SEARCH

SEARCH BY CITATION

Keywords:

  • family history;
  • cancer susceptibility;
  • sibling history;
  • smoking

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

BACKGROUND

Inherited genetic predispositions are important risk factors for the development of cancer in general. To determine genetic susceptibility for 14 common cancers, a case–control study of the impact of a family history of cancer in first-degree relatives was conducted. The authors further evaluated the effect modification by habitual smoking with adjustment for other confounding environmental factors.

METHODS

In total, 18,836 cancer cases and 28,125 age–matched and sex–matched controls, confirmed as being free of cancer, were recruited. Odds ratios (ORs) and 95% confidence intervals were determined by multiple logistic regression analysis, including stratification by family history for 14 cancer sites and interactions with a smoking history.

RESULTS

The associations between family history and risk of cancer were generally stronger at the same sites than across cancer sites. Risks to first-degree relatives at the same sites were found to be significantly elevated with 8 of 14 cancer sites; especially high ORs were found for prostate and thyroid cancers. Some across-site associations were observed; in particular, a reciprocal association between breast and prostate cancer was found. The interaction between family history and smoking history for breast cancer was found to be statistically significant. There was no statistical evidence for the interactions in other sites, but among subjects with a family history, the ORs were found to be higher in smokers compared with nonsmokers.

CONCLUSIONS

The results of the current study support the hypothesis of a genetic susceptibility to cancers in family members. For breast cancer, the interaction between family history and smoking history was observed to be significant. Cancer 2007. © 2007 American Cancer Society.

A family history of cancers has consistently been shown to increase the risk of neoplasia in several body sites. When first-degree relatives are involved, there is generally a 2-fold to 5-fold elevation in the likelihood of cancer development.1, 2

Most epidemiologic studies of family history have focused on the same anatomic sites among first-degree relatives, and to our knowledge only rarely has a systematic review of multiple cancer sites and sibling influence been conducted. Major registry-based studies, conducted in Western countries, have been published concerning family history across multiple sites,1, 3 but a limitation of these studies was the lack of controlling for important confounding factors, making it difficult to determine whether the familial aggregation observed was due to shared genetic background or environmental factors.

To assess familial cancer risks for common sites in Japan, we conducted a large-scale, case–control study with a further emphasis on interactions with smoking, as well as other confounding factors.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Study Subjects

Cases and controls were selected from the database of the Hospital-based Epidemiologic Research Program at Aichi Cancer Center (HERPACC), conducted at Aichi Cancer Center Hospital. Details of the HERPACC have been described elsewhere.4, 5 In brief, we enrolled patients in HERPACC-I between January 1988 and December 2000. Information regarding lifestyle factors as well as other relevant factors, including family history, were collected from all first-visit outpatients at our hospital ages 18–79 years using a self-administered questionnaire, with checks by trained interviewers. After HERPACC-I, HERPACC-II was continued between January 2001 and November 2005, in which all first-visit outpatients were asked to provide blood samples in addition to the information provided in the questionnaire. Each patient was asked for information when they were healthy or before the current symptoms developed. Similar to most general hospitals in Japan, our hospital accepts new outpatients who visit of their own volition, with or without a physician referral. Thus, even though the Cancer Center is called a cancer hospital, only 19% of all new outpatients have been diagnosed with cancer.6 Among noncancer outpatients, 45% present with no abnormal findings on clinical examination and 35% present with benign, nonspecific diseases.7 Our previous study showed that the lifestyle patterns of first-visit outpatients are in accordance with those in a randomly selected sample of the general population of Nagoya City.8 Therefore, there is a rationale for adopting noncancer outpatients at our hospital as appropriate controls for epidemiologic studies to evaluate cancer risk factors. The data were loaded into a HERPACC database and routinely linked with the hospital-based cancer registry system to update the data regarding cancer incidence. The study was approved by the Institutional Ethics Committee of Aichi Cancer Center.

