PDF
Citation
Print
Introduction
Kidney cancer is the 14th most frequently diagnosed cancer worldwide, with over 430,000 new cases reported in 2022 [1,2]. The incidence of kidney cancer has nearly doubled over the past few decades, with an annual increase of approximately 2% [1,2]. Established risk factors for kidney cancer include hypertension, smoking, obesity, diabetes, and chronic kidney disease (CKD) [3]. Elucidating potential risk factors associated with kidney cancer is essential for reducing the rapidly growing burden of the disease.
Although alcohol is classified as a carcinogen [4], the relationship between alcohol consumption and kidney cancer risk remains inconclusive, with studies reporting either an inverse [5-11] or null [12,13] association. The inverse association between alcohol consumption and kidney cancer risk is supported by evidence suggesting that moderate alcohol intake increases insulin sensitivity, reduces inflammation, and promotes diuresis [14,15]. However, the effect of alcohol on kidney cancer risk may vary depending on glycemic status. Hyperglycemia, including diabetes and prediabetes, is associated with insulin resistance, chronic inflammation, and renal dysfunction [16,17].
Although a potential interaction between alcohol’s insulin-sensitizing effect and hyperglycemia-related insulin resistance is proposed, no study has examined whether the dose-response relationship between alcohol consumption and kidney cancer risk varies based on glycemic status. Moreover, previous cohort studies were limited by small sample sizes [8-11] and insufficient adjustment for important confounders, such as diabetes and CKD [5-11]. Previous case–control studies on alcohol consumption and kidney cancer risk were significantly limited by recall, selection, and survivorship biases [18,19].
Global alcohol consumption has been steadily increasing, with per capita consumption rising from 5.9 L in 1990 to 6.5 L in 2017, and projections indicate that it could reach 7.6 L by 2030 [20]. Additionally, diabetes and prediabetes affect more than half of adults in most developed countries [21,22]. Investigating the interaction between alcohol consumption and hyperglycemia could enable personalized risk assessment and help reduce the growing burden of kidney cancer.
Thus, we performed a large-scale, nationwide cohort study to evaluate the dose-response relationship between alcohol consumption and kidney cancer risk by glycemic status (normoglycemia, prediabetes, and diabetes), adjusting for potential confounders. Over 9 million cancer-free participants who underwent the national health screening in Korea were followed for up to 10 years.
Materials and Methods
1. Data source and study population
The South Korean National Health Insurance Service (KNHIS) is a single-payer national health insurance system providing coverage for 97% of the South Korean population. The KNHIS also conducts biennial standardized national health screenings, with approximately 76% of the eligible population participating [23]. Data from the KNHIS were analyzed, comprising insurance claims data (including medical treatments, prescribed medications, and diagnoses coded using the International Classification of Diseases, 10th revision, clinical modification [ICD-10-CM]) and health examination data. The health examination data included laboratory test results, anthropometric measurements, and self-reported health-related behaviors.
The participant flowchart is shown in Fig. 1. Initially, 10,585,844 individuals aged 20 years or older who underwent national health screening between January 1, 2009, and December 31, 2009, were included in the study. Participants with a prior cancer diagnosis (n=162,512) were excluded. To minimize the potential for reverse causality and early detection bias, participants who were diagnosed with cancer or died within the first year following cohort enrollment were excluded from the analysis (n=93,862). Additionally, individuals missing data on variables of interest were excluded (n=837,139). Ultimately, 9,492,331 participants were followed until kidney cancer diagnosis, death, or December 31, 2018, whichever occurred first.
2. Definition of alcohol consumption
Alcohol consumption data were collected through self-administered standardized questionnaires during the national health screening. Participants provided self-reported data on alcohol consumption frequency, which was categorized into the following groups: 0, 1-2, 3-4, or 5-7 days per week. The amount of alcohol consumed per occasion was reported in terms of the number of glasses, with one glass of alcoholic beverage containing approximately 12 g of ethanol. The average daily alcohol consumption was calculated based on weekly consumption frequency and the amount consumed per occasion. Participants were classified into three groups based on their average daily alcohol consumption: non-drinkers, light-to-moderate drinkers (< 30 g/day for men; < 20 g/day for women), and heavy drinkers (≥ 30 g/day for men; ≥ 20 g/day for women) [24].
3. Definition of glycemic status
Fasting plasma glucose (FPG) levels were measured for each participant as part of the national health screening program. After an overnight fast, blood samples were collected by certified healthcare professionals at KNHIS-approved hospitals, adhering to quality control measures. Glycemic status was determined according to the American Diabetes Association criteria [25]. Normoglycemia was defined as an FPG level of < 100 mg/dL, prediabetes as an FPG level of 100-125 mg/dL, and diabetes as an FPG level of ≥ 126 mg/dL or the use of oral or injectable antidiabetic medications, with at least one insurance claim per year under ICD-10-CM codes E11–E14.
