https://orcid.org/0000-0001-9913-8883
Abstract
It has been unclear whether the risk of pancreatic cancer is different according to glucose levels.
To determine the association between fasting glucose levels and pancreatic cancer risk using prospectively collected nationwide population-based cohort data in Korea.
The National Health Insurance Service database of claims and preventive health check-up data recorded was used between 2009 and 2015.
A total of 25.4 million patients who had participated in a preventive health check-up between 2009 and 2013 were evaluated for pancreatic cancer incidence rates according to fasting glucose level.
The cumulative incidence rate for pancreatic cancer was calculated after grouping according to fasting glucose levels as follows: (i) low normal (<90 mg/dL), (ii) high normal (90 to 99 mg/dL), (iii) prediabetes level 1 (100 to 109 mg/dL), (iv) prediabetes level 2 (110 to 125 mg/dL), (v) diabetes (≥126 mg/dL), and (vi) diabetes on anti-diabetic medications.
The 5-year cumulative incidence rates (per 100,000) were as follows: (i) low normal = 32; (ii) high normal = 41; (iii) prediabetes level 1 = 50; (iv) prediabetes level 2 = 64; (v) diabetes = 75; and (vi) on anti-diabetic medications = 121. The risk of pancreatic cancer increased continuously with elevating fasting glucose levels (P < 0.0001). The incidence of pancreatic cancer increased significantly with increasing fasting blood glucose levels even after adjusting for age, sex, smoking, drinking, exercise, body mass index, and diabetes duration (P < 0.0001).
The cumulative incidence rate of pancreatic cancer significantly increased as the fasting glucose level elevated, even in populations with a normal glucose level range.
Pancreatic cancer is a highly fatal malignant neoplasm and has been associated with diabetes (1). Several cohort studies have evaluated the association between diabetes and pancreatic cancer and indicate diabetes is a risk factor rather than a consequence, although others have argued that diabetes is likely induced by pancreatic cancer (2–4).
A recent case-control study reported that patients diagnosed with pancreatic cancer had hyperglycemia for a mean period of 36 to 30 months before diagnosis, and there was an association between fasting glucose level and tumor volume or grade (5). However, it was not clear whether the risk of pancreatic cancer is different according to glucose levels in the general population.
Therefore, we used prospectively collected national cohort data in Korea to investigate the association between fasting glucose levels and pancreatic cancer risk in the general population.
Methods and Materials
Data sources
This study used the shared National Health Insurance Service (NHIS) database of claims and preventive health check-up data recorded between 2009 and 2015. The shared NHIS claims database contains medical information on all insurance claims for ∼50 million Koreans, which is >99% of the population. These data include each patient’s encrypted identification number, age, sex, primary diagnosis, secondary diagnoses, date of hospital visits, prescriptions received during inpatient and outpatient visits, hospital admissions, medical and surgical procedures, type of insurance (national health insurance or medical aid), and medical expenses. This study was approved by the Institutional Review Board of the Korean National Institute for Bioethics Policy. An informed-consent exemption was granted by the board. All methods were performed following relevant guidelines and regulations.
Definition of diabetes mellitus and pancreatic cancer
The diagnoses were coded according to the International Classification of Disease, 10th Revision (ICD-10). Patients with diabetes were defined from the insurance claims data as having at least one claim per year for the prescription of anti-diabetic medication under ICD-10 codes E11 to 14. Anti-diabetic medications included sulfonylureas, metformin, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, α-glucosidase inhibitors, meglitinides, and insulins. Type 1 diabetes patients who had claims under ICD-10 code E10 were excluded from this study. With the use of preventive health check-up data, collected between 2009 and 2013, the population was sorted into six groups, according to fasting glucose level, to calculate the hazard ratio (HR) for pancreatic cancer incidence as follows: (i) low normal (<90 mg/dL), (ii) high normal (90 to 99 mg/dL), (iii) prediabetes level 1 (100 to 109 mg/dL), (iv) prediabetes level 2 (110 to 125 mg/dL), (v) diabetes (≥126 mg/dL), and (vi) diabetes on anti-diabetic medications. Among the populations who had diabetes (≥126 mg/dL) or were treated with anti-diabetic medications, we defined the diabetes duration as the interval between the date of the first prescription of anti-diabetic medication and the date of health check-up.
The primary endpoint of this study was newly diagnosed pancreatic cancer, which was defined with the ICD-10 codes (C25, malignant neoplasm of pancreas) of the National Cancer Registry database. To avoid the enrollment of patients with pre-existing diseases, individuals diagnosed with pancreatic cancer during the preceding year were excluded.
Statistical analysis
Data are presented as the means ± SD, geometric mean (95% CI), or number (percentage). The baseline characteristics were compared, according to fasting glucose levels using t tests for continuous variables and the χ2 test for categorical variables. The incidence rates of pancreatic cancer, according to fasting glucose level, are expressed as the number of events per 100,000 persons per year. After an adjustment for possible confounders [age, sex, smoking, drinking, exercise, and body mass index (BMI)], the Cox proportional hazard models were used to estimate HRs and 95% CIs. Two-tailed P values ≤0.05 were considered statistically significant. All data were analyzed using SAS software version 9.3 (SAS Institute, Cary, NC).
Results
Among the source population of the NHIS database, a total of 25.4 million patients who had participated in a preventive health check-up between 2009 and 2013 were evaluated for pancreatic cancer incidence rates, according to fasting glucose level, and subjects who were younger than 30 years old (n = 2,489,999), had missing information (n = 96,202), or had a history of pancreatic cancer (n = 2074) were excluded (Fig. 1). During 116,995,600 person years of follow-up [median follow-up period was 5.45 years (interquartile range, 4.21 to 6.20)], 11,429 new cases of pancreatic cancer were identified (after excluding subjects whose follow-up period was <12 months).