The current study was based on data collected between January 1988 and December 2004, during which time a total of 119,938 patients visited our hospital as first-visit outpatients. Among them, 13,865 patients (11.6%) were not enrolled due to refusal, age ineligibility (age <18 years), interviewers'' absence, or other miscellaneous reasons. The questionnaire was ultimately administered to 106,073 patients. Finally, 104,801 subjects (98.8%) completed the questionnaire satisfactorily and were enrolled in the HERPACC.

In the current study, patients ages 20 to 79 years with 14 types of incident cancers and who were newly diagnosed between January 1988 and the end of 2004 were deemed eligible as case subjects. Information regarding cancer patients was obtained using a hospital-based cancer registry system. The 14 common cancers were defined according to the following codes of the International Classification of Diseases and Related Health Problems (ICD10): head and neck (C00-06, 10–14, and 30–32), esophagus (C15), stomach (C16), colorectum (C18-20), liver (C22), pancreas (C25), lung (C34), breast (C50), uterus (C53-55), ovary (C56), prostate (C61), bladder (C67), thyroid (C73), and lymphoma (C81-85). A total of 18,836 cancer patients (14,040 from HERPACC-I and 4796 from HERPACC-II) without a prior history of cancer were observed.

We randomly selected 1 or 2 controls for each case from among the 79,090 first-visit outpatients (64,501 from HERPACC-I and 14,589 from HERPACC-II) who were ages 20 to 79 years and confirmed to be free of cancer by diagnostic procedures at our hospital between 1988 and 2004, with matching for age and sex performed in each study. To increase the statistical power of the study, we conducted 2:1 matching if the control sample size was adequate in each age-sex group. Finally, 28,125 controls (22,007 from HERPACC-I and 6118 from HERPACC-II) were included in this study.

Exposure Data

The HERPACC questionnaire included items regarding demographic characteristics; individual medical history; height and weight; exercise, smoking, and drinking habits; and the consumption of selected foods and beverages as well as family history. All exposures were measured by a self-administered questionnaire given at the time of the first visit to our hospital and before any diagnostic procedures were conducted.

Information regarding family history of distant relatives is generally limited and less precise than for that closer or first-degree relatives. In this study, the existence of parents or siblings with cancer was considered as a positive family history. It was impossible to obtain a family history concerning type of uterine cancer (ie, endometrial or cervical) because of the categorization in the questionnaire.

Smoking history was entered under the 3 categories of never, former, and current. Former and current smokers were categorized as smokers. The cumulative smoking dose was evaluated as pack-years, the product of the number of packs consumed per day and years of smoking. Drinking history was entered into the 3 categories of never, former, and current. Former smokers and drinkers were defined as those who quit smoking and drinking at least 1 year before the survey.

Statistical Analyses

We assessed the effect of family history of cancer in terms of the odds ratios (ORs) and 95% confidence interval (95% CIs) calculated using age-matched and sex-matched conditional logistic models adjusted for potential confounders. For analysis of interaction between family history and smoking history, an unconditional logistic regression model was used because the matching was not retained after stratification by smoking history. Potential confounders considered in the multivariate analyses were age, sex, smoking history (never smokers, former smokers, or current smokers), drinking history (never drinkers, former drinkers, or current drinkers), body mass index (BMI) (<18.5 kg/m2, 18.5–24.9 kg/m2, or ≥25.0 kg/m2), regular exercise (yes or no), and referral pattern (patient's discretion, family recommendation, referral from other clinics, secondary screening after primary screening, or other). Interactions between family history and smoking history were evaluated under the multiplicative assumption. Products of scores for family history (0 indicates no and 1 indicates yes) and smoking history (0, never smoker; 1, smokers with ≤30 pack-years; and 2, smokers with >30 pack-years) were included as interaction terms. A P value <.05 was applied for statistical significance. All analyses were performed using STATA software (version 8; StataCorp, College Station, TX).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Table 1 summarizes the overall characteristics of the cases for the 14 common cancer sites and controls. A total of 18,836 cases and 28,125 controls that were matched with cases with regard to age and sex were included in the current study. Compared with the control group, the proportion of current smokers was higher in the group of cases. No difference was observed between the cases and controls with regard to drinking history or BMI. The percentage of subjects undertaking regular exercise was higher in the control group. Referral from other clinics was frequent, whereas patient discretion and secondary screening after primary screening was less common among the case group than the controls.

