https://orcid.org/0000-0003-4946-1257
Figures
Abstract
Glycemic variability (GV) is an emerging biomarker of glycemic control and may be a predictor for lung function impairment in persons with type 2 diabetes mellitus (T2DM). However, the associations between GV and lung function variables and lung function impairment have not been fully evaluated. The objective of this study was to assess the associations of glycemic variability (GV) with lung function impairment in persons with T2DM. A follow-up study was conducted on the data of 3,108 subjects collected from 2001 to 2020 using the diabetes care management program database in Taiwan. GV in fasting plasma glucose (FPG) was calculated using standard deviation (SD), average real variability (ARV), coefficient of variation (CV), variability independent of the mean (VIM), and slope of 1-year repeated measurements. A ratio of forced expiratory volume in 1 s (FEV1) to forced vital capacity (FVC) less than 0.70 was used to define lung function impairment. Multivariable linear and logistic regression models were applied to explore the relationships of GV with lung function variables and lung function impairment. A total of 359 (11.6%) subjects were defined as having lung function impairment. After multivariable adjustment, FPG‐SD, FPG-CV, FPG-AVR, FPG-VIM and were found to be negatively linked with FEV1, % predicted FEV1, and FVC but not FEV1/FVC. Relative to those for the first tertile, the odds ratios (ORs) of lung function impairment for the second and third tertiles were 1.37 (95% confidence interval [CI]: 1.01, 1.87) and 1.51 (1.10, 2.08) for FPG-CV, respectively; 1.59 (1.16, 2.17) and 1.73 (1.24, 2.40) for FPG‐SD, respectively; and 1.57 (1.15, 2.13) and 1.69 (1.22, 2.33) for FPG-AVR, respectively. GV, measured by CV, SD, VIM, and VIM, is linked with lung function impairment and all lung function variables, except for FEV1/FVC ratio. GV may serve as a useful biomarker for assessing lung function impairment in persons with T2DM.
Citation: Wu Y-H, Li C-I, Liu C-S, Lin C-H, Yang S-Y, Lin C-C, et al. (2025) Visit-to-visit glycemic variability is associated with lung function variables and lung function impairment in individuals with type 2 diabetes. PLoS One 20(12): e0337885. https://doi.org/10.1371/journal.pone.0337885
Editor: Mirna Alejandra Gonzalez-Gonzalez, Tecnologico de Monterrey, MEXICO
Received: July 29, 2025; Accepted: November 14, 2025; Published: December 1, 2025
Copyright: © 2025 Wu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and the supporting information files.
Funding: This study was funded by the Ministry of Science and Technology, Taiwan (MOST 109-2314-B-039-031-MY2, MOST 110-2314-B-039-021- & MOST 111-2314-B-039-018-), National Science and Technology Council, Taiwan (NSTC 112-2314-B-039-042- & NSTC 113-2314-B-039-042-) and China Medical University, Taiwan (CMU114-S-51). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviation: BMI, Body mass index; FEV1, Forced expiratory volume in 1second; FVC, Forced vital capacity; COPD, Chronic obstructive pulmonary disease; HbA1c, Glycated hemoglobin; FPG, Fasting plasma glucose; ICD-9-CM, International Classification of Disease, 9th Revision, Clinical Modification; DCMP, Diabetes Care Management Program; ICD-10-CM, International Classification of Disease, Tenth Revision, Clinical Modification; CMUH, China Medical University Hospital; TC, Total cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, Triglycerides; HDL-C, High-density lipoprotein cholesterol; eGFR, stimated glomerular filtration rate; CV, Coefficient of variation; SD, Standard deviation; VIM, Variability independent of the mean; ARV, Average real variability
Introduction
Diabetic microangiopathy affects the lungs as it does other organs; however, the recognition of the lungs as a target of diabetes is relatively recent. Individuals with type 1 diabetes exhibit reduced total lung capacity and indicators of pulmonary restrictive dysfunction compared with their nondiabetic peers [1]. A systemic review of cross-sectional studies indicated that the lung volumes of persons with type 2 diabetes are 3%–10% lower than those of individuals without diabetes, regardless of body mass index (BMI) and smoking status. However, findings from longitudinal studies are inconsistent. Some reported that individuals with type 2 diabetes experience an accelerated decline in lung function over time [2,3], but others did not observe such decline [4,5].
