Abstract
Background
The associations of laboratory-based ABO phenotypes with cancer risks and mortality have not been systematically determined.
Methods
The study subjects were 339,432 healthy individuals with laboratory-based blood types from a Taiwan cohort.
Results
Compared to blood type O, blood type A was significantly associated with an elevated risk of stomach cancer incidence (Hazard Ratio [HR], 1.38 [95% CI, 1.11–1.72]) and mortality (HR, 1.38 [95% CI, 1.02–1.86]) compared with blood type O, after adjusting for age, sex, education, smoking, alcohol drinking, physical activity, and body mass index. Non-O blood types were associated with an elevated risk of pancreatic cancer, with blood type B reaching statistical significance for incidence (HR, 1.59 [95% CI, 1.02–2.48]) and mortality (HR, 1.63 [95% CI, 1.02–2.60]). In contrast, kidney cancer risk was inversely associated with blood type AB (HR, 0.41 [95% CI, 0.18–0.93]) compared to type O.
Conclusion
Cancer risks vary in people with different ABO blood types, with elevated risks of stomach cancer associated with blood type A and pancreatic cancer associated with non-O blood types (A, B, and AB).
Keywords: ABO blood type, Cancer risk, Cancer mortality, Cohort
1. Introduction
ABO blood type has been extensively studied in relation to human diseases and blood transfusion complications since its discovery by Karl Landsteiner in 1900 [1,2]. An association between blood type and cancer was initially suggested through the observation that stomach cancer patients were more likely to have blood type A than control subjects [3]. A variety of other cancers have since been linked to ABO blood group [4], including cancers of the oral cavity [5], larynx [6], esophagus [7], lung [8], pancreas [9], breast [10], colorectum [11], bladder [12], prostate [13], and skin [14], but the associations generally have been inconsistent. For instance, some studies showed that blood type was not associated with pancreatic cancer [15,16]. Most previous studies were based on self-reporting of blood type which is prone to recall and information biases. For example, in a study of health professionals, about 10% of the participants incorrectly reported their blood types [9]. Additionally, the relative small size of most previous studies has limited the opportunity to comprehensively evaluate the link between ABO blood type and cancer risks, particularly for rare cancers. The majority of studies on blood type and cancer risks were conducted in the Western world [4], and published evidence in other populations is limited. ABO blood type distribution has been shown to vary substantially among populations [17,18]. Phenotypic frequencies of ABO type in two large series of hospital patients and blood donors showed a higher percentage of type A among Caucasians (38%) than Chinese (29%), but a lower percentage of type B among Caucasians (14%) than Chinese (27%) [19,20].
To address the gaps in findings from existing studies, we conducted a large cohort study with laboratory-based ABO blood type to investigate the association of blood type with cancer risk and mortality among Chinese in Taiwan.
2. Methods
Details of the study methods were reported previously [21,22]. Briefly, the cohort consists of healthy people at least 20 years of age who were provided health management by a private firm (MJ Health Management Institution, Taipei City, Taiwan) beginning in 1994. All cohort members took part in a standard medical screening program and donated an overnight fasting blood sample at baseline, and were followed until December 31, 2007. All participants underwent a standard panel of clinical tests, including hemogram, blood sugar tests, liver function tests, renal function tests, blood lipid tests, thyroid function tests, blood grouping, the presence of HBV surface antigen in blood (HBsAg), and the presence of HCV antibody in blood [23]. Individuals were defined as having diabetes if they had a history of diabetes or if they had positive diabetes screening results (fasting blood glucose concentration ≥7 mmol/L). This is to provide a complete picture of diabetes diagnosis as diabetic patients whose disease is under control may decide not to undergo screening test or the test results may be negative. Individuals with either type I or type II diabetes were included. The ABO blood typing (A, B, O and AB) was conducted using the standard forward antigen typing with antisera reagents (anti-A, anti-B) made from monoclonal antibodies. Every participant completed a self-administered questionnaire of their medical history and lifestyle information. All participants were encouraged to visit on a yearly basis; the same questionnaires were filled out on every visit, but results from the initial visit only were used for the current analysis. All study participants signed an informed consent. The study was approved by the Institutional Review Boards (IRB) of MJ Health Management Institution and National Health Research Institutes in Taiwan and MD Anderson Cancer Center.
Of 440,630 individuals enrolled in the MJ Health Management Institution between January 1, 1994 and December 31, 2007, 339,432 were included in the current analysis with a total of 2,969,357 person-years of follow-up. The participants were followed for an average of 8.75 years (s.d. 3.15) with a range of 1.0–12.9 years. Excluded from the current analysis were subjects who reported a history of any cancer at baseline (n = 3714) or were followed up for less than 1 year (n = 27,035) or death in less than 1 year (n = 447). Also excluded were subjects with no information on ABO blood type (n = 70,002).
The principal outcome variables were cancer incidence and cause-specific mortality, including the most common cancers occurring in Taiwan. The procedures to ascertain these outcomes have been described in detail previously [24]. Incident cancer and vital status were ascertained from the Taiwan National Cancer Registry and National Death File, respectively, via record linkage using the unique national identification numbers assigned to each Taiwan resident. Cancer cases were classified according to the International Classification of Diseases (ICD-9) and were based on histological discharge forms and oncology reports [22].
The Cox proportional hazards regression analysis was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for cancer incidence and mortality in relation to ABO blood type. The time of study entry was the date of enrollment, and the time of exit was the date of first cancer diagnosis, death, or the end of follow-up (December 31, 2007), whichever came first. Participants who died of any causes other than cancer were censored on the date of death. The proportional hazards assumption was evaluated by inspection of plots of log (−log[S]) against time [25], in which S was the estimated survival function. Baseline potential confounders were adjusted, including sex, age, education, physical activity, smoking, alcohol drinking, and body mass index (BMI). The age- and sex-adjusted model included 339,432 individuals and the fully adjusted model included 297,187 individuals after excluding those with missing values for covariates. In addition, HBV infection was included as a covariate for estimating risk of liver cancer incidence and mortality. The potential confounders were categorized as in Table 1. We also considered whether the association of ABO blood type with cancer varied with age, sex, or BMI group from the significance of interaction terms. All analyses were conducted using the STATA statistical software program (version 10.1; Stata Corp., College Station, TX, USA).