Table 1. Characteristic of All Cases of 14 Cancer Sites and Controls
 CasesControls
  • SD indicates standard deviation; BMI, body mass index.

  • *

    Former smokers and drinkers were defined as subjects who had quit smoking and drinking at least 1 year previously.

Total18,83628,125
Age, y (%)
 20–29275 (1.5)550 (2.0)
 30–391367 (7.3)2734 (9.7)
 40–493484 (18.5)6881 (24.5)
 50–595365 (28.5)9615 (34.2)
 60–695579 (29.6)5579 (19.8)
 70–792766 (14.7)2766 (9.8)
Sex (%)
 Male8874 (47.1)12,088 (43.0)
 Female9962 (52.9)16,037 (57.0)
Smoking history (%)
 Never9152 (48.6)16,081 (57.2)
 Former*3092 (16.4)4523 (16.1)
 Current6490 (34.5)7357 (26.2)
 Unknown102 (0.5)164 (0.6)
Drinking history (%)
 Never9354 (49.7)14,696 (52.3)
 Former*709 (3.8)739 (2.6)
 Current8701 (46.2)12630 (44.9)
 Unknown72 (0.4)60 (0.2)
Mean BMI (SD)22.5 (3.1)22.5 (2.9)
Regular exercise (%)
 Yes9902 (52.6)15,720 (55.9)
 No8826 (46.9)12,277 (43.7)
 Unknown108 (0.6)128 (0.5)
Referral pattern to our hospital (%)
 Patient's discretion3430 (18.2)10,326 (36.7)
 Family recommendation3441 (18.3)5599 (19.9)
 Referral from other clinics8788 (46.7)4911 (17.5)
 Secondary screening after primary screening2916 (15.5)6597 (23.5)
 Others117 (0.6)470 (1.7)
 Unknown144 (0.8)222 (0.8)

To examine the associations between family history of each tumor type and risks of cancers, we conducted analyses stratified by family history for 14 common cancer sites (Table 2). We observed stronger associations within the same cancer sites than across cancer sites. The risks at the same sites with first-degree relatives were found to be significantly elevated for 8 cancer sites. Especially high ORs were found for cancers of the prostate and thyroid. Some across-site associations were found for esophagus–liver, stomach–bladder, colorectum–lymphoma, and pancreas–ovary. Brest and prostate cancers exhibited a pattern of reciprocal association.

Table 2. Adjusted ORs* and 95% CIs for 14 Cancers With a Family History of the Cancers Reported in First-Degree Relatives
 Site of cancer in first-degree relatives
Head and neck (n = 759)Esophagus (n = 644)Stomach (n = 6988)Colorectum (n = 2712)Liver (n = 1806)Pancreas (n = 983)Lung (n = 2605)Breast (n = 1680)Uterus (n = 1504)Ovary (n = 218)Prostate (n = 384)Bladder (n = 402)Thyroid (n = 130)Lymphoma (n = 261)
ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)ORs(95% CIs)
  • ORs indicates odds ratio; 95% CI, 95% confidence interval; NA, data not available.

  • *

    Matched for age and sex and adjusted for smoking history, drinking, BMI, exercise habit, and referral pattern to our hospital.

  • OR was not available because of insufficient data.

  • A sum of 14 site cases is not equal to the number of total cases in Table 1 because there were cases of double primary cancers.

  • Bold-face type indicates same site ORs.

