Association of coronary artery calcium score with cardiovascular events: a retrospective study

Introduction

Coronary artery calcification (CAC) is a common pathological manifestation of atherosclerosis, directly reflecting the degree of calcification in coronary plaques and marking the overall burden of arteriosclerosis (1-3). Detection of CAC can be achieved through various imaging techniques, among which coronary computed tomography angiography (CCTA) is widely used in clinical practice with its high resolution and non-invasive nature (4,5). The coronary artery calcium score (CACS), as a quantitative indicator of CAC, has been shown to be closely related to the occurrence of coronary heart disease (6). In recent years, the role of CACS in assessing cardiovascular event risk has attracted widespread attention.

Previous studies have indicated a close relationship between high CACS and a significant increase in the incidence and mortality of coronary heart disease (7). In asymptomatic individuals, CACS can be used for long-term risk prediction of cardiovascular events, as it reflects the overall burden of coronary atherosclerosis (8). Additionally, CACS has high sensitivity and specificity in assessing the severity of coronary artery stenosis and predicting the occurrence of cardiovascular events (9). However, the variability in cardiovascular risk among patients with different CACS strata and the applicability of CACS in different patient populations have not been fully clarified, necessitating further research (10).

The relationship between CACS and other traditional cardiovascular risk factors (such as hypertension, diabetes, smoking, etc.) has also garnered interest among researchers. Studies have shown that the combined use of CACS with other risk factors can significantly improve the accuracy of cardiovascular event prediction (11). For example, certain high-risk patient populations (such as elderly individuals and diabetics) may experience a significant increase in cardiovascular event risk due to an elevation in CACS (12). Furthermore, the association between CACS and cardiovascular event risk may vary among different subgroups, providing a new perspective for studying the value of CACS in cardiovascular event prediction (11).

This study aims to further explore the association between CACS and the risk of cardiovascular events and assess the combined predictive effect of CACS with other risk factors by retrospectively analyzing data from 100 patients who underwent CCTA in Dangyang People’s Hospital between 2018 and 2023. By analyzing the stratified effect of CACS and its performance in different subgroups, this study aims to provide clinical practitioners with a more targeted risk assessment tool to aid in the early detection of high-risk patients and the implementation of corresponding prevention strategies. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-549/rc).

Methods

Study population

This retrospective study included 100 patients who underwent CCTA at Dangyang People’s Hospital from 2018 to 2023. Patients were selected based on the inclusion criteria of having complete medical records and a minimum follow-up period of 1 year. The primary clinical indications for CCTA were evaluation of chest pain (73% with stable angina), assessment of suspected coronary artery disease (CAD) (12% with unstable angina), and acute coronary syndrome workup (15% with non-ST or ST-elevation myocardial infarction). Patients with a history of CAD requiring revascularization or those with incomplete clinical data were excluded. Sample size adequacy was assessed based on the events-per-variable (EPV) rule of thumb for Cox regression. With 100 patients and 16 cardiovascular events observed during follow-up, the EPV ratio was 2.67 (16 events/6 covariates), which aligns with recommendations for exploratory analyses (13). While larger cohorts are ideal for definitive conclusions, this sample size provides sufficient power to detect moderate-to-large effect sizes in a hypothesis-generating context. For participants, characteristics including smoking status (never/former/current), smoking duration (years), diabetes severity [hemoglobin A1c (HbA1c) levels and disease duration], lipid profiles [low-density lipoprotein (LDL), triglycerides], and body mass index (BMI) have been collected. However, detailed smoking frequency (cigarettes/day) and longitudinal HbA1c measurements were not available in all patients. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Ethics Committee of Dangyang People’s Hospital (No. LL-2024-005-01), and individual consent for this retrospective analysis was waived.