Summary of participant selection.
A total of 22.8 million patients were evaluated and categorized into five groups according to their fasting glucose levels. The baseline characteristics of five group participants are summarized in Table 1. The 5-year cumulative incidence rates (per 100,000 with 95% CI) of the six groups for pancreatic cancer were as follows: (i) lower normal = 32 (95% CI, 31 to 33); (ii) high normal = 41 (40 to 43); (iii) prediabetes level 1 = 50 (47 to 52); (iv) prediabetes level 2 = 64 (60 to 68); (v) diabetes = 75 (69 to 82); and (vi) anti-diabetic medications = 121 (116 to 127). The risk of pancreatic cancer increased continuously as fasting glucose level elevated, and there was a significantly different risk according to fasting glucose level (log-rank test, P < 0.0001; Fig. 2).
Baseline Characteristics of Study Participants According to Fasting Glucose Levels (n = 22,801,607)
| Characteristics | Low Normal (<90 mg/dL) | High Normal (90–99 mg/dL) | Prediabetes Level 1 (100–109 mg/dL) | Prediabetes Level 2 (110–125 mg/dL) | Diabetes (≥126 mg/dL) | Anti-Diabetic Medications | Total | |
|---|---|---|---|---|---|---|---|---|
| n = 7,864,085 (34.5) | n = 7,227,941 (31.7) | n = 3,578,442 (15.7) | n = 1,814,242 (8.0) | n = 735,635 (3.2) | n = 1,581,352 (6.9) | n = 22,801,607 (100.0) | P Value | |
| Age, y | 46.8 ± 12.6 | 49.1 ± 12.7 | 50.8 ± 12.6 | 53.1 ± 12.6 | 53.4 ± 12.4 | 60.7 ± 10.8 | 49.9 ± 13.0 | <0.0001 |
| 30–49 | 4,874,226 (62.0) | 3,921,326 (54.3) | 1,721,861 (48.1) | 733,964 (40.5) | 290,942 (39.6) | 241,099 (15.2) | 11,783,418 (51.7) | <0.0001 |
| 50–59 | 1,641,786 (20.9) | 1,743,507 (24.1) | 955,126 (26.7) | 524,861 (28.9) | 221,221 (30.1) | 460,141 (29.1) | 5,546,642 (24.3) | <0.0001 |
| 60–69 | 844,135 (10.7) | 983,787 (13.6) | 565,209 (15.8) | 338,576 (18.7) | 135,761 (18.4) | 505,813 (32.0) | 3,373,281 (14.8) | <0.0001 |
| ≥70 | 503,938 (6.4) | 579,321 (8.0) | 336,246 (9.4) | 216,841 (11.9) | 87,711 (11.9) | 374,299 (23.7) | 2,098,356 (9.2) | <0.0001 |
| Sex, male | 3,541,676 (45.0) | 3,559,759 (49.3) | 2,023,355 (56.5) | 1,110,496 (61.2) | 495,609 (67.4) | 861,828 (54.5) | 11,592,723 (50.8) | <0.0001 |
| BMI, kg/m2 | 23.2 ± 3.1 | 23.8 ± 3.1 | 24.3 ± 3.2 | 24.8 ± 3.3 | 25.1 ± 3.5 | 25.1 ± 3.3 | 23.9 ± 3.2 | <0.0001 |
| Glucose, mg/dL | 82.3 ± 5.5 | 94.4 ± 2.9 | 103.8 ± 2.8 | 115.6 ± 4.5 | 154.8 ± 39.1 | 141.8 ± 49.0 | 98.6 ± 24.2 | <0.0001 |
| DM | — | — | — | — | 735,635 (100.0) | 1,581,352 (100.0) | 2,316,987 (10.3) | <0.0001 |
| SBP, mmHg | 120 ± 15 | 122 ± 15 | 125 ± 15 | 128 ± 15 | 130 ± 16 | 129 ± 16 | 122 ± 15 | <0.0001 |
| DBP, mmHg | 75 ± 10 | 76 ± 10 | 78 ± 10 | 79 ± 10 | 81 ± 11 | 78 ± 10 | 76 ± 10 | <0.0001 |
| HTN | 1,463,330 (18.6) | 1,755,900 (24.3) | 1,150,778 (32.2) | 742,790 (40.9) | 335,592 (45.6) | 1,034,517 (65.4) | 6,482,907 (28.4) | <0.0001 |
| TC, mg/dL | 192.1 ± 35.2 | 197.4 ± 35.7 | 202.0 ± 37.1 | 205.1 ± 38.9 | 208.1 ± 41.9 | 189.3 ± 41.6 | 196.7 ± 37.0 | <0.0001 |
| Dyslipidemia | 1,111,126 (14.1) | 1,315,948 (18.2) | 826,075 (23.1) | 505,981 (27.9) | 228,632 (31.1) | 799,427 (50.6) | 4,787,189 (21.0) | <0.0001 |
| Smoking | <0.0001 | |||||||
| None | 5,123,377 (65.2) | 4,550,748 (63.0) | 2,062,386 (57.6) | 979,511 (54.0) | 357,423 (48.6) | 953,954 (60.3) | 14,027,399 (61.5) | |
| Ex-smoker | 898,752 (11.4) | 994,961 (13.8) | 590,882 (16.5) | 328,278 (18.1) | 130,865 (17.8) | 287,849 (18.2) | 3,231,587 (14.2) | |
| Current smoker | 1,841,956 (23.4) | 1,682,232 (23.2) | 925,174 (25.9) | 506,453 (27.9) | 247,347 (33.6) | 339,549 (21.5) | 5,542,711 (24.3) | |
| Drinking | <0.0001 | |||||||
| None | 4,517,897 (57.5) | 3,992,472 (55.2) | 1,819,216 (50.8) | 879,785 (48.5) | 337,783 (45.9) | 1,035,777 (65.5) | 12,582,930 (55.2) | |