4. Definition of incident kidney cancer
The primary outcome of this study was the incidence of newly diagnosed kidney cancer cases recorded between January 2009 and December 2018. These diagnoses were based on the ICD-10-CM code (C64) during hospitalization, along with a cancer reimbursement code (V193). Physicians and medical institutions are required to use the V193 reimbursement code when certifying cancer diagnoses to facilitate a reduced co-payment rate of 5% for all cancer-related treatments and examinations. The National Cancer Registry is responsible for recording all patients with a confirmed cancer diagnosis using this V code.
5. Definition of the clinical variables
Anthropometric data, including height, weight, and waist circumference, were collected during the national health screening. Body mass index (BMI) was determined by dividing the participant’s weight (kg) by the square of their height (m²). According to Korean standards, obesity is defined as a BMI ≥ 25 kg/m2. Blood pressure measurements, including systolic and diastolic values, were taken in a seated position after a 5-minute rest. Serum levels of total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and creatinine were assessed. The estimated glomerular filtration rate (eGFR) was determined using the creatinine-based Modification of Diet in Renal Disease equation. CKD was defined as an eGFR of < 60 mL/min/1.73 m2.
Dyslipidemia was defined as either a total serum cholesterol level > 240 mg/dL or the presence of an ICD-10-CM code (E78) accompanied by lipid-lowering medication prescriptions. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or the presence of ICD-10-CM codes (I10-I13, I15) with antihypertensive medication prescriptions. The low-income group included participants who paid the lowest quartile of insurance fees or received free medical care.
Information on health-related behaviors was collected through a self-administered standardized questionnaire as part of the national health screening program. Smoking status was categorized as never a smoker (fewer than 100 cigarettes smoked in their lifetime), former smoker, or current smoker. Regular physical activity was defined as engaging in at least 20 minutes of vigorous-intensity physical activity three times per week or 30 minutes of moderate-intensity physical activity five times per week.
6. Statistical analysis
The incidence rate of kidney cancer was calculated per 100,000 person-years, stratified by alcohol consumption levels and glycemic status. The relationship between alcohol consumption and kidney cancer risk based on glycemic status was analyzed in a dose-response manner using hazard ratios (HRs) and 95% confidence intervals (CIs) obtained from multivariable Cox proportional hazards regression models. Model 1 was unadjusted, while model 2 adjusted for age and sex. Model 3 included adjustments for age, sex, smoking status, physical activity level, and income level. Model 4 was further adjusted for BMI, hypertension, dyslipidemia, and CKD. Stratified analysis was used to assess potential modification effects by covariates, such as age, sex, smoking status, physical activity level, income level, hypertension, dyslipidemia, CKD, and obesity status. The multiplicative interaction between alcohol and glycemic status was examined using the likelihood ratio test. All statistical tests were two-sided, and statistical significance was defined as p < 0.05. All analyses were conducted using SAS software ver. 9.3 (SAS Institute Inc.).
Results
1. Baseline characteristics
Over 78.1 million person-years of follow-up (mean duration, 8.2±0.9 years), a total of 12,381 individuals were newly diagnosed with kidney cancer. Table 1 presents the participants’ baseline characteristics. Patients diagnosed with kidney cancer were older and exhibited higher FPG levels and BMI (all p < 0.01). Additionally, they were more likely to be men, heavy drinkers, current smokers, and to have diabetes, dyslipidemia, hypertension, and CKD (all p < 0.01).
2. Dose-response association between alcohol consumption and kidney cancer risk according to glycemic status
Table 2 illustrates the dose-response association between alcohol consumption and kidney cancer risk by glycemic status. Among individuals with normoglycemia, light-to-moderate alcohol consumption was associated with a 6% reduced risk of kidney cancer (model 4: HR, 0.94; 95% CI, 0.89 to 0.99).
Individuals with diabetes who consumed light-to-moderate or heavy amounts of alcohol had an increased risk of kidney cancer in a dose-dependent manner compared to non-drinkers with normoglycemia (model 4: HR, 1.12; 95% CI, 1.03 to 1.22 and HR, 1.24; 95% CI, 1.09 to 1.42, respectively) (p for trend < 0.01).
S1 Table shows that consuming 3-4 glasses of alcohol on a single occasion or consuming alcohol 3-4 times per week was associated with a reduced risk of kidney cancer in those with normoglycemia (model 4: HR, 0.88; 95% CI, 0.81 to 0.95 and HR, 0.92; 95% CI, 0.84 to 0.997, respectively) but not in individuals with prediabetes or diabetes. Individuals with prediabetes or diabetes who consumed ≥ 8 glasses of alcohol on a single occasion had an increased risk of kidney cancer (model 4: HR, 1.11; 95% CI, 1.01 to 1.22 and HR, 1.31; 95% CI, 1.16 to 1.48, respectively).
3. Subgroup analysis of the dose-response association between alcohol consumption and kidney cancer risk by glycemic status
Table 3 presents the subgroup analysis of the dose-response association between alcohol consumption and kidney cancer risk by glycemic status. No significant differences were observed between the subgroups based on sex, smoking status, physical activity, income level, dyslipidemia, or CKD (all p for interaction > 0.05). The association between alcohol consumption and kidney cancer risk in relation to diabetes status differed by age, hypertension, and obesity (p for interaction < 0.001, < 0.001, and 0.002, respectively).