CCTA examination and CACS calculation

CCTA was performed using a 320-detector row CT scanner (Aquilion ONE, Canon Medical Systems) with prospective electrocardiogram (ECG) gating. Scanning parameters included: tube voltage 120 kV, tube current 300–500 mA (adjusted for BMI), and slice thickness 0.5 mm. Coronary artery stenosis severity was graded as follows: no stenosis (0%), mild stenosis (1–49%), moderate stenosis (50–69%), severe stenosis (≥70%). Plaque composition was classified as calcified, non-calcified, or mixed based on Hounsfield unit thresholds. CACS was quantified using the Agatston method, which sums the calcium scores of each coronary artery. Patients were stratified into four CACS groups: CACS =0, 1–100, 101–399, and ≥400. The CCTA images were independently analyzed by two experienced radiologists who were blinded to the clinical outcomes. CT image analysis of partial typical patients was shown in Figures S1,S2. The correlation analysis between CCTA and CACS scores was provided in Table S1.

Follow-up and outcome measures

Patients were followed for one year after CCTA, with regular clinic visits or phone interviews conducted at 3, 6, and 12 months. The primary outcome was all-cause mortality and the occurrence of MACE, including non-fatal myocardial infarction, unstable angina requiring hospitalization, and cardiovascular-related death. Secondary outcomes included survival stratified by clinical characteristics such as sex, smoking status, and the presence of hypertension.

Kaplan-Meier survival analysis

Survival probabilities were estimated using the Kaplan-Meier method, stratified by CACS groups. Log-rank tests were used to compare survival distributions across groups. Subgroup analyses were conducted to evaluate the effects of demographic and clinical factors, including sex, smoking history, and hypertension, on survival outcomes.

Cox proportional hazards regression

Cox proportional hazards regression was performed to identify independent predictors of mortality and MACE. The covariates included in the multivariate model were age, sex, smoking status, hypertension, diabetes, and dyslipidemia. Hazard ratios (HRs) with 95% confidence intervals (CIs) were reported for each variable. Model fit was assessed using the concordance index, and statistical significance was set at P<0.05.

Statistical analysis

All statistical analyses were conducted using R software (version 4.3.2). Kaplan-Meier curves were generated using the “survival” package, and Cox proportional hazards regression was performed using the “coxph” function. Statistical tests were two-sided, and P values <0.05 were considered statistically significant.

Results

Clinical characteristics of patients

A total of 100 patients who underwent CCTA and had at least 1 year of follow-up were included in this study, comprising 63 males and 37 females, with a mean age of 65.4±10.7 years. The baseline characteristics of the study population are summarized in Table 1. The prevalence of smoking history, hypertension, hyperlipidemia, and diabetes among the patients was 31%, 55%, 58%, and 33%, respectively. Additionally, 8% of the patients had chronic kidney dysfunction. Clinically, 73% of the patients were diagnosed with stable angina, 12% with unstable angina, 10% with acute non-ST elevation myocardial infarction, and 5% with acute ST elevation myocardial infarction (Table 1). As shown in Table S2, subgroup analyses revealed significant associations between CACS and current smoking status (56% in CACS ≥400 vs. 9% in CACS =0, P=0.023), diabetes duration (8.9±4.5 years in CACS ≥400 vs. 2.1±1.8 years in CACS =0, P=0.008), and LDL levels (3.9±1.1 mmol/L in CACS ≥400 vs. 2.8±0.6 mmol/L in CACS =0, P=0.015). However, HbA1c and BMI showed no significant trends across CACS groups (P>0.05).

Survival analysis stratified by CACS

The Kaplan-Meier survival curves for all patients, stratified by CACS groups (CACS =0, CACS =1–100, CACS =101–399, and CACS ≥400), are shown in Figure 1. A significant difference in survival probabilities between the groups was observed (log-rank P=0.011). Patients in the CACS ≥400 group exhibited the lowest survival probability, with survival decreasing more rapidly compared to the other groups. In contrast, patients with CACS =0 maintained the highest survival probability throughout the 12-month follow-up period. This finding suggests that higher levels of CAC are strongly associated with a reduced survival probability, highlighting the prognostic value of CACS in this patient cohort. Furthermore, CCTA analysis revealed a strong correlation between CACS and coronary stenosis severity (Table S1). All patients with CACS =0 had no detectable stenosis, while 60% of CACS ≥400 patients had severe stenosis (≥70%). Non-calcified plaques were more prevalent in higher CACS groups (73% in CACS ≥400 vs. 0% in CACS =0), suggesting that CACS reflects both calcified and total plaque burden.