Table 1.
Characteristics of the study participants by ABO blood type (column %).
| Characteristic | O | A | B | AB | Total |
|---|---|---|---|---|---|
| Sex | |||||
| Male | 67,190 (46.2) | 42,277 (46.5) | 38,711 (46.9) | 9358 (46.1) | 157,536 (46.4) |
| Female | 78,360 (53.8) | 48,695 (53.5) | 43,920 (53.1) | 10,921 (53.9) | 181,896 (53.6) |
| Age (years) | |||||
| <65 | 134,506 (92.4) | 83,652 (92.0) | 75,859 (91.8) | 18,636 (91.9) | 312,653 (92.1) |
| ≥65 | 11,044 (7.6) | 7320 (8.0) | 6772 (8.2) | 1643 (8.1) | 26,779 (7.9) |
| Education | |||||
| Middle school or lower | 44,421 (30.5) | 27,183 (29.9) | 24,572 (29.7) | 5908 (29.1) | 102,084 (30.1) |
| High school | 38,389 (26.4) | 23,996 (26.4) | 21,723 (26.3) | 5388 (26.6) | 89,496 (26.4) |
| Junior college | 28,565 (19.6) | 18,130 (19.9) | 16,469 (19.9) | 4096 (20.2) | 67,260 (19.8) |
| College or higher | 30,218 (20.8) | 19,251 (21.2) | 17,562 (21.2) | 4360 (21.5) | 71,391 (21.0) |
| Unknown | 3957 (2.7) | 2412 (2.6) | 2305 (2.8) | 527 (2.6) | 9201 (2.7) |
| BMI (kg/m2) | |||||
| ≤18.5 | 11,937 (8.2) | 7470 (8.2) | 6868 (8.3) | 1694 (8.4) | 27,969 (8.2) |
| <25 | 94,103 (64.7) | 58,636 (64.5) | 53,158 (64.3) | 13,111 (64.7) | 219,008 (64.6) |
| <30 | 34,019 (23.4) | 21,282 (23.4) | 19,415 (23.5) | 4665 (23.0) | 79,381 (23.4) |
| ≥30 | 5451 (3.7) | 3563 (3.9) | 3171 (3.9) | 800 (3.9) | 12,985 (3.8) |
| Unknown | 40 (0.0) | 21 (0.0) | 19 (0.0) | 9 (0.0) | 89 (0.0) |
| Smoking status | |||||
| Never-smoker | 95,653 (65.7) | 59,367 (65.2) | 53,779 (65.1) | 13,254 (65.4) | 222,053 (65.4) |
| Ex-smoker | 8445 (5.8) | 5534 (6.1) | 5011 (6.1) | 1197 (5.9) | 20,187 (5.9) |
| Smoker | 32,297 (22.2) | 20,366 (22.4) | 18,602 (22.5) | 4547 (22.4) | 75,812 (22.4) |
| Unknown | 9155 (6.3) | 5705 (6.3) | 5239 (6.3) | 1281 (6.3) | 21,380 (6.3) |
| Drinking status | |||||
| Never-drinker | 103,556 (71.1) | 64,487 (70.9) | 58,455 (70.7) | 14,358 (70.8) | 240,856 (71.0) |
| Ex-drinker | 4341 (3.0) | 2782 (3.0) | 2421 (2.9) | 594 (2.9) | 10,138 (3.0) |
| Regular drinker | 26,914 (18.5) | 17,095 (18.8) | 15,670 (19.0) | 3863 (19.1) | 63,542 (18.7) |
| Unknown | 10,739 (7.4) | 6608 (7.3) | 6085 (7.4) | 1464 (7.2) | 24,896 (7.3) |
| Physical activity | |||||
| Inactive | 74,051 (50.9) | 46,700 (51.3) | 42,606 (51.6) | 10,513 (51.8) | 173,870 (51.2) |
| Low | 42,853 (29.4) | 26,566 (29.2) | 23,933 (29.0) | 5886 (29.0) | 99,238 (29.3) |
| Middle | 16,324 (11.2) | 10,008 (11.0) | 9021 (10.9) | 2124 (10.5) | 37,477 (11.0) |
| High | 4306 (3.0) | 2740 (3.0) | 2492 (3.0) | 631 (3.1) | 10,169 (3.0) |
| Unknown | 8016 (5.5) | 4958 (5.5) | 4579 (5.5) | 1125 (5.6) | 18,678 (5.5) |
| Diabetes | |||||
| No | 141,449 (97.2) | 88,198 (97.0) | 80,200 (97.1) | 19,663 (97.0) | 329,510 (97.1) |
| Yes | 4100 (2.8) | 2774 (3.1) | 2431 (2.9) | 616 (3.0) | 9921 (2.9) |
| Unknown | 1 (0.00) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 1 (0.0) |
| Family history cancer | |||||
| No | 109,117 (75.0) | 67,523 (74.2) | 61,722 (74.7) | 15,230 (75.1) | 253,592 (74.7) |
| Yes | 36,433 (25.0) | 23,449 (25.8) | 20,909 (25.3) | 5049 (24.9) | 85,840 (25.3) |
| Unknown | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| HBV | |||||
| No | 103,242 (70.9) | 64,192 (70.6) | 58,761 (71.1) | 14,592 (72.0) | 240,787 (70.9) |
| Yes | 24,906 (17.1) | 16,022 (17.6) | 14,007 (17.0) | 3337 (16.4) | 58,272 (17.2) |
| Unknown | 17,402 (12.0) | 10,758 (11.8) | 9863 (11.9) | 2350 (11.6) | 40,373 (11.9) |
3. Results
Of the 339,432 participants, 42.9% had blood type O, and 26.8, 24.3, and 6.0% had blood types B, A, and AB, respectively. The mean and median ages at baseline were 41 and 39 years, with little variations in subgroups defined by blood type (Table 1). Only 7.9% of the cohort participants were age 65 years or older and 53.6% was female. About 42% of the cohort had attended at least some college. The majority were never smokers (69.8%) and never alcohol drinkers (76.6%). Overweight (BMI 25–<30) and obesity (BMI ≥ 30) were reported by 23.4 and 3.8% of the cohort, respectively. Over 85% were inactive or had low physical activity level. There were no meaningful differences in the distributions of the demographic variables and exposures by ABO blood groups.