| Moderate | 2,899,005 (36.9) | 2,725,923 (37.7) | 1,415,415 (39.6) | 715,392 (39.4) | 293,479 (39.9) | 419,487 (26.5) | 8,468,701 (37.1) | |
| Heavy, ≥30 g/day | 447,183 (5.6) | 509,546 (7.1) | 343,811 (9.6) | 219,065 (12.1) | 104,373 (14.2) | 126,088 (8.0) | 1,750,066 (7.7) | |
| Regular exercise | 1,320,733 (16.8) | 1,274,215 (17.6) | 653,638 (18.3) | 340,824 (18.8) | 132,511 (18.0) | 344,410 (21.8) | 4,066,331 (17.8) | <0.0001 |
| Characteristics | Low Normal (<90 mg/dL) | High Normal (90–99 mg/dL) | Prediabetes Level 1 (100–109 mg/dL) | Prediabetes Level 2 (110–125 mg/dL) | Diabetes (≥126 mg/dL) | Anti-Diabetic Medications | Total | |
|---|---|---|---|---|---|---|---|---|
| n = 7,864,085 (34.5) | n = 7,227,941 (31.7) | n = 3,578,442 (15.7) | n = 1,814,242 (8.0) | n = 735,635 (3.2) | n = 1,581,352 (6.9) | n = 22,801,607 (100.0) | P Value | |
| Age, y | 46.8 ± 12.6 | 49.1 ± 12.7 | 50.8 ± 12.6 | 53.1 ± 12.6 | 53.4 ± 12.4 | 60.7 ± 10.8 | 49.9 ± 13.0 | <0.0001 |
| 30–49 | 4,874,226 (62.0) | 3,921,326 (54.3) | 1,721,861 (48.1) | 733,964 (40.5) | 290,942 (39.6) | 241,099 (15.2) | 11,783,418 (51.7) | <0.0001 |
| 50–59 | 1,641,786 (20.9) | 1,743,507 (24.1) | 955,126 (26.7) | 524,861 (28.9) | 221,221 (30.1) | 460,141 (29.1) | 5,546,642 (24.3) | <0.0001 |
| 60–69 | 844,135 (10.7) | 983,787 (13.6) | 565,209 (15.8) | 338,576 (18.7) | 135,761 (18.4) | 505,813 (32.0) | 3,373,281 (14.8) | <0.0001 |
| ≥70 | 503,938 (6.4) | 579,321 (8.0) | 336,246 (9.4) | 216,841 (11.9) | 87,711 (11.9) | 374,299 (23.7) | 2,098,356 (9.2) | <0.0001 |
| Sex, male | 3,541,676 (45.0) | 3,559,759 (49.3) | 2,023,355 (56.5) | 1,110,496 (61.2) | 495,609 (67.4) | 861,828 (54.5) | 11,592,723 (50.8) | <0.0001 |
| BMI, kg/m2 | 23.2 ± 3.1 | 23.8 ± 3.1 | 24.3 ± 3.2 | 24.8 ± 3.3 | 25.1 ± 3.5 | 25.1 ± 3.3 | 23.9 ± 3.2 | <0.0001 |
| Glucose, mg/dL | 82.3 ± 5.5 | 94.4 ± 2.9 | 103.8 ± 2.8 | 115.6 ± 4.5 | 154.8 ± 39.1 | 141.8 ± 49.0 | 98.6 ± 24.2 | <0.0001 |
| DM | — | — | — | — | 735,635 (100.0) | 1,581,352 (100.0) | 2,316,987 (10.3) | <0.0001 |
| SBP, mmHg | 120 ± 15 | 122 ± 15 | 125 ± 15 | 128 ± 15 | 130 ± 16 | 129 ± 16 | 122 ± 15 | <0.0001 |
| DBP, mmHg | 75 ± 10 | 76 ± 10 | 78 ± 10 | 79 ± 10 | 81 ± 11 | 78 ± 10 | 76 ± 10 | <0.0001 |
| HTN | 1,463,330 (18.6) | 1,755,900 (24.3) | 1,150,778 (32.2) | 742,790 (40.9) | 335,592 (45.6) | 1,034,517 (65.4) | 6,482,907 (28.4) | <0.0001 |
| TC, mg/dL | 192.1 ± 35.2 | 197.4 ± 35.7 | 202.0 ± 37.1 | 205.1 ± 38.9 | 208.1 ± 41.9 | 189.3 ± 41.6 | 196.7 ± 37.0 | <0.0001 |
| Dyslipidemia | 1,111,126 (14.1) | 1,315,948 (18.2) | 826,075 (23.1) | 505,981 (27.9) | 228,632 (31.1) | 799,427 (50.6) | 4,787,189 (21.0) | <0.0001 |
| Smoking | <0.0001 | |||||||
| None | 5,123,377 (65.2) | 4,550,748 (63.0) | 2,062,386 (57.6) | 979,511 (54.0) | 357,423 (48.6) | 953,954 (60.3) | 14,027,399 (61.5) | |
| Ex-smoker | 898,752 (11.4) | 994,961 (13.8) | 590,882 (16.5) | 328,278 (18.1) | 130,865 (17.8) | 287,849 (18.2) | 3,231,587 (14.2) | |
| Current smoker | 1,841,956 (23.4) | 1,682,232 (23.2) | 925,174 (25.9) | 506,453 (27.9) | 247,347 (33.6) | 339,549 (21.5) | 5,542,711 (24.3) | |
| Drinking | <0.0001 | |||||||
| None | 4,517,897 (57.5) | 3,992,472 (55.2) | 1,819,216 (50.8) | 879,785 (48.5) | 337,783 (45.9) | 1,035,777 (65.5) | 12,582,930 (55.2) | |
| Moderate | 2,899,005 (36.9) | 2,725,923 (37.7) | 1,415,415 (39.6) | 715,392 (39.4) | 293,479 (39.9) | 419,487 (26.5) | 8,468,701 (37.1) | |
| Heavy, ≥30 g/day | 447,183 (5.6) | 509,546 (7.1) | 343,811 (9.6) | 219,065 (12.1) | 104,373 (14.2) | 126,088 (8.0) | 1,750,066 (7.7) | |
| Regular exercise | 1,320,733 (16.8) | 1,274,215 (17.6) | 653,638 (18.3) | 340,824 (18.8) | 132,511 (18.0) | 344,410 (21.8) | 4,066,331 (17.8) | <0.0001 |
Values are presented as n (%) or means ± SD. For dyslipidemia, total cholesterol (TC) ≥ 240 mg/dL or on medication.
