our subjects were Taiwanese, the generalisability of our results to other populations requires further validation. It is important to note that our dataset lacks records of individuals born outside Taiwan. Therefore, the potential influence of birthplace on TB risk could not be considered in the multivariable models. Exploring the impact of birth outside Taiwan could be considered for future studies.