During an average of 8.75 years of follow-up, 9164 participants developed cancer and 3803 participants had died of the disease. Compared to type O, type AB was associated with a slight decrease in risk of developing cancer overall (HR = 0.90, 95% CI 0.81–1.00; P = 0.049) after adjusting for age, sex, educational level, smoking, alcohol drinking, physical activity, and BMI (Table 2). When specific cancers were examined, a significant inverse association with type AB was observed only for kidney cancer (HR = 0.41, 95% CI 0.18–0.93). An inverse association relative to blood type O was also observed for other blood types, with kidney cancer risk for those with non-O blood type reduced 27% (HR = 0.73, 93% CI 0.56–0.96). In contrast, stomach cancer risk was significantly elevated for those with blood type A (HR = 1.38, 95% CI 1.11–1.72) compared to those with blood type O. No association was observed for blood type B or AB. For pancreas cancer, a significant excess risk associated with non-O blood type overall was observed (HR = 1.50, 95% CI 1.04–2.19), with an elevated risk found across all non-O blood types (HR for type A: 1.45; type B: 1.59; and type AB: 1.37).
Table 2.
HRs for over-all and cause-specific cancer incidence by ABO blood type.
| Cancer site | Blood type | N | Incidence rate (per 100,000 person-years) | HRa (95% CI) | HRb (95% CI) |
|---|---|---|---|---|---|
| Overall | O | 3921 | 339.86 | 1.00 (Reference) | 1.00 (Reference) |
| A | 2560 | 353.28 | 1.02 (0.97–1.07) | 1.04 (0.98–1.10) | |
| B | 2171 | 329.56 | 0.94 (0.89–0.99) | 0.97 (0.91–1.02) | |
| AB | 512 | 316.37 | 0.92 (0.84–1.01) | 0.90 (0.81–1.00) | |
| Non-O | 5243 | 339.30 | 0.97 (0.93–1.02) | 0.99 (0.95–1.04) | |
| Oral | O | 262 | 22.71 | 1.00 (Reference) | 1.00 (Reference) |
| A | 180 | 24.84 | 1.08 (0.89–1.30) | 1.06 (0.86–1.30) | |
| B | 147 | 22.31 | 0.96 (0.78–1.17) | 0.97 (0.77–1.21) | |
| AB | 38 | 23.48 | 1.03 (0.73–1.44) | 0.93 (0.62–1.37) | |
| Non-O | 365 | 23.62 | 1.02 (0.87–1.20) | 1.01 (0.84–1.20) | |
| Esophagus | O | 52 | 4.51 | 1.00 (Reference) | 1.00 (Reference) |
| A | 33 | 4.55 | 0.95 (0.61–1.48) | 1.08 (0.67–1.74) | |
| B | 27 | 4.10 | 0.85 (0.53–1.36) | 0.96 (0.58–1.60) | |
| AB | 8 | 4.94 | 1.08 (0.51–2.28) | 1.01 (0.43–2.39) | |
| Non-O | 68 | 4.40 | 0.92 (0.64–1.33) | 1.02 (0.69–1.53) | |
| Lung | O | 429 | 37.18 | 1.00 (Reference) | 1.00 (Reference) |
| A | 294 | 40.57 | 1.05 (0.91–1.22) | 1.01 (0.85–1.19) | |
| B | 281 | 42.66 | 1.08 (0.93–1.26) | 1.08 (0.92–1.28) | |
| AB | 61 | 37.69 | 0.97 (0.74–1.28) | 1.03 (0.77–1.38) | |
| Non-O | 636 | 41.16 | 1.06 (0.94–1.20) | 1.04 (0.91–1.19) | |
| Stomach | O | 215 | 18.64 | 1.00 (Reference) | 1.00 (Reference) |
| A | 180 | 24.84 | 1.28 (1.05–1.56) | 1.38 (1.11–1.72) | |
| B | 104 | 15.79 | 0.81 (0.64–1.03) | 0.82 (0.63–1.07) | |
| AB | 25 | 15.45 | 0.81 (0.54–1.23) | 0.92 (0.59–1.44) | |
| Non-O | 309 | 20.00 | 1.03 (0.87–1.23) | 1.09 (0.90–1.33) | |
| Liverc | O | 625 | 54.17 | 1.00 (Reference) | 1.00 (Reference) |
| A | 361 | 49.82 | 0.89 (0.79–1.02) | 0.92 (0.80–1.07) | |
| B | 339 | 51.46 | 0.92 (0.81–1.05) | 0.96 (0.83–1.12) | |
| AB | 69 | 42.64 | 0.77 (0.60–0.99) | 0.81 (0.61–1.07) | |
| Non-O | 769 | 49.77 | 0.89 (0.80–0.99) | 0.93 (0.82–1.04) | |
| Pancreas | O | 51 | 4.42 | 1.00 (Reference) | 1.00 (Reference) |
| A | 50 | 6.90 | 1.50 (1.02–2.22) | 1.45 (0.94–2.26) | |
| B | 45 | 6.83 | 1.47 (0.99–2.20) | 1.59 (1.02–2.48) | |
| AB | 10 | 6.18 | 1.36 (0.69–2.68) | 1.37 (0.64–2.93) | |
| Non-O | 105 | 6.80 | 1.48 (1.06–2.06) | 1.50 (1.04–2.19) | |