Figure 1 Kaplan-Meier survival curves depicting the relationship between CACS and survival probability over 12 months of follow-up. The four groups are stratified by CACS: 0, 1–100, 101–399, and ≥400. The survival rate decreases with higher CACS, with the highest CACS group (CACS ≥400) showing the poorest survival. The P value from the log-rank test indicates a significant difference in survival between the groups (P=0.011). CACS, coronary artery calcium score.

Association between CACS and cardiovascular event risk

Cox proportional hazards regression analysis was conducted to evaluate the association between CACS and cardiovascular event risk. The analysis included factors such as age, sex, smoking history, hypertension, diabetes, and dyslipidemia, revealing that CACS is an independent predictor of cardiovascular events. Specifically, higher CACS was associated with an increased risk of all-cause mortality and MACE (P<0.05). Kaplan-Meier survival curve analysis indicated significant differences among survival curves of different CACS strata (P<0.05), with the CACS ≥400 group showing a significantly higher incidence of cardiovascular events compared to other groups, while the CACS =0 group exhibited the lowest risk (Table 2).

Cox proportional hazards model

To quantify the effect of CACS on survival while adjusting for covariates such as age, gender, smoking status, hypertension, diabetes, and dyslipidemia, we performed a Cox proportional hazards model analysis. The model yielded a concordance index of 0.851, demonstrating a high degree of predictive accuracy (Table 3). The Likelihood ratio test (43.12, P=2e−06), Wald test (32.94, P=1e−04), and Log-rank test (46.13, P=6e−07) were all statistically significant, confirming the robustness of the model. In this multivariate analysis, higher CACS levels remained significantly associated with increased mortality risk, particularly in male and hypertensive patients. Conversely, in females and non-hypertensive patients, the impact of CACS on survival was less pronounced, suggesting that the prognostic value of CACS may vary across different demographic and clinical characteristics.

Subgroup survival analysis based on demographic and clinical characteristics

To further investigate the impact of CACS on survival within specific subgroups, we conducted Kaplan-Meier survival analyses stratified by gender, smoking status, and hypertension (Figure 2). In the male subgroup (Figure 2A), a significant difference in survival between CACS groups was observed (P=0.0031), with higher CACS levels being associated with progressively worse survival. Notably, patients with CACS ≥400 had the lowest survival rates. Conversely, the female subgroup (Figure 2B) did not show significant differences in survival between CACS groups (P=0.81), indicating that CACS might be less predictive of short-term survival in women compared to men. We then examined survival in patients stratified by smoking status. Among smokers (Figure 2C), although there was a trend of worse survival with increasing CACS, this did not reach statistical significance (P=0.12). Similarly, no significant survival differences were observed in non-smokers (Figure 2D, P=0.14), suggesting that smoking status may not strongly modify the relationship between CACS and survival in the short term. In the hypertensive group (Figure 2E), survival was significantly different across CACS groups (P=0.021), with the CACS ≥400 group having the poorest survival. However, in non-hypertensive individuals (Figure 2F), there was no significant survival difference between CACS groups (P=0.53), indicating that the prognostic value of CACS may be more pronounced in hypertensive patients.

Figure 2 Kaplan-Meier survival curves stratified by CACS across various subgroups. (A) Kaplan-Meier survival curve for men shows a significant difference between groups with different CACS (P=0.0031). (B) Kaplan-Meier survival curve for women reveals no significant difference between the groups based on CACS (P=0.81). (C) Kaplan-Meier survival curve for smokers indicates no significant difference between the CACS groups (P=0.12). (D) Kaplan-Meier survival curve for non-smokers shows no significant difference in survival across CACS groups (P=0.14). (E) Kaplan-Meier survival curve for individuals with hypertension shows significant survival differences across CACS groups (P=0.021). (F) Kaplan-Meier survival curve for individuals without hypertension reveals no significant survival differences between CACS groups (P=0.53). CACS, coronary artery calcium score.