Abbreviations: DM, diabetes mellitus (defined by fasting glucose ≥126 mg/dL or on medication); HTN, hypertension [systolic blood pressure (SBP) ≥ 140 mmHg, diastolic blood pressure (DBP) ≥ 90 mmHg, or on medication].
Baseline Characteristics of Study Participants According to Fasting Glucose Levels (n = 22,801,607)
| Characteristics | Low Normal (<90 mg/dL) | High Normal (90–99 mg/dL) | Prediabetes Level 1 (100–109 mg/dL) | Prediabetes Level 2 (110–125 mg/dL) | Diabetes (≥126 mg/dL) | Anti-Diabetic Medications | Total | |
|---|---|---|---|---|---|---|---|---|
| n = 7,864,085 (34.5) | n = 7,227,941 (31.7) | n = 3,578,442 (15.7) | n = 1,814,242 (8.0) | n = 735,635 (3.2) | n = 1,581,352 (6.9) | n = 22,801,607 (100.0) | P Value | |
| Age, y | 46.8 ± 12.6 | 49.1 ± 12.7 | 50.8 ± 12.6 | 53.1 ± 12.6 | 53.4 ± 12.4 | 60.7 ± 10.8 | 49.9 ± 13.0 | <0.0001 |
| 30–49 | 4,874,226 (62.0) | 3,921,326 (54.3) | 1,721,861 (48.1) | 733,964 (40.5) | 290,942 (39.6) | 241,099 (15.2) | 11,783,418 (51.7) | <0.0001 |
| 50–59 | 1,641,786 (20.9) | 1,743,507 (24.1) | 955,126 (26.7) | 524,861 (28.9) | 221,221 (30.1) | 460,141 (29.1) | 5,546,642 (24.3) | <0.0001 |
| 60–69 | 844,135 (10.7) | 983,787 (13.6) | 565,209 (15.8) | 338,576 (18.7) | 135,761 (18.4) | 505,813 (32.0) | 3,373,281 (14.8) | <0.0001 |
| ≥70 | 503,938 (6.4) | 579,321 (8.0) | 336,246 (9.4) | 216,841 (11.9) | 87,711 (11.9) | 374,299 (23.7) | 2,098,356 (9.2) | <0.0001 |
| Sex, male | 3,541,676 (45.0) | 3,559,759 (49.3) | 2,023,355 (56.5) | 1,110,496 (61.2) | 495,609 (67.4) | 861,828 (54.5) | 11,592,723 (50.8) | <0.0001 |
| BMI, kg/m2 | 23.2 ± 3.1 | 23.8 ± 3.1 | 24.3 ± 3.2 | 24.8 ± 3.3 | 25.1 ± 3.5 | 25.1 ± 3.3 | 23.9 ± 3.2 | <0.0001 |
| Glucose, mg/dL | 82.3 ± 5.5 | 94.4 ± 2.9 | 103.8 ± 2.8 | 115.6 ± 4.5 | 154.8 ± 39.1 | 141.8 ± 49.0 | 98.6 ± 24.2 | <0.0001 |
| DM | — | — | — | — | 735,635 (100.0) | 1,581,352 (100.0) | 2,316,987 (10.3) | <0.0001 |
| SBP, mmHg | 120 ± 15 | 122 ± 15 | 125 ± 15 | 128 ± 15 | 130 ± 16 | 129 ± 16 | 122 ± 15 | <0.0001 |
| DBP, mmHg | 75 ± 10 | 76 ± 10 | 78 ± 10 | 79 ± 10 | 81 ± 11 | 78 ± 10 | 76 ± 10 | <0.0001 |
| HTN | 1,463,330 (18.6) | 1,755,900 (24.3) | 1,150,778 (32.2) | 742,790 (40.9) | 335,592 (45.6) | 1,034,517 (65.4) | 6,482,907 (28.4) | <0.0001 |
| TC, mg/dL | 192.1 ± 35.2 | 197.4 ± 35.7 | 202.0 ± 37.1 | 205.1 ± 38.9 | 208.1 ± 41.9 | 189.3 ± 41.6 | 196.7 ± 37.0 | <0.0001 |
| Dyslipidemia | 1,111,126 (14.1) | 1,315,948 (18.2) | 826,075 (23.1) | 505,981 (27.9) | 228,632 (31.1) | 799,427 (50.6) | 4,787,189 (21.0) | <0.0001 |
| Smoking | <0.0001 | |||||||
| None | 5,123,377 (65.2) | 4,550,748 (63.0) | 2,062,386 (57.6) | 979,511 (54.0) | 357,423 (48.6) | 953,954 (60.3) | 14,027,399 (61.5) | |
| Ex-smoker | 898,752 (11.4) | 994,961 (13.8) | 590,882 (16.5) | 328,278 (18.1) | 130,865 (17.8) | 287,849 (18.2) | 3,231,587 (14.2) | |
| Current smoker | 1,841,956 (23.4) | 1,682,232 (23.2) | 925,174 (25.9) | 506,453 (27.9) | 247,347 (33.6) | 339,549 (21.5) | 5,542,711 (24.3) | |
| Drinking | <0.0001 | |||||||
| None | 4,517,897 (57.5) | 3,992,472 (55.2) | 1,819,216 (50.8) | 879,785 (48.5) | 337,783 (45.9) | 1,035,777 (65.5) | 12,582,930 (55.2) | |
| Moderate | 2,899,005 (36.9) | 2,725,923 (37.7) | 1,415,415 (39.6) | 715,392 (39.4) | 293,479 (39.9) | 419,487 (26.5) | 8,468,701 (37.1) | |
| Heavy, ≥30 g/day | 447,183 (5.6) | 509,546 (7.1) | 343,811 (9.6) | 219,065 (12.1) | 104,373 (14.2) | 126,088 (8.0) | 1,750,066 (7.7) | |