Discussion
In this large-scale nationwide cohort study, a modest U-shaped dose-response relationship between alcohol consumption and kidney cancer risk was observed exclusively among individuals with normoglycemia, after adjusting for potential confounding factors. Among individuals with prediabetes, no significant association was observed between alcohol consumption and kidney cancer risk. A dose-response association was observed in individuals with diabetes, with higher alcohol consumption levels corresponding to an increased risk of kidney cancer.
Despite the carcinogenic effects of alcohol, several previous studies have reported an inverse association between alcohol consumption and kidney cancer risk [5-11]. A cohort study conducted in the United States reported that, compared with alcohol consumption of > 0 to < 5 g per day, the multivariate relative risk (95% CI) for 15 to < 30 g per day and ≥ 30 g per day was 0.75 (0.63-0.90) and 0.71 (0.59-0.85), respectively [6]. Similarly, a pooled analysis of 12 cohort studies found that, compared with nondrinking, alcohol consumption of ≥ 15 g/day was associated with a decreased risk of kidney cancer (pooled multivariable relative risk, 0.72; 95% CI, 0.60 to 0.80) [5]. In contrast, another cohort study, which included U.S. residents, reported no significant association between alcohol consumption and kidney cancer risk [13]. Prior cohort studies were small-scale [8-11] or did not account for important confounders, such as diabetes or CKD [5-11]. Moreover, prior case–control studies were limited by survivorship, selection, and recall biases [18,19]. Notably, despite a possible interaction between alcohol consumption and hyperglycemia, whether the dose-response relationship between alcohol consumption and kidney cancer risk differs by glycemic status has not been investigated.
We provide new evidence that the association between alcohol consumption and kidney cancer risk varies based on the glycemic status of the individual. Healthcare providers should tailor patient education concerning alcohol consumption and kidney cancer risk, especially given the global rise in both alcohol consumption and diabetes prevalence.
The mechanisms underlying the differing dose-response effects of alcohol consumption on kidney cancer risk by glycemic status remain uncertain. However, we propose the following explanations. First, the mechanisms underlying these associations may involve the modulation of insulin signaling pathways by alcohol, which has been shown to affect cellular metabolism in ways that differ between normoglycemic and diabetic individuals. Light-to-moderate alcohol consumption may enhance insulin sensitivity in individuals with normoglycemia, potentially lowering their risk of kidney cancer [14,15]. Conversely, diabetes and prediabetes, conditions associated with insulin resistance and hyperinsulinemia, might attenuate the effect of alcohol on insulin sensitivity, thereby increasing kidney cancer risk. Second, the anti-inflammatory effects of light-to-moderate alcohol consumption may contribute to the decreased cancer risk observed in individuals with normoglycemia [26]. However, this effect may be absent in individuals with prediabetes or diabetes, who experience chronic inflammation and metabolic changes [27]. Finally, the diuretic effect of alcohol consumption might reduce kidney cancer risk by decreasing the contact time between carcinogens and renal epithelial cells. This is supported by studies demonstrating that coffee and tea consumption reduces kidney cancer risk [28]. However, the diuretic effects of alcohol might differ in individuals with hyperglycemia due to altered urine filtration [29].
Our study possesses several strengths. First, this large-scale nationwide cohort study utilized data from over 9 million individuals with up to 10 years of follow-up. The KNHIS database provides an accurate record of clinical progression of the cohort after enrollment. This allowed for a robust comparison of the dose-response relationship between alcohol consumption and kidney cancer risk across different glycemic statuses, resulting in reliable risk estimates for kidney cancer. Second, longitudinal data on lifestyle factors, blood test results, anthropometric measurements, and extensive medical records were collected. Alcohol consumption was assessed in more than 9 million participants using standardized questionnaires, minimizing the potential for recall bias. Third, to ensure high diagnostic accuracy, we identified kidney cancer cases using both the ICD-10-CM diagnostic code (C64) during hospitalization and the national registration code (V193). Fourth, our analyses were adjusted for potential confounding variables, including BMI, hypertension, smoking status, and CKD.
However, our study has some limitations. First, alcohol consumption was self-reported, which may introduce underreporting or misclassification, although such data are generally considered reliable [30]. Second, we lacked information on the type of alcohol consumed. Third, family history of kidney cancer was not included in our dataset. Fourth, we did not have access to pathological subtype data for kidney cancers; however, approximately > 90% of kidney cancers are renal cell carcinoma [3].
In conclusion, this nationwide cohort study demonstrated a modest U-shaped dose-response relationship between alcohol consumption and kidney cancer risk, which was observed exclusively in individuals with normoglycemia. Conversely, individuals with diabetes exhibited a dose-dependent increase in kidney cancer risk with alcohol consumption. These findings highlight the need for tailored patient education and personalized risk assessments concerning alcohol consumption and kidney cancer risk to address the growing burden of this disease. Given the increasing prevalence of both alcohol consumption and diabetes, healthcare professionals should incorporate these findings into preventive strategies aimed at reducing kidney cancer risk in at-risk populations. Additional studies are needed to elucidate the mechanisms by which alcohol consumption and hyperglycemia interact in the process of kidney carcinogenesis.
XML Download