Published online Feb 14, 2024.
https://doi.org/10.51789/cmsj.2024.4.e1
Associations Between Fasting Glucose and Cardiovascular Disease Mortality in Cancer Survivors: A Population-based Cohort Study
, MD, PhD,1,2
Sang-Wook Yi, MD, PhD,3
Sang Joon An, MD,4
Jee-Jeon Yi, MD,5
Sang-Hyun Ihm, MD, PhD,2,6
So-Young Lee, MD,1,2
Jong-Chan Youn, MD, PhD,1,2
Woo-Baek Chung, MD,1,2
Hae Ok Jung, MD, PhD,1,2
and Ho-Joong Youn, MD, PhD1,2
Abstract
Background and Objectives
Cardiovascular disease (CVD) prevention is an essential part of care among cancer survivors. Herein, we aimed to investigate the relationship between fasting glucose levels and the risk of CVD mortality among this population.
Methods
Using data from the Korea National Health Insurance database, we identified patients with cancer who underwent health examination during 2009–2010 and survived for more than 5 years as of December 31, 2010. Individuals with a known CVD were excluded from the study. Individuals were categorized into 5 groups based on their fasting glucose levels. The clinical endpoint was CVD-related mortality.
Results
Among 177,774 cancer survivors (mean age 60.7±11.9 years, 57.8% female), we observed a reverse L-shaped association between fasting glucose levels and overall CVD mortality. After adjusting for covariates, the hazard ratios (95% confidence intervals [CIs]) were 1.14 (0.72–1.82), 1.00 (reference), 1.02 (0.94–1.11), 1.40 (1.24–1.58), and 2.28 (1.88–2.77) for fasting glucose <70, 70–99, 100–125, 126–179, and ≥180 mg/dL, respectively. Stratifying the results by age, increased risk of hypoglycemia (fasting glucose <70 mg/dL) for CVD mortality was significant in younger adults (<65 years) but not in older adults (≥65 years). Among younger adult cancer survivors, the adjusted hazard ratios (95% CIs) were 2.97 (1.32–6.69), 1.00 (reference), 0.95 (0.74–1.21), 1.58 (1.13–2.21), and 1.97 (1.19–3.26) for fasting glucose <70, 70–99, 100–125, 126–179, and 180 mg/dL, respectively.
Conclusions
Elevated fasting glucose levels are associated with CVD mortality in cancer survivors. Additionally, hypoglycemia risk should not be overlooked in the younger adult subgroup.
GRAPHICAL ABSTRACT
INTRODUCTION
With the introduction and advances in cancer therapy, many patients with cancer have survived. In the US, there are approximately 18.1 million cancer survivors as of January 2022, representing 5.4% of the total population.1, 2 In Korea, the 5-year relative survival rate for patients with cancer is 70%.3 Thus, long-term management of cancer survivors, the so-called cancer survivorship care, is an urgent health care problem.4, 5 A previous cohort study involving more than 41,000 patients with cancer reported that cardiovascular disease (CVD) is the second most common cause of death in patients with cancer,6 which calls for regular surveillance and management of CVD in these patients.
Unfortunately, patients with cancer or their clinicians often overlook their metabolic and cardiovascular health, as most concerns tend to be focused on cancer treatment and the prevention of recurrence. An elevated fasting glucose level is a critical metabolic risk factor for CVD. Recent studies have shown that optimal fasting glucose levels may differ according to age and comorbidities.7, 8, 9, 10 However, the relationship between fasting glucose and CVD mortality in patients with cancer is relatively unexplored. Most previous studies have focused on the relationship between elevated fasting glucose levels and cancer-related outcomes, such as cancer incidence and cancer mortality, in the general population.11, 12, 13 Patients with cancer may have different metabolic profiles,14, 15 and the ideal glucose level for better CVD outcomes may differ from that of the general population.
From this perspective, we designed the current study to investigate the association between fasting glucose levels and CVD mortality in cancer survivors who were diagnosed with cancer and survived for more than five years after cancer diagnosis. In this study, we confined the study population to cancer survivors and not all patients with cancer, given that complications from CVD could be an important issue after cancer stabilization.