| Rectum | O | 205 | 17.77 | 1.00 (Reference) | 1.00 (Reference) |
| A | 159 | 21.94 | 1.20 (0.97–1.47) | 1.19 (0.94–1.51) | |
| B | 126 | 19.13 | 1.04 (0.83–1.30) | 1.09 (0.85–1.40) | |
| AB | 28 | 17.30 | 0.96 (0.65–1.42) | 1.00 (0.65–1.55) | |
| Non-O | 313 | 20.26 | 1.10 (0.93–1.32) | 1.13 (0.93–1.38) | |
| Colon | O | 328 | 28.43 | 1.00 (Reference) | 1.00 (Reference) |
| A | 226 | 31.19 | 1.06 (0.90–1.26) | 1.12 (0.93–1.36) | |
| B | 172 | 26.11 | 0.89 (0.74–1.07) | 0.96 (0.78–1.17) | |
| AB | 46 | 28.42 | 0.98 (0.72–1.34) | 1.01 (0.71–1.42) | |
| Non-O | 444 | 28.73 | 0.98 (0.85–1.13) | 1.04 (0.89–1.22) | |
| Kidney | O | 125 | 10.83 | 1.00 (Reference) | 1.00 (Reference) |
| A | 64 | 8.83 | 0.79 (0.58–1.07) | 0.82 (0.58–1.14) | |
| B | 58 | 8.80 | 0.78 (0.57–1.06) | 0.72 (0.50–1.03) | |
| AB | 12 | 7.42 | 0.67 (0.37–1.21) | 0.41 (0.18–0.93) | |
| Non-O | 134 | 8.67 | 0.77 (0.61–0.99) | 0.73 (0.56–0.96) | |
| Bladder | O | 85 | 7.37 | 1.00 (Reference) | 1.00 (Reference) |
| A | 68 | 9.38 | 1.21 (0.88–1.67) | 1.11 (0.75–1.63) | |
| B | 66 | 10.02 | 1.28 (0.93–1.77) | 1.32 (0.91–1.93) | |
| AB | 17 | 10.50 | 1.38 (0.82–2.32) | 1.58 (0.88–2.82) | |
| Non-O | 151 | 9.77 | 1.26 (0.97–1.64) | 1.25 (0.91–1.71) | |
| Leukemia | O | 67 | 5.81 | 1.00 (Reference) | 1.00 (Reference) |
| A | 38 | 5.24 | 0.89 (0.60–1.32) | 0.86 (0.55–1.34) | |
| B | 35 | 5.31 | 0.89 (0.59–1.35) | 0.97 (0.63–1.51) | |
| AB | 6 | 3.71 | 0.63 (0.27–1.46) | 0.64 (0.25–1.59) | |
| Non-O | 79 | 5.11 | 0.86 (0.62–1.20) | 0.89 (0.62–1.27) | |
| Breast (female) | O | 506 | 81.33 | 1.00 (Reference) | 1.00 (Reference) |
| A | 300 | 77.39 | 0.95 (0.82–1.09) | 0.95 (0.82–1.12) | |
| B | 271 | 77.31 | 0.94 (0.81–1.09) | 0.92 (0.78–1.09) | |
| AB | 70 | 80.21 | 0.99 (0.77–1.27) | 0.87 (0.65–1.17) | |
| Non-O | 641 | 77.65 | 0.95 (0.84–1.07) | 0.93 (0.82–1.06) | |
| Prostate (male) | O | 223 | 41.95 | 1.00 (Reference) | 1.00 (Reference) |
| A | 165 | 48.96 | 1.08 (0.88–1.32) | 1.22 (0.98–1.53) | |
| B | 122 | 39.58 | 0.87 (0.70–1.09) | 0.95 (0.74–1.22) | |
| AB | 32 | 42.92 | 0.95 (0.66–1.38) | 0.89 (0.58–1.38) | |
| Non-O | 319 | 44.32 | 0.98 (0.82–1.16) | 1.07 (0.88–1.30) |
Adjusted for age and sex.
Adjusted for age, sex, education level, smoking status, alcohol drinking status, physical activity, and BMI. Individuals with missing covariates were excluded from each model.
Also adjusted for HBV infection.
No consistent patterns of association were observed between ABO blood type and other cancers, including cancers of the oral cavity, esophagus, lung, colon, rectum, leukemia, breast, and prostate (Table 2). The findings for cancer mortality generally mirrored those for incident cancers (Table 3). Compared to blood type O, individuals with blood type A had a significantly elevated risk of stomach cancer mortality (HR = 1.38; 95% CI 1.02–1.86), whereas risk of pancreas cancer mortality was elevated for those with non-O blood type (HR = 1.47, 95% CI 0.98–2.19). Consistent with findings for incident cancer, the elevated risk for pancreas cancer mortality was found across all blood types, with statistical significance reached for blood type B (HR = 1.63, 95% CI 1.02–2.60). Furthermore, adjustment for other variables such as history of diabetes, HBV infection, or family history of cancer did not materially alter the findings for either cancer incidence or mortality (data not shown). Age, sex, or BMI did not modify the associations between ABO blood type and cancer incidence or mortality (data not show).
Table 3.
HRs for over-all and cause-specific cancer mortality by ABO blood type.