Head and neck (n = 1471)0.8(0.4–1.6)0.6(0.2–1.3)0.8(0.6–1.1)1.3(0.9–2.0)0.9(0.5–1.5)1.0(0.5–1.9)1.0(0.6–1.5)0.6(0.3–1.1)1.2(0.7–2.1)0.7(0.2–3.3)0.7(0.2–2.5)0.7(0.2–1.8)NA 3.1(0.6–15.7)
Esophagus (n = 826)1.9(0.9–4.2)2.7(1.1–6.7)1.0(0.7–1.5)0.6(0.4–1.0)2.2(1.2–4.1)0.7(0.3–1.6)1.5(0.9–2.6)0.7(0.4–1.5)1.2(0.6–2.6)0.3(0.1–2.8)1.2(0.2–7.1)2.2(0.4–12.3)NA 1.7(0.3–10.4)
Stomach (n = 3192)0.7(0.5–1.0)0.8(0.5–1.2)1.6(1.4–1.8)0.9(0.7–1.1)0.9(0.7–1.1)1.2(0.8–1.6)1.0(0.8–1.3)0.9(0.7–1.3)1.4(1.0–1.8)0.7(0.3–1.5)0.9(0.5–1.6)1.8(1.1–3.0)0.7(0.3–2.0)2.0(1.0–3.9)
Colorectum (n = 2189)1.1(0.7–1.9)0.6(0.3–1.0)1.1(0.9–1.3)1.6(1.3–2.0)0.8(0.5–1.0)0.9(0.6–1.5)1.0(0.8–1.3)0.9(0.6–1.3)0.9(0.7–1.3)1.7(0.6–4.7)1.0(0.5–1.9)0.9(0.5–1.9)1.2(0.4–3.5)2.7(1.2–6.4)
Liver (n = 612)1.2(0.5–3.1)1.2(0.4–3.6)0.8(0.6–1.2)0.7(0.4–1.3)3.4(1.9–6.1)1.5(0.6–3.4)0.6(0.4–1.2)0.7(0.3–1.7)1.1(0.6–2.3)1.2(0.1–12.3)0.2(0.1–0.9)1.5(0.5–5.2)2.0(0.1–41.2)0.3(0.1–1.9)
Pancreas (n = 389)1.8(0.5–6.0)2.2(0.5–11.1)0.7(0.5–1.2)1.4(0.7–2.5)1.1(0.5–2.7)1.2(0.5–2.9)1.0(0.5–1.9)0.9(0.4–2.3)1.5(0.6–3.6)8.2(1.2–57.5)9.9(1.0–103.2)0.4(0.1–4.0)0.8(0.1–50.3)4.0(0.1–346.4)
Lung (n = 2727)0.8(0.5–1.3)0.6(0.3–1.0)0.9(0.7–1.1)0.8(0.6–1.1)0.8(0.6–1.1)0.8(0.6–1.3)1.3(1.1–1.7)0.8(0.5–1.1)0.8(0.5–1.1)0.9(0.3–2.6)1.4(0.7–2.6)0.9(0.5–1.8)0.4(0.1–1.6)0.6(0.3–1.5)
Breast (n = 3861)1.0(0.7–1.5)0.7(0.5–1.1)1.0(0.9–1.1)0.8(0.7–1.0)0.8(0.6–1.0)1.1(0.8–1.5)0.8(0.7–1.0)1.5(1.3–1.9)0.9(0.7–1.2)1.0(0.6–1.8)1.7(1.2–2.6)0.8(0.5–1.3)0.9(0.4–1.9)1.2(0.7–2.0)
Uterus (n = 1718)0.6(0.3–1.0)1.0(0.5–1.9)0.9(0.7–1.1)0.8(0.6–1.1)1.3(0.9–1.9)0.8(0.5–1.4)1.0(0.7–1.4)0.9(0.6–1.3)1.2(0.8–1.9)0.8(0.3–2.2)0.7(0.3–1.8)0.5(0.2–1.1)NA 0.5(0.2–1.3)
Ovary (n = 325)1.8(0.4–7.8)2.9(0.7–12.5)0.9(0.6–1.5)0.8(0.4–1.6)1.2(0.6–2.6)0.9(0.3–3.0)1.0(0.5–2.0)1.3(0.6–2.6)2.3(0.8–6.6)1.7(0.4–7.9)0.6(0.1–32.0)2.1(0.2–20.8)NA 0.4(0.1–5.2)
Prostate (n = 257)1.0(0.1–7.1)1.9(0.4–8.4)0.8(0.5–1.3)1.8(0.9–3.9)0.8(0.3–2.0)0.6(0.2–1.7)0.7(0.3–1.5)3.6(1.1–11.7)0.7(0.3–1.7)NA 5.6(1.5–20.5)NA NA 1.0(0.1–7.5)
Bladder (n = 195)2.9(0.5–15.6)4.2(0.7–24.1)0.5(0.3–1.0)1.4(0.6–3.3)0.4(0.1–2.4)1.1(0.3–3.9)2.5(0.8–7.5)0.6(0.2–2.4)2.2(0.6–8.5)NA 2.8(0.4–19.0)3.4(0.5–24.3)NA NA 
Thyroid (n = 400)1.0(0.2–3.7)1.0(0.3–3.8)1.2(0.7–1.9)0.8(0.4–1.6)1.5(0.6–3.4)0.6(0.2–2.3)1.4(0.8–2.5)1.2(0.6–2.3)1.0(0.4–2.4)0.2(0.1–1.9)0.9(0.1–7.1)0.8(0.1–4.5)13.3(1.5–114.7)2.5(0.4–17.2)
Lymphoma (n = 685)0.8(0.3–2.2)0.4(0.1–1.4)1.2(0.8–1.8)1.2(0.7–2.0)1.0(0.5–2.2)0.6(0.2–1.5)1.1(0.6–2.0)0.8(0.4–1.6)1.7(0.8–3.5)2.4(0.1–66.0)1.2(0.3–4.6)0.4(0.1–1.9)0.4(0.1–4.2)0.9(0.3–3.2)