Discussion

In this study, we demonstrated that CACS is a significant independent predictor of survival in patients undergoing CCTA. For participants in our study, smoking intensity and LDL levels correlate with higher CACS, aligning with another study showing oxidative stress and lipid deposition drive coronary calcification (14). The association between longer diabetes duration and higher CACS highlights the role of chronic glycemic exposure in accelerating atherosclerosis. While CACS is often studied in asymptomatic populations for long-term risk stratification, our study focused on patients with symptoms prompting CCTA evaluation (e.g., stable angina in 73% of cases). This design aligns with clinical practice, where CACS is increasingly used to guide management in symptomatic patients with intermediate pre-test probability of CAD. Our findings align with previous studies that have shown a strong correlation between elevated CACS and adverse cardiovascular outcomes, including MACE and all-cause mortality (15-17). Specifically, patients with a CACS ≥400 exhibited significantly lower survival probabilities compared to those with lower scores, reinforcing the value of CACS as a non-invasive marker for risk stratification. This underscores the clinical utility of CACS in predicting future cardiovascular events, particularly in high-risk populations. Overall, CACS quantifies calcified plaque burden, CCTA provides comprehensive anatomical data, including non-calcified plaque detection and stenosis severity grading. Our findings align with studies showing that CACS ≥400 correlates with severe stenosis (60% of cases), highlighting its utility in identifying high-risk anatomy. However, CCTA’s ability to detect non-calcified plaques (73% in CACS ≥400) underscores its complementary role in risk assessment, particularly in patients with moderate CACS [101–399] who may harbor unstable non-calcified lesions. Future studies should integrate both CACS and CCTA-derived plaque characteristics into predictive models. Meanwhile, our findings highlighted that even in this symptomatic cohort, CACS still retained significant prognostic value, particularly in high-risk subgroups such as hypertensive patients. Future studies should compare these results with asymptomatic cohorts to assess generalizability. This finding aligns with prior evidence suggesting that calcium deposition can also contribute to vascular inflammation in other cardiovascular regions (18).

The gender-specific analysis revealed that CACS is a more robust predictor of survival in men compared to women. In male patients, the risk of cardiovascular events increased substantially with higher CACS levels, as reflected in both Kaplan-Meier survival analysis and Cox proportional hazards regression. Conversely, no significant association was observed between CACS and short-term survival in female patients. This may suggest that other factors, such as hormonal differences or alternative pathophysiological mechanisms, could influence cardiovascular risk in women, diminishing the predictive power of CACS. Although we acknowledged the limited sample size (n=100) as a potential limitation, which may compromise statistical power for detecting small effect sizes or subgroup interactions, the observed associations—such as the 4.76-fold increased risk (adjusted HR 4.76, 95% CI: 1.82–12.45, P=0.012) for CACS ≥400 versus CACS =0—demonstrate robustness within this cohort. This is further supported by high model concordance (C-index 0.851; 95% CI: 0.792–0.910). Our results align with prior studies of comparable size (n=80–150) that established the prognostic role of CACS in symptomatic populations (17,19). To address these limitations, future research should prioritize: (I) multicenter collaborations to expand cohort diversity and statistical power for subgroup analyses (e.g., females with limited event rates); (II) refined risk stratification models incorporating CACS with biomarkers (e.g., high-sensitivity troponin) to enhance precision in gender-specific prognostication.

Hypertension was another significant modifier of the relationship between CACS and survival. In hypertensive patients, those with a CACS ≥400 had the poorest survival, while non-hypertensive individuals did not exhibit significant differences in survival across CACS groups. This suggests that hypertension may amplify the detrimental effects of CAC, possibly due to synergistic mechanisms that accelerate atherosclerosis progression. Given that hypertension is a common comorbidity in patients at risk of cardiovascular disease, our results highlight the importance of incorporating both CACS and blood pressure management into comprehensive risk assessment models for this population.