| Regular exercise | 1,320,733 (16.8) | 1,274,215 (17.6) | 653,638 (18.3) | 340,824 (18.8) | 132,511 (18.0) | 344,410 (21.8) | 4,066,331 (17.8) | <0.0001 |
| Characteristics | Low Normal (<90 mg/dL) | High Normal (90–99 mg/dL) | Prediabetes Level 1 (100–109 mg/dL) | Prediabetes Level 2 (110–125 mg/dL) | Diabetes (≥126 mg/dL) | Anti-Diabetic Medications | Total | |
|---|---|---|---|---|---|---|---|---|
| n = 7,864,085 (34.5) | n = 7,227,941 (31.7) | n = 3,578,442 (15.7) | n = 1,814,242 (8.0) | n = 735,635 (3.2) | n = 1,581,352 (6.9) | n = 22,801,607 (100.0) | P Value | |
| Age, y | 46.8 ± 12.6 | 49.1 ± 12.7 | 50.8 ± 12.6 | 53.1 ± 12.6 | 53.4 ± 12.4 | 60.7 ± 10.8 | 49.9 ± 13.0 | <0.0001 |
| 30–49 | 4,874,226 (62.0) | 3,921,326 (54.3) | 1,721,861 (48.1) | 733,964 (40.5) | 290,942 (39.6) | 241,099 (15.2) | 11,783,418 (51.7) | <0.0001 |
| 50–59 | 1,641,786 (20.9) | 1,743,507 (24.1) | 955,126 (26.7) | 524,861 (28.9) | 221,221 (30.1) | 460,141 (29.1) | 5,546,642 (24.3) | <0.0001 |
| 60–69 | 844,135 (10.7) | 983,787 (13.6) | 565,209 (15.8) | 338,576 (18.7) | 135,761 (18.4) | 505,813 (32.0) | 3,373,281 (14.8) | <0.0001 |
| ≥70 | 503,938 (6.4) | 579,321 (8.0) | 336,246 (9.4) | 216,841 (11.9) | 87,711 (11.9) | 374,299 (23.7) | 2,098,356 (9.2) | <0.0001 |
| Sex, male | 3,541,676 (45.0) | 3,559,759 (49.3) | 2,023,355 (56.5) | 1,110,496 (61.2) | 495,609 (67.4) | 861,828 (54.5) | 11,592,723 (50.8) | <0.0001 |
| BMI, kg/m2 | 23.2 ± 3.1 | 23.8 ± 3.1 | 24.3 ± 3.2 | 24.8 ± 3.3 | 25.1 ± 3.5 | 25.1 ± 3.3 | 23.9 ± 3.2 | <0.0001 |
| Glucose, mg/dL | 82.3 ± 5.5 | 94.4 ± 2.9 | 103.8 ± 2.8 | 115.6 ± 4.5 | 154.8 ± 39.1 | 141.8 ± 49.0 | 98.6 ± 24.2 | <0.0001 |
| DM | — | — | — | — | 735,635 (100.0) | 1,581,352 (100.0) | 2,316,987 (10.3) | <0.0001 |
| SBP, mmHg | 120 ± 15 | 122 ± 15 | 125 ± 15 | 128 ± 15 | 130 ± 16 | 129 ± 16 | 122 ± 15 | <0.0001 |
| DBP, mmHg | 75 ± 10 | 76 ± 10 | 78 ± 10 | 79 ± 10 | 81 ± 11 | 78 ± 10 | 76 ± 10 | <0.0001 |
| HTN | 1,463,330 (18.6) | 1,755,900 (24.3) | 1,150,778 (32.2) | 742,790 (40.9) | 335,592 (45.6) | 1,034,517 (65.4) | 6,482,907 (28.4) | <0.0001 |
| TC, mg/dL | 192.1 ± 35.2 | 197.4 ± 35.7 | 202.0 ± 37.1 | 205.1 ± 38.9 | 208.1 ± 41.9 | 189.3 ± 41.6 | 196.7 ± 37.0 | <0.0001 |
| Dyslipidemia | 1,111,126 (14.1) | 1,315,948 (18.2) | 826,075 (23.1) | 505,981 (27.9) | 228,632 (31.1) | 799,427 (50.6) | 4,787,189 (21.0) | <0.0001 |
| Smoking | <0.0001 | |||||||
| None | 5,123,377 (65.2) | 4,550,748 (63.0) | 2,062,386 (57.6) | 979,511 (54.0) | 357,423 (48.6) | 953,954 (60.3) | 14,027,399 (61.5) | |
| Ex-smoker | 898,752 (11.4) | 994,961 (13.8) | 590,882 (16.5) | 328,278 (18.1) | 130,865 (17.8) | 287,849 (18.2) | 3,231,587 (14.2) | |
| Current smoker | 1,841,956 (23.4) | 1,682,232 (23.2) | 925,174 (25.9) | 506,453 (27.9) | 247,347 (33.6) | 339,549 (21.5) | 5,542,711 (24.3) | |
| Drinking | <0.0001 | |||||||
| None | 4,517,897 (57.5) | 3,992,472 (55.2) | 1,819,216 (50.8) | 879,785 (48.5) | 337,783 (45.9) | 1,035,777 (65.5) | 12,582,930 (55.2) | |
| Moderate | 2,899,005 (36.9) | 2,725,923 (37.7) | 1,415,415 (39.6) | 715,392 (39.4) | 293,479 (39.9) | 419,487 (26.5) | 8,468,701 (37.1) | |
| Heavy, ≥30 g/day | 447,183 (5.6) | 509,546 (7.1) | 343,811 (9.6) | 219,065 (12.1) | 104,373 (14.2) | 126,088 (8.0) | 1,750,066 (7.7) | |
| Regular exercise | 1,320,733 (16.8) | 1,274,215 (17.6) | 653,638 (18.3) | 340,824 (18.8) | 132,511 (18.0) | 344,410 (21.8) | 4,066,331 (17.8) | <0.0001 |
Values are presented as n (%) or means ± SD. For dyslipidemia, total cholesterol (TC) ≥ 240 mg/dL or on medication.