Pulmonary function is most commonly assessed using the noninvasive technique of spirometry [6]. Key indicators such forced vital capacity (FVC), forced expiratory volume in 1 second (FEV1), and FEV1/FVC ratio are crucial measurements for assessing lung function. FEV1, which indicates the volume of air exhaled in the first second of a forced breath, is critical for evaluating airflow obstruction. FVC measures the total volume of air exhaled forcefully after a full inhalation and is used to assess lung volume. The FEV1/FVC ratio, whose low value indicates obstruction, helps differentiate between restrictive and obstructive (such as chronic obstructive pulmonary disease, chronic obstructive pulmonary disease [COPD] or asthma) lung diseases. The importance of these lung function measures lies in their ability to facilitate the early diagnosis of respiratory disease and condition, track disease progression to therapy, and ultimately improve management for individuals with respiratory conditions.
In addition to single glucose measurements (glycated hemoglobin [HbA1c] or fasting plasma glucose [FPG]), glycemic variability has emerged as a significant factor in blood glucose regulation and a key target in diabetes care. Growing body of evidence suggests that glucose fluctuations may serve as a novel predictor of mortality and microvascular and macrovascular morbidity [7–12]. A previous study assessed the association between glycemic variability and COPD risk in individuals with type 2 diabetes by utilizing International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM) codes from insurance claims data [13]. However, the relationship between glycemic variability and lung function variables or lung function impairment as determined by spirometry measurements has never been explored.
Materials and methods
Study design and criteria for study participants
This retrospective cohort study examined the enrollees of the Diabetes Care Managed Program (DCMP), which has been operational since November 2001. The DCMP aims to enhance quality of diabetes care and reduce complications via close monitoring, continuous support, and promotion of healthy lifestyle choices. Study subjects were enrolled during the period from November 2001 through December 2020. Eligibility criteria were as follows: diabetes diagnosis (ICD-9-CM code 250 prior to 2016 or International Classification of Disease, Tenth Revision, Clinical Modification [ICD-10-CM] code E0800 thereafter), aged 30 years and above, and completion of lung function examination. Initially, 4,128 individuals were enrolled at a teaching medical center in Taiwan. The exclusion criteria were as follows: type 1 diabetes (ICD-9-CM codes 250.x1/x3; ICD-10-CM code E10.9), ICD-10-CM code O24.419), pregnancy-associated diabetes (ICD-9-CM code 648.83; age under 30 years, and no data for glycemic variability (n = 986). After these exclusions were applied, 3,142 participants remained. Further exclusions due to data attrition on anthropometric measurements, comorbidities, and biochemical tests lowered the final sample size to 3,108 study subjects for analysis (S1 Fig). The reference date for each participant was the date of their first lung function assessment; when multiple assessments were conducted, only the first examination was considered. Measurements for other variables were collected on or near the reference date, and glycemic variability measures were derived from data gathered in the year prior to the reference date. Approval for this study was granted by the Research Ethics Committee in China Medical University Hospital (CMUH), and all procedures adhered to the relevant guidelines and regulatory standards. The Research Ethics Committee granted a waiver of inform consent (CMUH112-REC2–180), because the dataset employed in this study comprised de-identified secondary data made available for research.
Data source
Data were sourced from the electronic database of the DCMP at CMUH in Taichung, Taiwan. This program includes persons with type 2 diabetes diagnosed in accordance with the American Diabetes Association guidelines. Healthcare providers involved in the DCMP take part in training and clinical education activities to enhance their skills. The healthcare team is composed of specialists from various fields, including family medicine, endocrinology, cardiology, internal medicine, and nephrology. Ongoing education and training within the DCMP aim to standardize clinical practices related to assessing and tracking blood sugar levels and diagnosing complications of diabetes. Collaborative care is delivered by doctor-led interdisciplinary teams, which include doctors and case managers who adhere to evidence-based diabetes care management. The DCMP database contains comprehensive diabetes care information including yearly self-care assessments and instruction, yearly eye examinations, and quarterly screenings for cholesterol, blood sugar, and kidney function.