| Cancer site | Blood type | N | Mortality rate (per 100,000 person-years) | HRa (95% CI) | HRb (95% CI) |
|---|---|---|---|---|---|
| Overall | O | 1608 | 11.61 | 1.00 (Reference) | 1.00 (Reference) |
| A | 1090 | 12.53 | 1.04 (0.96–1.12) | 1.04 (0.95–1.13) | |
| B | 900 | 11.39 | 0.94 (0.87–1.02) | 0.96 (0.87–1.05) | |
| AB | 205 | 10.56 | 0.89 (0.77–1.03) | 0.90 (0.77–1.06) | |
| Non-O | 2195 | 11.84 | 0.98 (0.92–1.05) | 0.99 (0.92–1.06) | |
| Oral | O | 94 | 0.68 | 1.00 (Reference) | 1.00 (Reference) |
| A | 59 | 0.68 | 0.97 (0.70–1.34) | 0.82 (0.57–1.18) | |
| B | 46 | 0.58 | 0.83 (0.58–1.18) | 0.87 (0.60–1.26) | |
| AB | 8 | 0.41 | 0.60 (0.29–1.23) | 0.51 (0.22–1.17) | |
| Non-O | 113 | 0.61 | 0.87 (0.66–1.15) | 0.81 (0.60–1.09) | |
| Esophagus | O | 38 | 0.27 | 1.00 (Reference) | 1.00 (Reference) |
| A | 24 | 0.28 | 0.96 (0.57–1.60) | 1.05 (0.60–1.83) | |
| B | 19 | 0.24 | 0.83 (0.48–1.44) | 0.96 (0.53–1.75) | |
| AB | 6 | 0.31 | 1.09 (0.46–2.57) | 1.15 (0.45–2.96) | |
| Non-O | 49 | 0.26 | 0.92 (0.60–1.40) | 1.02 (0.64–1.63) | |
| Lung | O | 319 | 2.30 | 1.00 (Reference) | 1.00 (Reference) |
| A | 213 | 2.45 | 1.02 (0.86–1.21) | 0.95 (0.78–1.15) | |
| B | 199 | 2.52 | 1.04 (0.87–1.24) | 1.05 (0.86–1.27) | |
| AB | 46 | 2.37 | 1.00 (0.73–1.36) | 1.03 (0.73–1.44) | |
| Non-O | 458 | 2.47 | 1.03 (0.89–1.18) | 1.00 (0.85–1.17) | |
| Stomach | O | 110 | 0.79 | 1.00 (Reference) | 1.00 (Reference) |
| A | 101 | 1.16 | 1.40 (1.07–1.84) | 1.38 (1.02–1.86) | |
| B | 60 | 0.76 | 0.91 (0.67–1.25) | 0.79 (0.55–1.14) | |
| AB | 9 | 0.46 | 0.57 (0.29–1.12) | 0.70 (0.35–1.38) | |
| Non-O | 170 | 0.92 | 1.11 (0.87–1.41) | 1.06 (0.81–1.38) | |
| Liverc | O | 404 | 2.92 | 1.00 (Reference) | 1.00 (Reference) |
| A | 247 | 2.84 | 0.94 (0.80–1.10) | 1.00 (0.84–1.20) | |
| B | 213 | 2.69 | 0.89 (0.75–1.05) | 0.91 (0.75–1.09) | |
| AB | 48 | 2.47 | 0.83 (0.61–1.12) | 0.80 (0.56–1.13) | |
| Non-O | 508 | 2.74 | 0.91 (0.79–1.03) | 0.94 (0.81–1.09) | |
| Pancreas | O | 45 | 0.33 | 1.00 (Reference) | 1.00 (Reference) |
| A | 43 | 0.49 | 1.47 (0.97–2.23) | 1.34 (0.83–2.17) | |
| B | 40 | 0.51 | 1.49 (0.98–2.29) | 1.63 (1.02–2.60) | |
| AB | 8 | 0.41 | 1.24 (0.59–2.64) | 1.38 (0.61–3.09) | |
| Non-O | 91 | 0.49 | 1.46 (1.02–2.08) | 1.47 (0.98–2.19) | |
| Rectum | O | 52 | 0.38 | 1.00 (Reference) | 1.00 (Reference) |
| A | 46 | 0.53 | 1.36 (0.91–2.02) | 1.42 (0.91–2.21) | |
| B | 41 | 0.52 | 1.32 (0.88–1.99) | 1.49 (0.94–2.34) | |
| AB | 7 | 0.36 | 0.94 (0.43–2.07) | 1.21 (0.54–2.70) | |
| Non-O | 94 | 0.51 | 1.30 (0.93–1.82) | 1.42 (0.97–2.08) | |
| Colon | O | 112 | 0.81 | 1.00 (Reference) | 1.00 (Reference) |
| A | 78 | 0.90 | 1.07 (0.80–1.43) | 1.22 (0.89–1.69) | |
| B | 48 | 0.61 | 0.72 (0.51–1.01) | 0.79 (0.54–1.16) | |
| AB | 15 | 0.77 | 0.94 (0.55–1.61) | 1.00 (0.55–1.84) | |
| Non-O | 141 | 0.76 | 0.90 (0.71–1.16) | 1.02 (0.77–1.35) | |
| Kidney | O | 39 | 0.28 | 1.00 (Reference) | 1.00 (Reference) |
| A | 22 | 0.25 | 0.86 (0.51–1.45) | 0.95 (0.53–1.71) | |
| B | 25 | 0.32 | 1.07 (0.65–1.77) | 1.05 (0.58–1.89) | |
| AB | 7 | 0.36 | 1.25 (0.56–2.80) | 0.97 (0.34–2.76) | |
| Non-O | 54 | 0.29 | 0.99 (0.66–1.50) | 1.00 (0.62–1.61) | |
| Bladder | O | 24 | 0.17 | 1.00 (Reference) | 1.00 (Reference) |
| A | 17 | 0.20 | 1.06 (0.57–1.98) | 1.17 (0.51–2.67) | |
| B | 19 | 0.24 | 1.30 (0.71–2.37) | 1.30 (0.57–2.97) | |
| AB | 3 | 0.15 | 0.86 (0.26–2.84) | 0.54 (0.07–4.10) | |
| Non-O | 39 | 0.21 | 1.14 (0.69–1.90) | 1.16 (0.58–2.32) | |
| Leukemia | O | 32 | 0.23 | 1.00 (Reference) | 1.00 (Reference) |
| A | 17 | 0.20 | 0.82 (0.46–1.48) | 0.76 (0.39–1.46) | |
| B | 14 | 0.18 | 0.74 (0.40–1.39) | 0.83 (0.43–1.61) | |
| AB | 2 | 0.10 | 0.44 (0.11–1.84) | 0.52 (0.12–2.19) | |
| Non-O | 33 | 0.18 | 0.75 (0.46–1.22) | 0.76 (0.45–1.29) | |
| Breast (female) | O | 50 | 0.67 | 1.00 (Reference) | 1.00 (Reference) |
| A | 36 | 0.77 | 1.15 (0.75–1.77) | 1.05 (0.64–1.71) | |
| B | 35 | 0.83 | 1.23 (0.80–1.90) | 1.06 (0.64–1.75) | |
| AB | 7 | 0.67 | 1.01 (0.46–2.23) | 0.89 (0.35–2.26) | |
| Non-O | 78 | 0.79 | 1.17 (0.82–1.67) | 1.03 (0.69–1.55) | |
| Prostate (male) | O | 45 | 0.71 | 1.00 (Reference) | 1.00 (Reference) |
| A | 35 | 0.87 | 1.14 (0.73–1.77) | 1.11 (0.66–1.84) | |
| B | 23 | 0.62 | 0.82 (0.50–1.36) | 0.76 (0.42–1.37) | |
| AB | 5 | 0.56 | 0.74 (0.29–1.85) | 0.77 (0.27–2.18) | |
| Non-O | 63 | 0.73 | 0.96 (0.66–1.41) | 0.92 (0.59–1.44) |
Adjusted for age and sex.