Abbreviations: DM, diabetes mellitus (defined by fasting glucose ≥126 mg/dL or on medication); HTN, hypertension [systolic blood pressure (SBP) ≥ 140 mmHg, diastolic blood pressure (DBP) ≥ 90 mmHg, or on medication].
Cumulative incidence curves of pancreatic cancer according to different fasting glucose levels.
In a multiadjusted analysis, the incidence of pancreatic cancer increased significantly with increasing fasting blood glucose levels compared with the low normal glucose level group (<90 mg/dL). In Model 1, which was adjusted for age and sex, the HRs for pancreatic cancer incidence, according to fasting blood glucose, were 1.17 (95% CI, 1.11 to 1.23) for the high normal group, 1.27 (1.20 to 1.35) for the prediabetes level 1 group, 1.51 (1.41 to 1.62) for the prediabetes level 2 group, 1.74 (1.59 to 1.90) for the diabetes group, and 2.15 (2.03 to 2.28) for the anti-diabetic medications group (P < 0.0001; Table 2). Moreover, similar significant associations were also observed between fasting glucose level and pancreatic cancer incidence in Model 2 after adjusting for age, sex, smoking, drinking, exercise, BMI, and diabetes duration.
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Fasting Glucose Level
| Glucose Level, mg/dL | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1a HR (95% CI) | Model 2a HR (95% CI) |
|---|---|---|---|---|---|---|
| Low normal (<90) | 7,864,085 | 2652 | 40,413,653 | 6.56 | Reference | Reference |
| High normal (90–99) | 7,227,941 | 3122 | 37,054,791 | 8.43 | 1.17 (1.11–1.23) | 1.14 (1.08–1.20) |
| Prediabetes level 1 (100–109) | 3,578,442 | 1875 | 18,413,285 | 10.18 | 1.27 (1.20–1.35) | 1.22 (1.15–1.29) |
| Prediabetes level 2 (110–125) | 1,814,242 | 1249 | 9,372,162 | 13.33 | 1.51 (1.41–1.62) | 1.42 (1.33–1.52) |
| Diabetes (≥126) | 735,635 | 587 | 3,729,125 | 15.74 | 1.74 (1.59–1.90) | 1.62 (1.48–1.77) |
| Anti-diabetic medications | 1,581,352 | 1944 | 8,012,586 | 24.26 | 2.15 (2.03–2.28) | 1.91 (1.75–2.08) |
| Glucose Level, mg/dL | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1a HR (95% CI) | Model 2a HR (95% CI) |
|---|---|---|---|---|---|---|
| Low normal (<90) | 7,864,085 | 2652 | 40,413,653 | 6.56 | Reference | Reference |
| High normal (90–99) | 7,227,941 | 3122 | 37,054,791 | 8.43 | 1.17 (1.11–1.23) | 1.14 (1.08–1.20) |
| Prediabetes level 1 (100–109) | 3,578,442 | 1875 | 18,413,285 | 10.18 | 1.27 (1.20–1.35) | 1.22 (1.15–1.29) |
| Prediabetes level 2 (110–125) | 1,814,242 | 1249 | 9,372,162 | 13.33 | 1.51 (1.41–1.62) | 1.42 (1.33–1.52) |
| Diabetes (≥126) | 735,635 | 587 | 3,729,125 | 15.74 | 1.74 (1.59–1.90) | 1.62 (1.48–1.77) |
| Anti-diabetic medications | 1,581,352 | 1944 | 8,012,586 | 24.26 | 2.15 (2.03–2.28) | 1.91 (1.75–2.08) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, BMI, and diabetes duration.
Wald test, P < 0.0001.