Measurements
Prior to enrolling in the DCMP, the study subjects underwent urine, blood, and body measurements and provided responses in interviews by a case manager using a standardized electronic questionnaire. This assessment was conducted annually or quarterly to obtain information about lifestyle, dietary habits, and medical history, including current or previous diseases. Sociodemographic information included age at entry, sex, and family medical history of hypertension, diabetes, obesity, and hyperlipidemia. Assessed lifestyle behaviors included alcohol consumption, smoking status, and leisure-time physical activity, which were categorized according to study subjects’ self-reports (yes versus no).
Baseline coexisting conditions included hypertension, hyperlipidemia, and obesity. Complications of diabetes were grouped into chronic and acute categories. Chronic complications encompassed peripheral vascular disease, neuropathy, stroke, nephropathy, amputation, diabetic retinopathy, and diabetic foot. Acute complications consisted of diabetic ketoacidosis, hypoglycemia, and hyperosmolar nonketotic coma. Coexisting conditions and complications were documented using a yes/no classification.
Diabetes-related variables included duration of diabetes condition. Anti-diabetes medication was classified into two main categories: insulin therapy and oral drugs. Oral drugs were classified into 7 classes: sulfonylureas, meglitinides, biguanides, dipeptidyl peptidase 4 inhibitors, α-glucosidase inhibitors, other compounds, and insulin sensitizers. Pharmacological variables encompassed drugs for hyperlipidemia (e.g., statins [HMG-CoA reductase inhibitors]), elevated blood pressure (e.g., calcium channel blockers), and nephropathy. The participants were categorized according to drug uses as recorded in their computerized medical records (yes or no).
Blood specimens were collected via antecubital venipuncture during morning hours following a 12-hour period of fast and processed within four hours after sample collection. Biochemical tests included the measurement of FPG, HbA1c, total cholesterol (TC), serum creatinine, serum low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and urinary protein concentration. A biochemical auto-analyzer (Beckman Coulter Synchron system, Lx-20, Fullerton, CA, USA) was used to analyze these biochemical markers. The estimated glomerular filtration rate (eGFR) was derived from serum creatinine measurements using the Chronic Kidney Disease Epidemiology Collaboration formula [14].
Anthropometric measurements included height, weight, and BMI. Height and weight were recorded with an automated anthropometric device (Super-view HW-666); the study subjects were instructed to take off their shoes and dress in light attire. BMI was derived using the formula: weight (kg)/ (height (m)².
Glycemic variability evaluated by FPG variability were calculated as coefficient of variation (CV), variability independent of the mean (VIM), standard deviation (SD), average real variability (ARV), and slope. To account for considering for the potential impact of varying numbers of measurements per individual, the CV is divided by , as recommended from prior study [15]. These variables were calcaulated as observing at least two measurements of FPG in one-year intervals.
Outcome measures
A spirometry test (EasyOne spirometer; NDD [New Diagnostic Design], Zurich, Switzerland) was conducted to measure pulmonary function by qualified and trained nurses following a standardized protocol [16]. Spirometry evaluates the coordinated mechanical performance of the chest wall, lungs, airways, and respiratory muscles by assessing the total exhaled air volume forcefully from fully inflated lungs to full expiration. During the test, the study subjects sat down and fitted with a nose clip. They were directed to take a deep breath, followed by a brief pause (<1 s), and exhale fully until no additional air could be released while remaining upright. Each subject underwent a minimum of three tests, and the highest values for FEV1, % predicted FEV1, FEV, and FEV1/FVC ratio were recorded for analysis.