METHODS
Study population
We used data from the Korean National Health Insurance Service (NHIS) database linked to the Korean General Health Screening Database. The NHIS offers compulsory health insurance coverage for 97% of the Korean population along with biennial health screenings conducted nationwide. The NHIS database encompasses a range of information, including demographic characteristics, diagnoses coded by the International Classification of Disease (ICD), prescriptions, mortality data, and details from health screening examinations.16
To identify cancer survivors, our initial step was to identify those aged 18–99 years who were diagnosed with cancer between 2002 and 2005 and who underwent health screening examinations in 2009–2010 (n=190,232). We defined an individual as having cancer if they had been hospitalized with an ICD code for cancer (C00-97) or had received outpatient or inpatient care with a critical condition code for cancer (V193, V194). We excluded individuals with missing information on health examinations (n=593), known CVD at the time of health examination (n=9,538), and death prior to the index date of January 1, 2011 (n=2,327). The final analysis included a total of 177,774 cancer survivors. A detailed flowchart illustrating the study population is provided in Supplementary Figure 1. The current study was approved by the Institutional Review Board (HDT2021-03-010), and the requirement for informed consent was waived because the NHIS provided anonymized data under a strict confidentiality protocol.
Data collection
Venous samples were collected after overnight fasting. Trained staff members measured the blood pressure at least twice after a minimum of 5 minutes of rest. Body mass index (BMI) was determined by dividing weight by the square of height (kg/m2). Information on alcohol consumption, smoking status, physical activity, and pre-diagnosis of CVD and cancer was gathered using a questionnaire completed by the participants. Health examinations and data collection adhered to the standard protocols established by the Ministry of Health and Welfare. Further details regarding NHIS health examinations can be found elsewhere.16, 17 Household income status was categorized based on the quartiles of all NHIS beneficiaries. The causes of death were tracked through national death records obtained using the Microdata Integrated Service. This service links health-screening data with statistics from the National Statistical Office using unique personal identification numbers. The study population was followed up until the end of the study in December 2018.
Exposure and the study outcomes
In the current analysis, the exposure was the fasting glucose level, and the study outcome was CVD mortality. The fasting glucose level was categorized into 5 levels as follows: <70, 70–99 (reference), 100–125, 126–179, and ≥180 mg/dL. The primary outcome was death from overall CVD using ICD codes (I00-I99). The secondary outcome was defined as death from CVD. ICD-10 codes for individual CVD are as follows: hypertensive disease (I10-I13), ischemic heart disease (I20-I25), acute myocardial infarction (I21), total stroke (I60-I69), ischemic stroke (I63), hemorrhagic stroke (I60-I62), subarachnoid hemorrhage (I60), intracerebral hemorrhage (I61-I62), heart failure (I50), and sudden cardiac death (I46).
Statistical analysis
Baseline characteristics were reported as means and standard deviations for continuous variables and numbers with percentages for categorical variables. To determine the effect of fasting glucose on CVD mortality, cause-specific Cox regression analysis was employed while accounting for potential covariates (age, sex, smoking status, alcohol consumption frequency, income status, physical activity, systolic blood pressure, BMI, triglycerides, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol). The results are presented as hazard ratios (HRs) and 95% confidence intervals (CIs). Data of individuals who died of other causes or reached the end of the follow-up period were treated as censored data. Subgroup analyses were conducted according to age and sex. Sensitivity analyses were performed by excluding those who died within the first 3 years and those with a prior history of diabetes. Additionally, assuming a linear relationship, the risk associated with every 18 mg/dL increase in fasting glucose was assessed for CVD mortality. Restricted cubic spline curves were plotted to explore the shape of the association between fasting glucose level and CVD mortality, adjusting for covariates with 4 knots, with a reference point set at a fasting glucose level of 90 mg/dL. All statistical analyses were performed using the SAS version 9.4 software (SAS Institute, Cary, NC, USA). All p-values were two-sided, and a p-value of less than 0.05 indicated statistically significance.
RESULTS
Baseline characteristics of the study population
In total, 177,774 cancer survivors were included in the current analysis. The mean age of the study population was 60.7±11.9 years, and 57.8% were female. The mean fasting glucose level was 100.3±24.9 mg/dL, and the mean BMI was 23.4±3.2 kg/m2. In the study population, 10.6% (n=18,927) had a history of diabetes. By cancer site, stomach cancer accounted for 21.3% of the cases, thyroid cancer for 15.9%, breast cancer for 12.5%, colorectal cancer for 10.7%, cervical cancer for 5.4%, lung cancer for 3.9%, leukemia/lymphoma for 3.9%, liver cancer for 3.8%, bladder cancer for 2.9%, prostate cancer for 1.8%, and other types of cancers for 17.9%. The baseline characteristics of the study population are summarized in Table 1.