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Fasting Glucose Level
| Glucose Level, mg/dL | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1a HR (95% CI) | Model 2a HR (95% CI) |
|---|---|---|---|---|---|---|
| Low normal (<90) | 7,864,085 | 2652 | 40,413,653 | 6.56 | Reference | Reference |
| High normal (90–99) | 7,227,941 | 3122 | 37,054,791 | 8.43 | 1.17 (1.11–1.23) | 1.14 (1.08–1.20) |
| Prediabetes level 1 (100–109) | 3,578,442 | 1875 | 18,413,285 | 10.18 | 1.27 (1.20–1.35) | 1.22 (1.15–1.29) |
| Prediabetes level 2 (110–125) | 1,814,242 | 1249 | 9,372,162 | 13.33 | 1.51 (1.41–1.62) | 1.42 (1.33–1.52) |
| Diabetes (≥126) | 735,635 | 587 | 3,729,125 | 15.74 | 1.74 (1.59–1.90) | 1.62 (1.48–1.77) |
| Anti-diabetic medications | 1,581,352 | 1944 | 8,012,586 | 24.26 | 2.15 (2.03–2.28) | 1.91 (1.75–2.08) |
| Glucose Level, mg/dL | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1a HR (95% CI) | Model 2a HR (95% CI) |
|---|---|---|---|---|---|---|
| Low normal (<90) | 7,864,085 | 2652 | 40,413,653 | 6.56 | Reference | Reference |
| High normal (90–99) | 7,227,941 | 3122 | 37,054,791 | 8.43 | 1.17 (1.11–1.23) | 1.14 (1.08–1.20) |
| Prediabetes level 1 (100–109) | 3,578,442 | 1875 | 18,413,285 | 10.18 | 1.27 (1.20–1.35) | 1.22 (1.15–1.29) |
| Prediabetes level 2 (110–125) | 1,814,242 | 1249 | 9,372,162 | 13.33 | 1.51 (1.41–1.62) | 1.42 (1.33–1.52) |
| Diabetes (≥126) | 735,635 | 587 | 3,729,125 | 15.74 | 1.74 (1.59–1.90) | 1.62 (1.48–1.77) |
| Anti-diabetic medications | 1,581,352 | 1944 | 8,012,586 | 24.26 | 2.15 (2.03–2.28) | 1.91 (1.75–2.08) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, BMI, and diabetes duration.
Wald test, P < 0.0001.
We further analyzed to investigate the relationship between the diabetes duration and pancreatic cancer and whether using insulin could show reduced the risk of pancreatic cancer in diabetic populations. The pancreatic cancer incidence was separated according to diabetes duration (Table 3): the incidence was high in the population <1 year duration of diabetes, and then, the incidence rate was consistently increased as the diabetes duration increased. There was a J-shaped association between the duration of diabetes and the pancreatic cancer incidence. This finding suggests that among the populations who had diabetes or were treated with anti-diabetic medications, there was a correlation between the pancreatic cancer incidence and the diabetes duration (>1 year), although some patients might have hyperglycemia as a consequence from pancreatic cancer (in the <1-year group). Among the populations who were treated with anti-diabetic medications, there was no significant difference in the incidence of pancreatic cancer between the patients who received insulin and those who did not (Table 4).
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Diabetes Duration
| Diabetes Duration | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No medication | 735,635 | 587 | 3,729,125 | 15.7 | Reference | Reference |
| <1 y | 274,616 | 387 | 1,406,064 | 27.5 | 1.60 (1.41–1.82) | 1.60 (1.40–1.82) |
| <3 y | 242,938 | 224 | 1,256,152 | 17.8 | 0.99 (0.85–1.16) | 0.99 (0.85–1.15) |
| <5 y | 243,557 | 299 | 1,273,118 | 23.5 | 1.24 (1.08–1.43) | 1.24 (1.08–1.43) |
| ≥5 y | 820,241 | 1034 | 4,077,252 | 25.4 | 1.25 (1.13–1.39) | 1.26 (1.14–1.40) |
| Diabetes Duration | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No medication | 735,635 | 587 | 3,729,125 | 15.7 | Reference | Reference |
| <1 y | 274,616 | 387 | 1,406,064 | 27.5 | 1.60 (1.41–1.82) | 1.60 (1.40–1.82) |
| <3 y | 242,938 | 224 | 1,256,152 | 17.8 | 0.99 (0.85–1.16) | 0.99 (0.85–1.15) |
| <5 y | 243,557 | 299 | 1,273,118 | 23.5 | 1.24 (1.08–1.43) | 1.24 (1.08–1.43) |
| ≥5 y | 820,241 | 1034 | 4,077,252 | 25.4 | 1.25 (1.13–1.39) | 1.26 (1.14–1.40) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, and BMI.
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Diabetes Duration
| Diabetes Duration | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No medication | 735,635 | 587 | 3,729,125 | 15.7 | Reference | Reference |
| <1 y | 274,616 | 387 | 1,406,064 | 27.5 | 1.60 (1.41–1.82) | 1.60 (1.40–1.82) |
| <3 y | 242,938 | 224 | 1,256,152 | 17.8 | 0.99 (0.85–1.16) | 0.99 (0.85–1.15) |
| <5 y | 243,557 | 299 | 1,273,118 | 23.5 | 1.24 (1.08–1.43) | 1.24 (1.08–1.43) |
| ≥5 y | 820,241 | 1034 | 4,077,252 | 25.4 | 1.25 (1.13–1.39) | 1.26 (1.14–1.40) |
| Diabetes Duration | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No medication | 735,635 | 587 | 3,729,125 | 15.7 | Reference | Reference |
| <1 y | 274,616 | 387 | 1,406,064 | 27.5 | 1.60 (1.41–1.82) | 1.60 (1.40–1.82) |
| <3 y | 242,938 | 224 | 1,256,152 | 17.8 | 0.99 (0.85–1.16) | 0.99 (0.85–1.15) |
| <5 y | 243,557 | 299 | 1,273,118 | 23.5 | 1.24 (1.08–1.43) | 1.24 (1.08–1.43) |
| ≥5 y | 820,241 | 1034 | 4,077,252 | 25.4 | 1.25 (1.13–1.39) | 1.26 (1.14–1.40) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, and BMI.