For the measurements to be considered repeatable, the two largest values for FVC and FEV1 must differ by no more than 0.150 L, unless the maximum measurement for either parameter falls short of 1 L. FVC represents the full amount of air expelled during a forceful exhalation after the deepest possible breath. FEV1 refers to the air volume breathed out in the initial second of a forced exhale. The % predicted FEV1 was derived by expressing the patient’s FEV1 as a ratio of the expected value for a healthy individual with comparable sex, age, and body characteristics. The FEV1/FVC ratio represents the fraction of the FEV1 relative to FVC. An FEV1/FVC ratio less than 0.7 was used to indicate impaired lung function [17].
Statistical analysis
Descriptive statistics such as standard deviations, and means were summarized as continuous variables. Meanwhile, proportions and frequencies were used for categorical variables. An independent t-test was applied to continuous variables and Chi-square tests to categorical variables for bivariate analyses. The following steps were conducted to build linear models with multiple predictors: first, univariate models were built for main exploratory variables (glucose variability in HbA1c and FPG) and all covariates, and variables with p-values <0.25 were selected for further analysis [18,19]. Next, glucose variability in HbA1c and FPG were included one at a time to multiple models with the selected covariates from the previous process to assess their statistical significance (p-value <0.05). Regression coefficients were subsequently estimated using multiple linear regression models. SAS version 9.4 (SAS, Cary, NC) was used for all analyses, with two-tailed p-values and significance set at p < 0.05.
Results
Among the 3,108 persons with type 2 diabetes, the average age was 63.1 years with a SD of 11.2. Among these participants, 1,920 (61.8%) were men. Table 1 shows that individuals with lung function impairment had a significantly higher age (p < 0.001), had a higher proportion of smoking (p = 0.02), and had a reduced mean BMI (p < 0.001) than those without. In addition, the men exhibited lower prevalence of hyperlipidemia (p = 0.005) and obesity (p = 0.02) compared with the women. Furthermore, individuals with lung function impairment had a higher likelihood of using cardiovascular medication (p = 0.03) than those without.
Prior to fitting the linear regression between glucose variation and lung function variables, their relationships were assessed and found to be linear as shown in Figs 1–3. Table 2 presents the relationships between glucose variation and lung function variables FEV1, % predicted FEV1, FEV1/FVC, and FVC. In models adjusted for age and sex, all FPG variation variables were significantly linked with FEV1, except for FPG-slope. These relationships continued to be significant after controlling for lifestyle behaviors and baseline blood glucose. Further controlling for coexisting conditions and drug use did not alter these significant relationships. These results indicate a negative association, meaning that great glucose variation values are linked to low FEV1 [FPG-CV (b=−76.66, p < 0.001), FPG-SD (b=−66.88, p < 0.001), FPG-VIM (b=−52.05, p < 0.001), and FPG-ARV (b=−38.91, p < 0.001)]. Similarly, glucose variation variables (except FPG-slope) were significantly associated with %predicted FEV1 [FPG-CV (b=−2.98, p < 0.001), FPG-SD (b=−2.65, p < 0.001), FPG-VIM (b=−1.98, p < 0.001), and FPG-ARV (b=−1.24, p < 0.001)] and FVC [FPG-CV (b=−84.76, p < 0.001), FPG-SD (b=−73.43, p < 0.001), FPG-VIM (b=−57.38, p < 0.001), and FPG-ARV (b=−43.08, p < 0.001)]. None of the glucose variation variables were significantly associated with FEV1/FVC.
The ORs and 95% CI estimated from the three lung function impairment models based on glycemic variability measures are presented in Table 3. FPG-SD, FPG-CV, and FPG-ARV were significant in the model controlled for age and sex (model 1). They remained significant after accounting for diabetes-related factors, lifestyle behaviors, FPG, and HbA1c (model 2) and after further adjusting for the baseline status of complications and medication use (model 3). The ORs of lung function impairment for the second and third tertiles of FPG-CV in model 3 were 1.37 (95% CI: 1.01, 1.87, p < 0.05) and 1.51 (1.10, 2.08, p < 0.05), respectively; for FPG-SD, they were 1.59 (1.16, 2.17, p < 0.01) and 1.73 (1.24, 2.40, p < 0.01), respectively; and for FPG-ARV, they were 1.57 (1.15, 2.13, p < 0.01), and 1.69 (1.22, 2.33, p < 0.01), respectively. The linear trends of FPG-CV and FPG-SD were significant [per 1 SD: 1.15 (1.02, 1.30, p < 0.05) and 1.18 (1.04, 1.32, p < 0.01), respectively], but the effect of FPG-ARV was not significant.
