Metformin use and the risk of bacterial pneumonia in patients with type 2 diabetes

Polymorphonuclear neutrophils adhere to vascular endothelium and transmigrate through blood vessels in response to chemotactic gradients to phagocytose and kill invading pathogens¹. Reports suggest that persons with type 2 diabetes (T2D) have neutrophil dysfunction and disturbed cytokine production due to cumulated hyperglycemia and excess oxidative stress², with a higher risk of infection than those without diabetes³. Studies have demonstrated that persons with T2D have a higher risk of the following than those without T2D: 3.0- to 4.3-fold for urinary tract infections, 1.8- to 2.0-fold for cellulitis, 1.2- to 2.6-fold for pneumonia, and 2.0- to 3.3-fold for sepsis⁴. People with T2D have a higher risk of developing comorbidities and vascular complications, and the presence of infection can increase morbidity and mortality^3,4. The diabetes guidelines for infection are scarce⁵.

Persons with T2D may have pulmonary microangiopathy and impaired lung function⁶, with an increased risk of respiratory failure or death due to pneumonia³. According to Taiwan’s Diabetes Atlas report (2019), the odds ratios for pneumonia hospitalization in patients with T2D versus non-diabetes ranged from 2.86 to 2.93 between 2005 and 2014, with a significantly increasing trend⁷. Pneumonia constituted 3.52–7.27% of the leading cause of death in patients with T2D from 2005 to 2014, with a significantly rising trend⁸.

Metformin has long been considered the first-line medication for T2D and a key regulator of metabolism. Activation of adenosine 5'-monophosphate-activated protein kinase (AMPK) by metformin can activate neutrophils and regulate the secretion of cytokines with anti-inflammatory and antibacterial effects⁹. In diabetic mice, pretreatment with metformin could modify glucose flux across the airway epithelium and limit hyperglycemia-induced bacterial growth¹⁰. Therefore, this study aimed to compare the risks of pneumonia and respiratory death between metformin users and nonusers.

Our study showed that metformin use in persons with T2D was associated with significantly lower risks of bacterial pneumonia, IMV, and respiratory death than metformin nonuse. Moreover, a longer duration of metformin use tended to offer better protection against bacterial pneumonia and respiratory outcomes. Specifically, all the outcomes mentioned above had statistically significant protective effects when metformin use exceeded 364 days.

A study from the Emerging Risk Factors Collaboration revealed that cardiovascular disease and cancer were the main causes of death in persons with diabetes. However, infectious diseases [HR 2.39 (1.95–2.93)] and pneumonia [HR, 1.67 (1.45–1.92)] also accounted for a higher risk of death compared to non-diabetes persons¹¹. Specifically, pneumonia is the most important infectious disease in patients with T2D³. Preclinical studies demonstrated that sputum and bronchial aspirates were enriched with bacteria in animals with hyperglycemia¹⁰. Diabetes may narrow the capillary lumen of the lung and impair pulmonary function⁶. Thus, diabetes may be a unique risk factor associated with an increased incidence of pneumonia^1,4 and higher risks of hospitalization and mortality due to pneumonia than the non-diabetic state^1,3. The American Diabetes Association guidelines recommend that patients with diabetes receive anti-influenza and pneumococcal vaccines to decrease the risks of influenza and pneumonia⁵. However, coverage rates for vaccines are inadequate, especially pneumococcal vaccination¹². More efforts directed to prevent pneumonia are urgently needed to reduce the risk of progression to respiratory failure and death.

Studies show that metformin can prevent and mitigate tuberculosis infection¹³. A study on mice with diabetes showed that metformin could limit hyperglycemia-induced S. aureus growth in the airway¹⁰. One nested case-control study found no differences in the incidence of community-acquired pneumonia between oral antidiabetic drugs in monotherapy and metformin in patients with T2D¹⁴. However, a Danish population-based cohort study showed a lower risk of pneumonia hospitalization in patients with T2D and initiation with metformin therapy versus sulfonylurea or insulin¹⁵. Subgroup analyses conducted in the two studies mentioned above assessed the impact of metformin use in pneumonia risk. Yang et al. conducted a retrospective cohort study, aiming to compare the risk of pneumonia between metformin use and nonuse in individuals with T2D from the Hong Kong Diabetes Registry. This study disclosed that long-term use of metformin was associated with reduced risks of pneumonia and pneumonia-related death¹⁶. Our study is consistent with Yang’s study, showing that metformin use was associated with an 11% lower risk of bacterial pneumonia, and a longer duration and a higher dose of metformin use tended to lower the risk of bacterial pneumonia further. The comparative analysis of metformin versus sulfonylurea, and the time-varying analysis of metformin exposure also demonstrated that metformin use was associated with a lower risk of bacterial pneumonia. Additionally, the three age groups analysis revealed that younger metformin users (20–40 years) were associated with a higher risk of pneumonia (aHR1.81, 95% CI 1.28–2.57), probably because the number of younger metformin users was relatively small, and they may have suboptimal compliance of clinic visits and taking medications. However, metformin use seems to have a protective effect against pneumonia in older patients. For the older people (61–80 years), metformin users were less likely to suffer from bacterial pneumonia (aHR 0.84, 95% 0.78–0.9).