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Insulin Treatment
| Insulin | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No insulin | 1,385,643 | 1702 | 7,066,775 | 24.1 | Reference | Reference |
| Insulin treatment | 195,709 | 242 | 945,810 | 25.6 | 1.05 (0.91–1.20) | 1.05 (0.92–1.21) |
| Insulin | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No insulin | 1,385,643 | 1702 | 7,066,775 | 24.1 | Reference | Reference |
| Insulin treatment | 195,709 | 242 | 945,810 | 25.6 | 1.05 (0.91–1.20) | 1.05 (0.92–1.21) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, and BMI.
Adjusted HR With 95% CI for Pancreatic Cancer Incidence According to Insulin Treatment
| Insulin | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No insulin | 1,385,643 | 1702 | 7,066,775 | 24.1 | Reference | Reference |
| Insulin treatment | 195,709 | 242 | 945,810 | 25.6 | 1.05 (0.91–1.20) | 1.05 (0.92–1.21) |
| Insulin | Populations, n | Events, n | Duration, Person Year | Incidence Rate, Per 100,000 | Model 1 HR (95% CI) | Model 2 HR (95% CI) |
|---|---|---|---|---|---|---|
| No insulin | 1,385,643 | 1702 | 7,066,775 | 24.1 | Reference | Reference |
| Insulin treatment | 195,709 | 242 | 945,810 | 25.6 | 1.05 (0.91–1.20) | 1.05 (0.92–1.21) |
Model 1, adjusted for age and sex; Model 2, adjusted for age, sex, smoking, drinking, exercise, and BMI.
Discussion
We evaluated pancreatic cancer incidence in Korea, according to fasting glucose levels, using a national cohort database. As the fasting glucose level elevated, the cumulative incidence rate of pancreatic cancer significantly increased, not only in diabetic populations but also in those with prediabetes or normal range of fasting blood glucose levels.
Diabetes is one of the established risk factors for pancreatic cancer (6). In addition to diabetes, fasting or postprandial glucose level has been evaluated for association with pancreatic cancer; however, most studies have focused on the mortality of pancreatic cancer (6–8). Whether hyperglycemia could increase the risk of pancreatic cancer remains unclear, as previous studies have showed inconsistent results (7, 9). These findings could be associated with a very small amount of pancreatic cancer incidence in the prediabetic range of glucose levels. In addition, as the hyperglycemia was first noted 36 to 30 months before diagnosis, some investigators have suggested that hyperglycemia might be an effect of pancreatic cancer as a paraneoplastic syndrome (5, 10).
In this study, we found a linear relationship between fasting glucose levels and pancreatic cancer incidence in both prediabetes, including normal glucose levels, and diabetes, even after an adjustment for well-known risk factors, including age, sex, smoking, drinking, exercise, and BMI. A recent prospective Chinese cohort study (n = 512,000) showed that each 1 mmol/L (=18 mg/dL) increase in blood glucose level was associated with a 15% increase in risk of pancreatic cancer incidence (11). In a Japanese cohort study (n = 46,387), the fasting blood glucose level also showed a dose-response relationship with pancreatic cancer mortality (12). Therefore, the risk of pancreatic cancer can increase progressively with worsening fasting glucose levels.
As the incidences of hyperglycemia and diabetes are rising globally with an aging population and increasing age-specific prevalence, diabetic patients in Korea have increased from 6% to 10% during the last 10 years (13, 14). As two previous randomized trials have shown that a lifestyle intervention could decrease fasting glucose levels and reduce the risk of diabetes (15, 16), early detection of hyperglycemia in preventive health check-ups could offer a critical opportunity for lowering the risk of pancreatic cancer. Therefore, efforts made toward early detection of hyperglycemia and lifestyle modification to improve glucose profile may provide a practical strategy to reduce the increasing risk of pancreatic cancer.
The major strengths of our study are its prospective nature with a large national population, specific fasting glucose category, and adjustment for well-known confounding risk factors for pancreatic cancer. However, there were several limitations as follows: (i) the duration of diabetes could be underestimated, as there were no available data before the first health check-up year as a result of nonprospective design; (ii) we could not evaluate the correlation between the hemoglobin (Hb)A1c levels of diabetic populations and pancreatic cancer incidence, as the NHIS database did not collect the HbA1c levels; therefore, further study should be warranted as to whether the diabetic patients who are more intensively treated have a lower incidence of pancreatic cancer using a comparison of glucose levels and HbA1c levels; and (iii) we do not evaluate the mortality of pancreatic cancer, as the NHIS database did not provide exact information on whether a patient’s death was associated with prediagnosed cancer.
In conclusion, the cumulative incidence rate of pancreatic cancer significantly increased with elevating fasting glucose level in both diabetic and prediabetic populations, including those with a normal range of fasting blood glucose levels.
Acknowledgments
Financial Support: This work was supported by the Medical Research Funds from Kangbuk Samsung Hospital (Seoul, Korea).
Author Contributions: Study concept and design were done by D.-H.K., K.-D.H., and C.-Y.P. Analysis and interpretation of data were completed by K.-D.H., D.-H.K., and C.-Y.P. Drafting of the manuscript was done by D.-H.K. and C.-Y.P. Statistical analysis was performed by K.-D.H. Funding was obtained by C.-Y.P.
Additional Information
Disclosure Summary: The authors have nothing to disclose.
Data Availability:
Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.
Abbreviations:
- BMI
body mass index
- HR
hazard ratio
- ICD-10
International Classification of Diseases, 10th Revision
- NHIS
National Health Insurance Service