Finally, the multivariate Cox proportional hazards model confirmed that CACS remained an independent predictor of mortality after adjusting for age, smoking status, diabetes, and dyslipidemia. Interestingly, while smoking and hypertension were also significant predictors of survival, other factors like diabetes and dyslipidemia did not show a strong independent association with increased mortality risk in this cohort. This suggests that while these conditions are traditionally associated with cardiovascular risk, their impact on short-term survival may be more variable, particularly when considered in the context of CACS. Future studies should aim to further explore these interactions to optimize cardiovascular risk prediction models.

Conclusions

This study demonstrates that CACS is a significant independent predictor of survival in patients undergoing CCTA. Higher CACS levels, particularly CACS ≥400, are associated with a significantly increased risk of cardiovascular events and all-cause mortality, especially in male and hypertensive patients. As a non-invasive risk assessment tool, CACS can effectively assist clinicians in identifying high-risk patients and formulating more targeted treatment strategies. However, the prognostic value of CACS is relatively weaker in female and non-hypertensive patients, suggesting that further research is needed to better understand the cardiovascular risk characteristics in these subgroups. This study has several limitations requiring consideration. First, the modest sample size (n=100) may restrict generalizability and underpower subgroup analyses in key demographic subgroups, including female patients (37% of cohort) and normotensive individuals. Second, the single-center retrospective design poses inherent selection bias risks, though prospective follow-up protocols [median 4.2 years, interquartile range (IQR), 3.1–5.0] and centralized CACS adjudication by ≥2 experienced radiologists attenuated potential confounding. Third, as previously emphasized, the inclusion of patients with symptoms necessitating CCTA may constrain the generalizability of our findings to asymptomatic populations. Notwithstanding this, the study design intentionally mirrors contemporary clinical practice where CACS serves as a refinement tool for cardiovascular risk stratification in symptomatic cohorts. In summary, CACS combined with other clinical characteristics provides a more comprehensive risk stratification, contributing to improved management and prognosis of cardiovascular diseases.

Acknowledgments

We appreciate the institutional support provided by Department of Radiology, Dangyang People’s Hospital, Hubei, China.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-549/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-549/dss