Table 1
Baseline characteristics of the study population by fasting glucose levels
Association between fasting glucose and CVD mortality among cancer survivors
After adjustment for covariates, HR (95% CI) for overall CVD mortality was 1.14 (0.72–1.82), 1.00, 1.02 (0.94–1.11), 1.40 (1.24–1.58), and 2.28 (1.88–2.77) for fasting glucose levels of <70, 70–99 (reference), 100–125, 126–179, and ≥180 mg/dL, respectively (Table 2). We performed subgroup analysis by age groups (18–64 and ≥65 years), which demonstrated generally similar findings except for a significantly elevated risk of hypoglycemia in the younger group (adjusted HR, 2.97; 95% CI, 1.32–6.69) compared to the older group (adjusted HR, 0.86; 95% CI, 0.49–1.51; Table 3). Additionally, we performed a subgroup analysis by sex, which exhibited a similar pattern of risk profile across the spectrum of fasting glucose levels between the sexes (Supplementary Table 1). When we repeated the same analysis after excluding those who died within the first 3 years and those who had diabetes, similar results were obtained (Supplementary Table 2).
Table 2
HRs for overall cardiovascular mortality by fasting glucose
Table 3
Age-specific HRs for overall cardiovascular mortality by fasting glucose
Assuming a linear association in the range of 70–300 mg/dL, each 18 mg/dL (1 mmol/L) elevation in fasting glucose level was associated with a 11% increase in overall CVD mortality (adjusted HR, 1.11; 95% CI, 1.08–1.13). The associations were not substantially different between men (adjusted HR, 1.11; 95% CI, 1.08–1.14) and women (adjusted HR, 1.09; 95% CI, 1.05-1.14) or between younger (adjusted HR, 1.10; 95% CI, 1.03–1.17) and older (adjusted HR, 1.11; 95% CI, 1.08–1.14) adults.
Restricted cubic spline curves demonstrated a J-shaped-or reverse L-shaped association between fasting glucose levels and overall and individual CVD mortality (Figure 1). Particularly for atherosclerotic cardiovascular disease, which includes ischemic heart disease and ischemic stroke, a J-shaped association indicating an elevated risk toward both ends of the fasting glucose level was evident. The optimal fasting glucose level associated with the lowest risk of CVD death was approximately 90 mg/dL in cubic spline analysis. Outside this threshold, the risk of CVD deaths increased linearly. Consistent with the categorical analysis, the risk of hypoglycemia was more pronounced in adults aged <65 years (Figure 2).
Figure 1
CVD = cardiovascular disease.
Shape of association between fasting glucose and cardiovascular disease mortality in cancer survivors
Restricted spline curves for overall and individual CVD mortality are plotted by adjusting for covariates with four knots and with a reference point set at a fasting glucose level of 90 mg/dL. Overall, J-shaped- or reverse-L-shaped associations are observed between fasting glucose levels and CVD mortality.
Figure 2
Age-specific or sex-specific shape of association between fasting glucose and CVD mortality in cancer survivors.
CVD = cardiovascular disease.
DISCUSSION
In this population-based cohort study, 177,774 cancer survivors were examined to elucidate the association between fasting glucose levels and CVD mortality in patients with cancer. Our findings revealed that both hyperglycemia and hypoglycemia increased the risk of CVD death among cancer survivors. Specifically, we observed a graded risk elevation with hyperglycemia in the diabetic range (fasting glucose ≥126 mg/dL), which was consistent across age and sex. Notably, the risk of hypoglycemia-related complications varied according to age group and was found to be significant only among younger adult cancer survivors. To our knowledge, this is the first study to explicitly explore the association between fasting blood glucose levels and CVD mortality in patients with cancer. The current findings imply that regular surveillance and meticulous management of glucose levels are necessary even in individuals who have survived cancer.
The increasing survival of patients with cancer has led to a growing global interest in cardio-oncology, which focuses on addressing the cardiovascular and metabolic health of these individuals.4, 18, 19, 20, 21 The coverage range of cardio-oncology has expanded from a narrow spectrum of chemotherapy-induced cardiac dysfunction to a broader spectrum of cardiovascular health. It encompasses various aspects such as pre-cancer therapy CVD screening and risk assessment, cardiovascular monitoring during and after cancer therapy, management of cardiovascular side effects during and after cancer treatment, and long-term survivorship care. Among these aspects, long-term survivorship care is still easily overlooked or forgotten, despite its relevance.4, 18 A recent cohort study in Korea revealed that non-adherence to antihypertensive medication was common (2 in 3) among patients with cancer with hypertension. Importantly, non-adherence was associated with increased cardiovascular events and mortality.18 Considering that CVD is the second most common cause of death in patients with cancer,16 the prevention and management of CVD are mandatory parts of survivorship care for these patients.