Adjusted for age, sex, education level, smoking status, alcohol drinking status, physical activity, and BMI. Individuals with missing covariates were excluded from each model.
Also adjusted for HBV infection.
4. Discussion
In this large cohort of Chinese men and women with laboratory typing of ABO blood group at baseline, we found that after an average follow-up of 8.75 years, subsequent risk of cancer varied in people with different blood types. Compared to people with blood type O, those with non-O blood type (A, B, and AB combined as well as separately) had a reduced risk of kidney cancer but an elevated risk of pancreas cancer. In addition, blood type A was associated with an excess risk of stomach cancer whereas type AB was associated with a marginal reduction in overall cancer risk.
Since Aird et al. first reported an association between blood type A and gastric cancer 60 years ago [3], this observation has been confirmed in many case-control studies and a cohort of blood donors in Sweden and Denmark [18,26]. Assessment of ABO genotype, which is highly concordant with ABO phenotype [27], has also linked an increased risk of gastric cancer with AA genotype [28]. Our observation of an elevated risk of both gastric cancer incidence and mortality associated with blood type A add to the limited evidence from cohort study. Recently, a study of intestinal metaplasia and dysplasia, both precancerous gastric lesions, found that the association with ABO type was modified by Helicobacter pylori (H. pylori) and cagA status, a known risk factor for gastric cancer [29]. An elevated risk of these lesions was associated with blood type A only among carriers of the virulent cagA+ strains of H. pylori, hypothesized to be due to differential binding of the bacterium to the blood group antigens. Our study did not collect data on H. pylori and cagA status and therefore was unable to confirm an interactive effect of H. pylori and ABO blood group.
Our observation of an increased risk of pancreatic cancer associated with all non-O blood types (A, B, or AB) relative to blood type O is consistent with findings reported in two prospective cohort studies of health professionals in the USA [9]. These findings, however, differ from results in a recent European consortium study [30] and a large population-based case-control study in Shanghai, China [27]. Assessing ABO blood type based on genotype, these studies found an increased risk of pancreatic cancer associated only with blood type A, and not type B or AB, when compared with type O. To explore reasons for the diverse findings, Risch et al. conducted a meta-analysis of 24 studies worldwide totaling 10,415 cases and 869,044 controls [31]. When the studies were stratified by cagA+ H. pylori endemic (mostly East Asian populations) versus cagA non-endemic populations, an increased risk of pancreatic cancer was associated with blood type A in all populations, but with blood type B or AB only in cagA non-endemic populations (mostly Western populations) [31]. However, this observation is not supported in our cohort study in Taiwan, which is a cagA+ H. pylori endemic population [32]. The reasons for the diverse findings are unclear, but further assessment of interactions between ABO blood group and H. pylori carriage and other exposures could be informative. We were unable to examine H. pylori status as these data were not available in our study. The distributions of other potential cancer risk factors, including smoking and alcohol drinking, body mass index, and physical activity level, were not meaningfully different among people of various ABO blood groups in our cohort.
The inverse associations of kidney cancer with blood types A, B, and AB in our study are unexpected and in contrast to the elevated risk of renal cell cancer, the predominant form of kidney cancer, associated with the respective blood types in two prospective cohorts of health professionals [33]. This difference in findings cannot be explained by the inclusion of renal pelvis cancer in our study since the risk of bladder cancer, which has similar histology as renal pelvis cancer, is positively associated with blood types A, B, and AB. A registry-based study comparing kidney cancer cases to other cancer patients reported no association with ABO blood type [4]. It is premature to link kidney cancer with ABO blood type based on the limited and inconsistent evidence.
Our study also found a weak inverse association between overall cancer risk and blood type AB compared to type O. To our knowledge, this observation has not been reported previously. Confirmation in other studies is needed before we can draw any inference from this finding.
The ABO blood group antigens are glycoproteins expressed on the surface of red blood cells and other tissues, including epithelial cells of the gastrointestinal tract [34]. Glycoconjugates are mediators of intercellular adhesion and membrane signaling and therefore can affect cell proliferation and malignant progression [9,34]. These cell surface molecules are also recognized by host immune response and may play a role in immunosurveillance of malignant cells [9,34,35]. Furthermore, individuals with various ABO blood types have been shown to have different inflammatory responses to H. pylori infection, indicated by the release of different levels of proinflammatory cytokines [36]. The cagA strains of H. pylori have been shown to modify the association between single nucleotide polymorphisms at the ABO locus and risk of pancreatic and gastric cancers and precancerous lesions [28–31]. Together these observations suggest that ABO blood type may influence risk of gastric and pancreatic cancers through differential immune function and inflammatory responses. Although mechanisms have been proposed for other cancers, such as kidney cancer, the evidence linking these cancers to ABO blood type is tenuous. Further evidence is needed before mechanisms can be fully considered.
Our study has several notable strengths. The large study size provides an opportunity to comprehensively examine the association between ABO blood type and cancer risks, including risk of uncommon cancers. Misclassification of ABO blood type is minimized by using a laboratory-based measurement. The relatively homogenous and stable study population diminishes potential confounding by other factors. The unique personal identification number and the availability of population-based cancer registry and vital statistics registry, along with continuing health management for the cohort, enhance the completeness and accuracy of follow-up for cancer incidence and mortality. Furthermore, MJ is a homogenous cohort of individuals belonging to the same ethnicity (Han Chinese). There are 98% Han Chinese in Taiwan. The original minority is only 2%. The distribution of blood type of the study population is similar to the Han Chinese: A 20–39%; B 20–39%, and O 30–40% [37]. Such a homogenous population could minimize potential confounding but may limit the generalizability of our study findings to other populations with more diverse prevalence of exposures. Our study is limited by the availability of only phenotype, and not genotype, for the classification of individuals into ABO blood type. It has been shown that compared with people with OO genotype, pancreatic cancer risk was higher among those with AA or BB genotype than those with AO or BO genotype, respectively [28,38]. Similar observations were also reported for gastric cancer, comparing groups with AA or AO with those with OO genotype [31]. Since our study was unable to classify cohort participants by ABO genotype, the cancer risk we observed for subjects with blood group A or B phenotype may be underestimated relative to those with homozygous A or B genotype. Our study is also limited by not having information on H. pylori and cag A status, which could be an effect modifier in the relation between ABO blood group and risk of gastric and pancreatic cancers. Lastly, the number of cases with blood type AB is small for the less common cancers, despite the large study population.