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-549/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Ethics Committee of Dangyang People’s Hospital (No. LL-2024-005-01), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827-32. [Crossref] [PubMed]
2. Gupta A, Bera K, Kikano E, Pierce JD, Gan J, Rajdev M, Ciancibello LM, Gupta A, Rajagopalan S, Gilkeson RC. Coronary Artery Calcium Scoring: Current Status and Future Directions. Radiographics 2022;42:947-67. [Crossref] [PubMed]
3. Kwok CS, Bennett S, Lip GYH. Coronary artery calcium score and its association with stroke: A systematic review and meta-analysis. Eur J Clin Invest 2023;53:e13892. [Crossref] [PubMed]
4. Choi HYJ. Coronary Artery Calcium Scoring for Prevention of Cardiovascular Disease. Am Fam Physician 2022;106:93-4.
5. Menown IB, Davies S, Gupta S, Kalra PR, Lang CC, Morley C, Padmanabhan S. Resting heart rate and outcomes in patients with cardiovascular disease: where do we currently stand? Cardiovasc Ther 2013;31:215-23. [Crossref] [PubMed]
6. Khan SS, Post WS, Guo X, Tan J, Zhu F, Bos D, Sedaghati-Khayat B, van Rooij J, Aday A, Allen NB, Bos MM, Uitterlinden AG, Budoff MJ, Lloyd-Jones DM, Mosley JD, Rotter JI, Greenland P, Kavousi M. Coronary Artery Calcium Score and Polygenic Risk Score for the Prediction of Coronary Heart Disease Events. JAMA 2023;329:1768-77. [Crossref] [PubMed]
7. Zhu Y, Wu J, Wei D, Huang G, Jiang Y, Feng X. Application of computed tomography angiography imaging in coronary heart disease screening and analysis of associated risk factors. Quant Imaging Med Surg 2024;14:8227-37. [Crossref] [PubMed]
8. Valenti V, Ó Hartaigh B, Heo R, Cho I, Schulman-Marcus J, Gransar H, Truong QA, Shaw LJ, Knapper J, Kelkar AA, Sandesara P, Lin FY, Sciarretta S, Chang HJ, Callister TQ, Min JK. A 15-Year Warranty Period for Asymptomatic Individuals Without Coronary Artery Calcium: A Prospective Follow-Up of 9,715 Individuals. JACC Cardiovasc Imaging 2015;8:900-9. [Crossref] [PubMed]
9. Lei MH, Wu YL, Chung SL, Chen CC, Chen WC, Hsu YC. Coronary Artery Calcium Score Predicts Long-Term Cardiovascular Outcomes in Asymptomatic Patients with Type 2 Diabetes. J Atheroscler Thromb 2021;28:1052-62. [Crossref] [PubMed]
10. Mahmood T, Shapiro MD. Coronary artery calcium testing in low-intermediate risk symptomatic patients with suspected coronary artery disease: An effective gatekeeper to further testing? PLoS One 2020;15:e0240539. [Crossref] [PubMed]
11. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, Himmelfarb CD, Khera A, Lloyd-Jones D, McEvoy JW, Michos ED, Miedema MD, Muñoz D, Smith SC Jr, Virani SS, Williams KA Sr, Yeboah J, Ziaeian B. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation 2019;140:e596-646. [Crossref] [PubMed]
12. Liang S, Wang C, Zhang J, Liu Z, Bai Y, Chen Z, Huang H, He Y. Triglyceride-glucose index and coronary artery disease: a systematic review and meta-analysis of risk, severity, and prognosis. Cardiovasc Diabetol 2023;22:170. [Crossref] [PubMed]
13. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol 2007;165:710-8. [Crossref] [PubMed]
14. Beck J, Ferrucci L, Sun K, Fried LP, Varadhan R, Walston J, Guralnik JM, Semba RD. Circulating oxidized low-density lipoproteins are associated with overweight, obesity, and low serum carotenoids in older community-dwelling women. Nutrition 2008;24:964-8. [Crossref] [PubMed]
15. Nazifova-Tasinova NF, Atanasov AA, Pasheva MG, Yotov YT, Gerova DI, Vankova DG, Todorova MN, Ivanova DG, Kiselova-Kaneva YD, Galunska BT. Circulating uncarboxylated matrix Gla protein in patients with atrial fibrillation or heart failure with preserved ejection fraction. Arch Physiol Biochem 2022;128:1619-29. [Crossref] [PubMed]
16. Ambale-Venkatesh B, Yang X, Wu CO, Liu K, Hundley WG, McClelland R, Gomes AS, Folsom AR, Shea S, Guallar E, Bluemke DA, Lima JAC. Cardiovascular Event Prediction by Machine Learning: The Multi-Ethnic Study of Atherosclerosis. Circ Res 2017;121:1092-101. [Crossref] [PubMed]
17. Nasir K, Cainzos-Achirica M. Role of coronary artery calcium score in the primary prevention of cardiovascular disease. BMJ 2021;373: [Crossref] [PubMed]
18. Mongiardo A, Curcio A, Spaccarotella C, Parise S, Indolfi C. Molecular mechanisms of restenosis after percutaneous peripheral angioplasty and approach to endovascular therapy. Curr Drug Targets Cardiovasc Haematol Disord 2004;4:275-87. [Crossref] [PubMed]
19. Azmiardi A, Murti B, Febrinasari RP, Tamtomo DG. The effect of peer support in diabetes self-management education on glycemic control in patients with type 2 diabetes: a systematic review and meta-analysis. Epidemiol Health 2021;43:e2021090. [Crossref] [PubMed]