Discussion
This study thoroughly examined the independent relationships between various measures of glycemic variability in FPG with distinct lung function variables and lung function impairment determined by the FEV1/FEV ratio. Except for the FEV1/FVC ratio, all the lung function variables tested were significantly linked with glycemic variability markers in FPG, with the exception of the FPG-slope. The strongest association was observed with FPG-CV, and the weakest was with FPG-ARV. FPG-CV and FPG-SD were the only two measures of variability found to be associated with lung function impairment as determined by the FEV1/FVC ratio after multivariate adjustment.
Prior studies explored the associations between diabetes status and lung function variables using cross-sectional [2,3,5,20,21] or longitudinal designs [3–5] and examined the relationship between glucose measures such as FPG and lung function variables in the general population [22–24] or persons with type 2 diabetes [25]. Most of the cross-sectional studies on the relationships between diabetes status and lung function variables found significant associations [2–5,20,21]. Similarly, most longitudinal studies reported that diabetes status is linked with changes in lung function variables [3,4], with the exception of one research conducted in older adults [5]. The studies examining the associations between blood glucose variables and lung function found that FPG [22,25] was associated with lung function variables. However, none of these studies investigated the relationships between visit-to-visit glycemic variability and lung function markers. Visit-to-visit glycemic variability refers to fluctuations in a person’s blood glucose levels across multiple healthcare visits, typically assessed over months. This metric is important for diabetes care because it provides crucial insights into the stability of a patient’s blood glucose control over time. It may offer an accurate reflection of overall glycemic control and potential risks of hypoglycemia and hyperglycemia. In addition, visit-to-visit glycemic variability is associated with the long-term risk of complications [7,11,12,26].
A potential biological mechanism may explain the link between glycemic variability and lung function impairment. Chronic hyperglycemia can lead to the glycosylation of serum and tissue proteins, resulting in the formation of advanced glycation end-products [27]. Once deposited in tissues, these glycosylated proteins trigger pro-inflammatory responses, contributing to microangiopathic complications, such as damage to the microvascular units of the alveolar–capillary network. In addition, chronic hyperglycemia may stimulate inflammatory responses [28,29], which can ultimately cause structural changes in lung tissue and impair lung function. Systemic and local chronic inflammation likely play a significant role, as hyperglycemia rapidly increases circulating cytokine levels through an oxidative mechanism. In vivo studies showed that chronic hyperglycemia can induce structural changes in lung tissue through oxidative stress, leading to restricted gas exchange [30].
This study had strengths. It was a follow-up study conducted with a hospital sample of participants enrolled in DCMP who regularly had outpatient visits and completed multiple clinical examinations. This approach enabled us to gather extensive data on their lung function variables, comorbidities, medication use, and lifestyle factors, allowing us to adjust for confounding variables effectively. This study also had some limitations. First, as this study is observational in nature, its findings may be affected by potential residual and unrecognized confounding variables. Second, the lung function variables were taken at one point in time, without accounting for trends over time. Measurement errors could have occurred due to the extensive data collected from clinical practice. Third, we had access to 1-year FPC measurements in the DCMP dataset, which restricted our ability to evaluate the impact of FPC-CV on lung function during follow-up. Fourth, the study did not include data on respiratory diseases such as bronchiolitis, asthma, COPD, and lung cancer. As a result, we were unable to assess potential associations between glycemic variability and these conditions. Last, our analysis was based on data from a single hospital, which may limit the external generalizability of the findings to broader populations of individuals with diabetes, particularly those in different geographic or healthcare settings. Nevertheless, the findings remain relevant and applicable to populations with type 2 diabetes who share similar demographic and clinical characteristics to our study cohort.