5. Conclusion
In this cohort study with laboratory measured ABO phenotype, we confirmed previous observations linking blood type A to an excess risk of stomach cancer. Compared to people with blood type O, our study suggested that those with blood type A, B, or AB have an elevated risk of pancreas cancer. Further studies are needed to identify the mechanisms through which ABO blood type may influence risk of various cancers in different populations.
Acknowledgments
This work was supported by the MD Anderson Research Trust and MD Anderson Institutional support for the Center for Translational and Public Health Genomics (to X.W.); and Taiwan Department of Health Clinical Trial and Research Center of Excellence (DOH102-TD-B-111-004 to C.P.W.).
Footnotes
Conflicts of interest
The authors declare no conflict of interest.
Authorship contribution
CW, XW, and WC contributed to study conception and design; WS, CW, JL, CW, XP, MH, MT, CT, XW, and WC contributed to acquisition of data, or analysis and interpretation of data; WS, CW, JL, XW, and WC contributed to drafting or revision of the article for important intellectual content; and all authors have approved the final version of the manuscript for submission.
References
- 1.Kruskall MS, AuBuchon JP. Making Landsteiner’s discovery superfluous: safety and economic implications of a universal group O red blood cell supply. Transfus Sci. 1997;18(4):613–20. doi: 10.1016/s0955-3886(97)00061-1. [DOI] [PubMed] [Google Scholar]
- 2.Anstee DJ. The relationship between blood groups and disease. Blood. 2010;115(23):4635–43. doi: 10.1182/blood-2010-01-261859. [DOI] [PubMed] [Google Scholar]
- 3.Aird I, Bentall HH, Roberts JA. A relationship between cancer of stomach and the ABO blood groups. Br Med J. 1953;1(4814):799–801. doi: 10.1136/bmj.1.4814.799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Iodice S, Maisonneuve P, Botteri E, Sandri MT, Lowenfels AB. ABO blood group and cancer. Eur J Cancer. 2010;46(18):3345–50. doi: 10.1016/j.ejca.2010.08.009. [DOI] [PubMed] [Google Scholar]
- 5.Cerovic R, et al. Examining the presence of ABO(H) antigens of blood types in the saliva of patients with oral cancer. Coll Antropol. 2008;32(2):509–12. [PubMed] [Google Scholar]
- 6.Cerovic R, Juretic M, Balen S, Belusic M, Caser L, Rogic M. ABO blood groups in the patients with laryngeal cancer. Otolaryngol Pol. 2000;54(Suppl 31):209–11. [PubMed] [Google Scholar]
- 7.Maur M, Toranzo JC, Caffarello J. Blood groups ABO and cancer of the Esophagus. Sem Med. 1964;124:253–4. [PubMed] [Google Scholar]
- 8.Suadicani P, Hein HO, Gyntelberg F. ABO phenotypes and inflammation-related predictors of lung cancer mortality: the Copenhagen male study—a 16-year follow-up. Eur Respir J. 2007;30(1):13–20. doi: 10.1183/09031936.00062506. [DOI] [PubMed] [Google Scholar]
- 9.Wolpin BM, Chan AT, Hartge P, Chanock SJ, Kraft P, Hunter DJ. ABO blood group and the risk of pancreatic cancer. J Natl Cancer Inst. 2009;101(6):424–31. doi: 10.1093/jnci/djp020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gates MA, Wolpin BM, Cramer DW, Hankinson SE. ABO blood group and incidence of epithelial ovarian cancer. Int J Cancer. 2011;128(2):482–6. doi: 10.1002/ijc.25339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Slater G, Itzkowitz S, Azar S, Aufses AH., Jr Clinicopathologic correlations of ABO and Rhesus blood type in colorectal cancer. Dis Colon Rectum. 1993;36(1):5–7. doi: 10.1007/BF02050293. [DOI] [PubMed] [Google Scholar]
- 12.Chihara Y, Sugano K, Kobayashi A, Kanai Y, Yamamoto H, Nakazono M, et al. Loss of blood group A antigen expression in bladder cancer caused by allelic loss and/or methylation of the ABO gene. Lab Invest. 2005;85(7):895–907. doi: 10.1038/labinvest.3700268. [DOI] [PubMed] [Google Scholar]
- 13.Kvist E, Krogh J, Hjortberg P. Prognostic variables in patients with prostate cancer: influence of blood group ABO (H), the Rhesus system, age, differentiation, tumour stage and metastases. Int Urol Nephrol. 1992;24(4):417–23. doi: 10.1007/BF02550636. [DOI] [PubMed] [Google Scholar]
- 14.Xie J, Qureshi AA, Li Y, Han J. ABO blood group and incidence of skin cancer. PLoS ONE. 2010;5(8):e11972. doi: 10.1371/journal.pone.0011972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gong Y, Yang YS, Zhang XM, Su M, Wang J, Han JD, et al. ABO blood type: diabetes and risk of gastrointestinal cancer in northern China. World J Gastroenterol. 2012;18(6):563–9. doi: 10.3748/wjg.v18.i6.563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wang DS, Wang ZQ, Zhang L, Qiu MZ, Luo HY, Ren C, et al. Are risk factors associated with outcomes in pancreatic cancer? PLoS ONE. 2012;7(7):e41984. doi: 10.1371/journal.pone.0041984. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Xu WH, Zheng W, Xiang YB, Shu XO. ABO blood type is associated with endometrial cancer risk in Chinese women. Chin J Cancer. 2011;30(11):766–71. doi: 10.5732/cjc.011.10305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Edgren G, Hjalgrim H, Rostgaard K, Norda R, Wikman A, Melbye M, et al. Risk of gastric cancer and peptic ulcers in relation to ABO blood type: a cohort study. Am J Epidemiol. 2010;172(11):1280–5. doi: 10.1093/aje/kwq299. [DOI] [PubMed] [Google Scholar]