Several studies have suggested that preadmission metformin use may reduce mortality risk in patients with T2D and sepsis¹⁵. Experimental studies have demonstrated that metformin can alleviate acute lung injury and various sepsis-induced organ injuries¹⁷. Our study showed that metformin use was associated with a 23% lower risk for invasive mechanical ventilation than metformin nonuse in persons with T2D. Although metformin use in patients with hypoxia or critical condition may cause lactic acidosis¹⁷, these studies suggest that metformin use may have beneficial effects in such patients.

Metformin use was associated with significantly lower mortality risk in women with obesity or T2D who underwent hospitalization for COVID-19 infection¹⁸. Our study also showed that metformin use was associated with a 36% lower risk of respiratory causes of death than metformin nonuse in patients with T2D. This study has shown that metformin use was associated with lower risks of bacterial pneumonia and IMV use, contributing to the lower risk of respiratory death. In brief, metformin may reduce the risk of death from respiratory causes in patients with T2D.

The possible explanations for the role of metformin in preventing the development and progression of bacterial pneumonia are as follows: (1) Metformin inhibits mitochondrial respiratory-chain complex-1 and activates the liver kinase B1 (LKB1)/ AMPK pathway to facilitate neutrophil-dependent bacterial uptake and killing, and promote innate immune response¹⁸. (2) But it can suppress pro-inflammatory markers of high sensitivity C-reactive protein, interferon-α(IFN-α)¹⁹, tumor necrosis factor-α (TNF-α), and interleukin -6 (IL-6); and inhibits neutrophil activation and chemotaxis, improves neutrophil to lymphocyte ratio¹⁸, reduces B-cell intrinsic inflammation, increases antibody response, and stabilizes mast cells¹⁸. Metformin also can boost levels of the anti-inflammatory marker IL-10¹⁸. (3) The inhibition of mitochondrial complex-1 and electron transport can also suppress the energy production required for bacterial growth. Metformin inhibits bacterial gluconeogenesis, and the limited utilization of glycerol in the Kreb’s cycle reduces bacterial virulence; the anti-folate effect of metformin may inhibit the folate cycle of bacteria and limit bacterial growth²⁰. Thus, metformin may attenuate the risk of bacterial pneumonia by its metabolic, immunologic, and antibacterial effects.

This study had some limitations. First, this dataset lacked information on family history, dietary patterns, physical activity, alcohol use, smoking habits, and vaccination status; it did not include data on hemoglobin A1C, biochemical tests, renal function, immune status, and pulmonary function test, precluding an accurate understanding of patient health status, and the severity of T2D. However, we matched the demographic information on age and sex to achieve a balance between the study and control groups; we matched the type and number of oral antidiabetic drugs, insulin use, and DCSI scores to balance the severity of T2D; we have matched the comorbidity of CKD to decrease the influence of renal function on outcomes, and further increase the comparability between the study and control groups. Second, metformin is contraindicated in patients with eGFR < 30 ml/min/1.73 m². Because the NHIRD dataset lacks the information of renal function, we only excluded patients on dialysis or kidney transplant to avoid the effects of confounding by indication. Third, this study was conducted on Taiwanese people, and the results may not be generalizable to other ethnicities. Finally, cohort studies are likely to be influenced by some unknown and unmeasured confounding factors, and randomized controlled studies are warranted to confirm our results.

Standards of diabetes care have gradually become more comprehensive. A longer life expectancy in patients with T2D may contribute to the rise of several non-communicable and communicable diseases needing vigilance. Specifically, pneumonia is a critical communicable complication in patients with T2D; however, there are few recommendations on preventing pneumonia. Our study demonstrated that metformin was associated with lower risks of bacterial pneumonia, IMV use, and respiratory death. Metformin can play a beneficial role in reducing the risk and progression of pneumonia.