Conclusions
Except for the FEV1/FVC ratio, all the lung function variables tested and lung function impairment were significantly linked with glycemic variability measures in FPG, with the exception of the FPG-slope. Understanding the link between glycemic variability and lung function impairment underscores the importance of effectively managing blood glucose levels, particularly in patients with diabetes and respiratory conditions, to prevent or mitigate the deterioration of their lung health.
References
- 1. Boulbou MS, Gourgoulianis KI, Klisiaris VK, Tsikrikas TS, Stathakis NE, Molyvdas PA. Diabetes mellitus and lung function. Med Princ Pract. 2003;12(2):87–91. pmid:12634462
- 2. Davis TM, Knuiman M, Kendall P, Vu H, Davis WA. Reduced pulmonary function and its associations in type 2 diabetes: the Fremantle Diabetes Study. Diabetes Res Clin Pract. 2000;50(2):153–9. pmid:10960726
- 3. Yeh H-C, Punjabi NM, Wang N-Y, Pankow JS, Duncan BB, Cox CE, et al. Cross-sectional and prospective study of lung function in adults with type 2 diabetes: the Atherosclerosis Risk in Communities (ARIC) study. Diabetes Care. 2008;31(4):741–6. pmid:18056886
- 4. Lange P, Parner J, Schnohr P, Jensen G. Copenhagen City Heart Study: longitudinal analysis of ventilatory capacity in diabetic and nondiabetic adults. Eur Respir J. 2002;20(6):1406–12. pmid:12503696
- 5. Litonjua AA, Lazarus R, Sparrow D, Demolles D, Weiss ST. Lung function in type 2 diabetes: the Normative Aging Study. Respir Med. 2005;99(12):1583–90. pmid:16291079
- 6. Ranu H, Wilde M, Madden B. Pulmonary function tests. Ulster Med J. 2011;80(2):84–90. pmid:22347750
- 7. Lin C-C, Yang C-P, Li C-I, Liu C-S, Chen C-C, Lin W-Y, et al. Visit-to-visit variability of fasting plasma glucose as predictor of ischemic stroke: competing risk analysis in a national cohort of Taiwan Diabetes Study. BMC Med. 2014;12:165. pmid:25255837
- 8. Muggeo M, Verlato G, Bonora E, Bressan F, Girotto S, Corbellini M, et al. The Verona diabetes study: a population-based survey on known diabetes mellitus prevalence and 5-year all-cause mortality. Diabetologia. 1995;38(3):318–25. pmid:7758879
- 9. Muggeo M, Zoppini G, Bonora E, Brun E, Bonadonna RC, Moghetti P, et al. Fasting plasma glucose variability predicts 10-year survival of type 2 diabetic patients: the Verona Diabetes Study. Diabetes Care. 2000;23(1):45–50. pmid:10857967
- 10. Nalysnyk L, Hernandez-Medina M, Krishnarajah G. Glycaemic variability and complications in patients with diabetes mellitus: evidence from a systematic review of the literature. Diabetes Obes Metab. 2010;12(4):288–98. pmid:20380649
- 11. Yang C-P, Lin C-C, Li C-I, Liu C-S, Lin C-H, Hwang K-L, et al. Fasting plasma glucose variability and HbA1c are associated with peripheral artery disease risk in type 2 diabetes. Cardiovasc Diabetol. 2020;19(1):4. pmid:31910828
- 12. Yang Y-F, Li T-C, Li C-I, Liu C-S, Lin W-Y, Yang S-Y, et al. Visit-to-Visit Glucose Variability Predicts the Development of End-Stage Renal Disease in Type 2 Diabetes: 10-Year Follow-Up of Taiwan Diabetes Study. Medicine (Baltimore). 2015;94(44):e1804. pmid:26554779
- 13. Chiu H-T, Li T-C, Li C-I, Liu C-S, Lin W-Y, Lin C-C. Visit-to-visit glycemic variability is a strong predictor of chronic obstructive pulmonary disease in patients with type 2 diabetes mellitus: Competing risk analysis using a national cohort from the Taiwan diabetes study. PLoS One. 2017;12(5):e0177184. pmid:28489885
- 14. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–12. pmid:19414839
- 15. Kilpatrick ES, Rigby AS, Atkin SL. A1C variability and the risk of microvascular complications in type 1 diabetes: data from the Diabetes Control and Complications Trial. Diabetes Care. 2008;31(11):2198–202. pmid:18650371
- 16. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, et al. Standardisation of spirometry. Eur Respir J. 2005;26(2):319–38. pmid:16055882
- 17. Mohamed Hoesein FAA, Zanen P, Lammers J-WJ. Lower limit of normal or FEV1/FVC < 0.70 in diagnosing COPD: an evidence-based review. Respir Med. 2011;105(6):907–15. pmid:21295958
- 18. Bendel RB, Afifi AA. Comparison of Stopping Rules in Forward “Stepwise” Regression. Journal of the American Statistical Association. 1977;72(357):46.