- 19.Guo N, Wang J, Ness P, Yao F, Dong X, Bi X, et al. Demographics of apheresis platelet donors in five blood centers in China. Transfusion. 2012;52(3):560–6. doi: 10.1111/j.1537-2995.2011.03328.x. [DOI] [PubMed] [Google Scholar]
- 20.Fang C, Cohen HW, Billett HH. Race, ABO blood group, and venous thromboembolism risk: not black and white. Transfusion. 2013;53(1):187–92. doi: 10.1111/j.1537-2995.2012.03665.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wen CP, Wai JP, Tsai MK, Yang YC, Cheng TY, Lee MC, et al. Minimum amount of physical activity for reduced mortality and extended life expectancy: a prospective cohort study. Lancet. 2011;378(9798):1244–53. doi: 10.1016/S0140-6736(11)60749-6. [DOI] [PubMed] [Google Scholar]
- 22.Wen CP, Cheng TY, Tsai MK, Chang YC, Chan HT, Tsai SP, et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan. Lancet. 2008;371(9631):2173–82. doi: 10.1016/S0140-6736(08)60952-6. [DOI] [PubMed] [Google Scholar]
- 23.Wen CP, Lin J, Yang YC, Tsai MK, Tsao CK, Etzel C, et al. Hepatocellular carcinoma risk prediction model for the general population: the predictive power of transaminases. J Natl Cancer Inst. 2012;104(20):1599–611. doi: 10.1093/jnci/djs372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wen CP, Lee JH, Tai YP, Wen C, Wu SB, Tsai MK, et al. High serum iron is associated with increased cancer risk. Cancer Res. 2014;74(22):6589–97. doi: 10.1158/0008-5472.CAN-14-0360. [DOI] [PubMed] [Google Scholar]
- 25.Cox DR. Regression models and life tables (with discussion) J R Statist Soc B. 1972;34:187–220. [Google Scholar]
- 26.Wang Z, Liu L, Ji J, Zhang J, Yan M, Zhang J, et al. ABO blood group system and gastric cancer: a case-control study and meta-analysis. Int J Mol Sci. 2012;13(10):13308–21. doi: 10.3390/ijms131013308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Risch HA, Lu L, Wang J, Zhang W, Ni Q, Gao YT, et al. ABO blood group and risk of pancreatic cancer: a study in Shanghai and meta-analysis. Am J Epidemiol. 2013;177(12):1326–37. doi: 10.1093/aje/kws458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Nakao M, Matsuo K, Ito H, Shitara K, Hosono S, Watanabe M, et al. ABO genotype and the risk of gastric cancer: atrophic gastritis, and Helicobacter pylori infection. Cancer Epidemiol Biomarkers Prev. 2011;20(8):1665–72. doi: 10.1158/1055-9965.EPI-11-0213. [DOI] [PubMed] [Google Scholar]
- 29.Rizzato C, Kato I, Plummer M, Munoz N, Stein A, Jan van Doorn L, et al. Risk of advanced gastric precancerous lesions in Helicobacter pylori infected subjects is influenced by ABO blood group and cagA status. Int J Cancer. 2013;133(2):315–22. doi: 10.1002/ijc.28019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Rizzato C, Campa D, Pezzilli R, Soucek P, Greenhalf W, Capurso G, et al. ABO blood groups and pancreatic cancer risk and survival: results from the PANcreatic Disease ReseArch (PANDoRA) consortium. Oncol Rep. 2013;29(4):1637–44. doi: 10.3892/or.2013.2285. [DOI] [PubMed] [Google Scholar]
- 31.Nakao M, Matsuo K, Hosono S, Ogata S, Ito H, Watanabe M, et al. ABO blood group alleles and the risk of pancreatic cancer in a Japanese population. Cancer Sci. 2011;102(5):1076–80. doi: 10.1111/j.1349-7006.2011.01907.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lai CH, Perng CL, Lan KH, Lin HJ. Association of IS605 and cag-PAI of Helicobacter pylori isolated from patients with gastrointestinal diseases in Taiwan. Gastroenterol Res Pract. 2013;2013:356217. doi: 10.1155/2013/356217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Joh HK, Cho E, Choueiri TK. ABO blood group and risk of renal cell cancer. Cancer Epidemiol. 2012;36(6):528–32. doi: 10.1016/j.canep.2012.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hakomori S. Antigen structure and genetic basis of histo-blood groups A, B and O: their changes associated with human cancer. Biochim Biophys Acta. 1999;1473(1):247–66. doi: 10.1016/s0304-4165(99)00183-x. [DOI] [PubMed] [Google Scholar]
- 35.Hakomori S. Tumor-associated carbohydrate antigens defining tumor malignancy: basis for development of anti-cancer vaccines. Adv Exp Med Biol. 2001;491:369–402. doi: 10.1007/978-1-4615-1267-7_24. [DOI] [PubMed] [Google Scholar]
- 36.Alkout AM, Blackwell CC, Weir DM. Increased inflammatory responses of persons of blood group O to Helicobacter pylori. J Infect Dis. 2000;181(4):1364–9. doi: 10.1086/315375. [DOI] [PubMed] [Google Scholar]
- 37.Kilgore B, JHL Iso-agglutination test of Chinese bloods for transfusion compatibility. China Med J. 1918;32:21. [Google Scholar]
- 38.Wolpin BM, Kraft P, Gross M, Helzlsouer K, Bueno-de-Mesquita HB, Steplowski E, et al. Pancreatic cancer risk and ABO blood group alleles: results from the pancreatic cancer cohort consortium. Cancer Res. 2010;70(3):1015–23. doi: 10.1158/0008-5472.CAN-09-2993. [DOI] [PMC free article] [PubMed] [Google Scholar]