In the current study, hyperglycemia in the diabetic range (≥126 mg/dL) was significantly associated with CVD mortality in cancer survivors, regardless of age and sex. Hyperglycemia in the impaired fasting glucose range (100–125 mg/dL) was also associated with increased CVD mortality, although the difference was not statistically significant. As the reference group in the categorical analysis had a relatively wide range of fasting glucose levels, including a lower normal value (70–99 mg/dL), we further examined the shape of the association using a cubic spline model. The cubic spline curve revealed that the nadir for overall CVD mortality occurred at a fasting glucose level of approximately 90 mg/dL. The curve demonstrated that the ideal fasting glucose level was relatively narrow, ranging from 85 to 99 mg/dL. Above or below this range, there was a sharp increase in CVD mortality, with a narrower 95% CI in the hyperglycemic than in the hypoglycemic range. Notably, the subsequent age-specific cubic spline curve clearly demonstrated that the risk associated with hyperglycemia was evident and consistent (as reflected by the narrow 95% CI) among older adult cancer survivors (aged ≥65 years) who are commonly encountered in clinical practice. Even a slight increase in fasting glucose levels was associated with a gradual increase in CVD mortality. These findings suggest that older cancer survivors with elevated fasting glucose levels may be at higher risk of subsequent CVD, necessitating close monitoring and proper glucose control. In fact, these results were somewhat unexpected because previous studies have indicated that older adults or individuals with comorbidities, who are typically considered a vulnerable population, tend to have a slightly higher ideal glucose level compared to the general healthy population.7, 9 Yi et al.7 previously reported that middle-aged and older adults with diabetes had a higher ideal fasting glucose level for mortality (90–130 mg/dL) than non-diabetic individuals or younger adults with diabetes (80–95 mg/dL). Similarly, Lee et al.9 reported that older adults (≥65 years) with diabetes had optimal fasting glucose levels of 110–124 mg/dL for all-cause mortality and 95–124 mg/dL for cardiovascular events. Unlike these studies, our study revealed that cancer survivors had an ideal glucose range of 85–99 mg/dL, similar to that in the general population.22 It is well known that patients with cancer pose a higher risk for developing diabetes in age- and sex-matched general population.23, 24 This is partly due to the shared risk factors between diabetes and cancer (for example, unhealthy diet, physical inactivity, and alcohol consumption) as well as cancer therapy itself such as steroids, radiotherapy, and androgen deprivation therapy. Regular surveillance of the glucose state and strict control of hyperglycemia may be an important health issue, even in patients with cancer.
The risk associated with hypoglycemia also requires careful examination and interpretation.25, 26 In older adult cancer survivors, the risk of hypoglycemia was not statistically significant. However, among younger adult cancer survivors (aged <65 years), the risk of CVD mortality increased nearly threefold in those with fasting glucose levels below 70 mg/dL compared to that in the reference group (with fasting glucose levels of 70–99 mg/dL). Notably, the risk associated with hypoglycemia was higher than that associated with hyperglycemia among younger adult cancer survivors (HRs for fasting glucose <70 mg/dL and ≥180 mg/dL were 2.94 and 1.89, respectively). Similar results were obtained in the sensitivity analysis, which excluded those who died within the first 3 years and those with known diabetes, although statistical significance was lost owing to the small number of CVD deaths. The precise mechanism underlying the increased risk of hypoglycemia in younger adults with cancer remains unclear. The coexistence of type 1 diabetes was considered a possible cause, but the elevated risk of hypoglycemia in younger adults seemed to persist even after excluding those with diabetes. Hypoglycemia triggers counter-regulatory responses involving the sympathetic nervous and adrenal systems, leading to an elevated heart rate, platelet and coagulation system activation, and increased susceptibility to cardiac arrhythmia, all of which are substrates for CVD events and mortality.27, 28 These counter-regulatory responses may be more pronounced in younger than in older adults. Although further data are required to validate our findings, our data suggest that hypoglycemia could be a dangerous sign, warranting meticulous glucose control (stringent glucose control while avoiding hypoglycemia), particularly in younger adult cancer survivors.