- 19. Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989;129(1):125–37. pmid:2910056
- 20. Chance WW, Rhee C, Yilmaz C, Dane DM, Pruneda ML, Raskin P, et al. Diminished alveolar microvascular reserves in type 2 diabetes reflect systemic microangiopathy. Diabetes Care. 2008;31(8):1596–601. pmid:18492945
- 21. Walter RE, Beiser A, Givelber RJ, O’Connor GT, Gottlieb DJ. Association between glycemic state and lung function: the Framingham Heart Study. Am J Respir Crit Care Med. 2003;167(6):911–6. pmid:12623860
- 22. Kabeya Y, Kato K, Tomita M, Katsuki T, Oikawa Y, Shimada A. Association of glycemic status with impaired lung function among recipients of a health screening program: a cross-sectional study in Japanese adults. J Epidemiol. 2014;24(5):410–6. pmid:24998953
- 23. Lee W-H, Wu D-W, Chen Y-C, Liu Y-H, Liao W-S, Chen S-C, et al. Association of Pulmonary Function Decline over Time with Longitudinal Change of Glycated Hemoglobin in Participants without Diabetes Mellitus. J Pers Med. 2021;11(10):994. pmid:34683134
- 24. Maan HB, Meo SA, Al Rouq F, Meo IMU, Gacuan ME, Alkhalifah JM. Effect of Glycated Hemoglobin (HbA1c) and Duration of Disease on Lung Functions in Type 2 Diabetic Patients. Int J Environ Res Public Health. 2021;18(13):6970. pmid:34209922
- 25. Zhang R-H, Zhou J-B, Cai Y-H, Shu L-P, Simó R, Lecube A. Non-linear association between diabetes mellitus and pulmonary function: a population-based study. Respir Res. 2020;21(1):292. pmid:33148273
- 26. Li C-I, Cheng H-M, Liu C-S, Lin C-H, Lin W-Y, Wang M-C, et al. Association between glucose variation and lower extremity amputation incidence in individuals with type 2 diabetes: a nationwide retrospective cohort study. Diabetologia. 2020;63(1):194–205. pmid:31686118
- 27. Orasanu G, Plutzky J. The pathologic continuum of diabetic vascular disease. J Am Coll Cardiol. 2009;53(5 Suppl):S35-42. pmid:19179216
- 28. Esposito K, Nappo F, Marfella R, Giugliano G, Giugliano F, Ciotola M, et al. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation. 2002;106(16):2067–72. pmid:12379575
- 29. Ghowsi M, Qalekhani F, Farzaei MH, Mahmudii F, Yousofvand N, Joshi T. Inflammation, oxidative stress, insulin resistance, and hypertension as mediators for adverse effects of obesity on the brain: A review. Biomedicine (Taipei). 2021;11(4):13–22. pmid:35223415
- 30. Forgiarini LA Jr, Kretzmann NA, Porawski M, Dias AS, Marroni NAP. Experimental diabetes mellitus: oxidative stress and changes in lung structure. J Bras Pneumol. 2009;35(8):788–91. pmid:19750332