Generally, patients with cancer are considered a vulnerable population that requires careful attention to their nutritional support. Often, overnutrition or metabolic unhealthiness is overlooked in this population and rarely addressed in the clinical setting. Nevertheless, the significance of these factors may differ from that at the initial stage of cancer diagnosis and treatment, particularly among patients with cancer who have survived for more than 5 years. The present study clearly demonstrated that even a small elevation in glucose levels was linked to increased CVD mortality. Adopting proactive lifestyle modifications, such as regular physical activity, a balanced diet while limiting the consumption of sugary and processed foods, and ideal weight maintenance, should be encouraged in cancer survivors. Our study also reminds us that the risk of hypoglycemia in younger adults is substantial. Although one is prone to focus on strict glucose control in younger cancer survivors given their long-term exposure to the risk of hyperglycemia, we should bear in mind the risk of hypoglycemia.
The current study was conducted in a large population-based cohort; however, it has some limitations. We lacked information regarding other measures of glucose homeostasis, such as glucose tolerance, HbA1C, and glucose variability, which warrants additional studies. Additionally, our investigation was based on a one-time fasting glucose measurement at baseline. Nevertheless, it is crucial to recognize that the consequences of extended exposure to uncontrolled hyperglycemia, factoring in the duration of exposure, could potentially influence the results. Our data were based on the entire cancer population; therefore, specific cancer-type-related associations could not be identified. Finally, our study sample was exclusively composed of Koreans; thus, the findings may not be directly applicable to other ethnicities.
In this nationwide cohort of more than 170 thousand cancer survivors, both elevated and low glucose levels were associated with elevated CVD mortality. The optimal glucose level range was relatively narrow (85–99 mg/dL). Deviations from this range were linearly associated with elevated CVD mortality. Stringent management of hyperglycemia, including lifestyle modifications, may be warranted for older adult cancer survivors, who are frequently encountered in clinical settings. Among younger adult cancer survivors, meticulous glucose control while avoiding hypoglycemia would be beneficial. Further clinical attention and in-depth evidence are required to achieve long-term cardiovascular and metabolic health in cancer survivors.
SUPPLEMENTARY MATERIALS
Sex-specific HRs for overall cardiovascular mortality by the fasting glucoseSupplementary Table 1
HRs for overall cardiovascular mortality by the fasting glucose after excluding those who died within 3 yearsSupplementary Table 2
Study flow chart.Supplementary Figure 1
Funding:This work was supported by a grant from the Korean Society of CardioMetabolic Syndrome (2021-04).
Conflict of Interest:The authors have no conflicts of interest.
Author Contributions:
Conceptualization: Jung MH, Yi SW, An SJ, Lee SY, Chung WB, Youn HJ.
Data curation: Yi SW.
Formal analysis: Yi SW, Yi JJ.
Funding acquisition: Jung MH.
Investigation: Yi SW, An SJ, Ihm SH, Lee SY, Youn JC.
Methodology: Jung MH, Yi SW, Chung WB, Jung HO, Youn HJ.
Supervision: Yi SW, An SJ, Yi JJ, Ihm SH, Youn JC, Chung WB, Jung HO, Youn HJ.
Validation: Jung MH.
Writing - original draft: Jung MH.
Writing - review & editing: Yi SW, An SJ, Yi JJ, Ihm SH, Lee SY, Youn JC, Chung WB, Jung HO, Youn HJ.
ACKNOWLEDGMENTS
The authors thank the staff at the Big Data Steering Department of the National Health Insurance Service (NHIS) for providing data and support.
References
-
Kim H, Chung WB, Cho KI, et al. Diagnosis, treatment, and prevention of cardiovascular toxicity related to anti-cancer treatment in clinical practice: an opinion paper from the working group on cardio-oncology of the Korean Society of Echocardiography. J Cardiovasc Ultrasound 2018;26:1–25.
-
-
Lyon AR, López-Fernández T, Couch LS, et al. 2022 ESC guidelines on cardio-oncology developed in collaboration with the European Hematology Association (EHA), the European Society for Therapeutic Radiology and Oncology (ESTRO) and the International Cardio-Oncology Society (IC-OS). Eur Heart J Cardiovasc Imaging 2022;23:e333–e465.
-
-
Rodríguez-Gutiérrez R, Montori VM. Glycemic control for patients with type 2 diabetes mellitus: our evolving faith in the face of evidence. Circ Cardiovasc Qual Outcomes 2016;9:504–512.
-
Cited by
MeSH Terms
Figures
1 / 2
Tables
1 / 3