Risks for the same sites according to a family history in parents and siblings are shown in Table 3. Significantly increased risks were found for 5 sites in parents and 6 sites in siblings. Pancreatic and lung cancers demonstrated significantly elevated ORs with siblings only and esophageals cancer demonstrated a significantly elevated OR with parents only. No clear association was observed among subjects with both a parent and sibling family history due to small sample size (data not shown).

Table 3. Adjusted ORs and 95% CIs for 14 Cancers With Family History at the Same Site According to Parents and Siblings
 ParentsSiblings
ORs*(95% CIs)ORs*(95% CIs)
  • ORs indicates odds ratios; 95% CIs, 95% confidence intervals; NA, not available.

  • *

    Matched for age and sex and adjusted for smoking history, drinking history, body mass index, exercise habit, and referral pattern to our hospital.

  • NA indicates that the OR was not available because of insufficient data.

Head and neck0.59(0.27–1.26)3.72(0.88–15.73)
Esophagus4.81(1.51–15.33)0.68(0.14–3.43)
Stomach1.40(1.20–1.63)1.76(1.39–2.23)
Colorectum1.56(1.17–2.08)1.52(1.08–2.14)
Liver3.33(1.48–7.52)2.92(1.26–6.77)
Pancreas0.44(0.14–1.44)6.11(1.23–30.32)
Lung1.16(0.85–1.57)1.53(1.07–2.18)
Breast1.58(1.19–2.10)1.53(1.20–1.96)
Uterus1.03(0.63–1.69)1.79(0.82–3.88)
Ovary0.46(0.06–3.60)9.26(0.58–146.81)
Prostate7.52(0.93–60.96)2.49(0.57–10.87)
Bladder3.23(0.44–23.77)NA 
Thyroid4.63(0.47–45.74)NA 
Lymphoma0.22(0.02–2.47)1.84(0.38–8.98)

Data regarding the impact of smoking history in combination with a family history are shown in Table 4. The interaction was found to be statistically significant in breast cancer (P = .01). There was no statistical evidence of interactions at other sites, but among subjects with a family history, the ORs were found to be higher in smokers than in nonsmokers.

Table 4. Association Between Smoking History and Risk of Cancers By Family History At Same Site in First-Degree Relatives
 Family history in same siteSmoking history
NonsmokerSmokerInteraction P
≤ 30 Pack-Years> 30 Pack-Years
ORs*(95% CIs)ORs*(95% CIs)ORs*(95% CIs)
  • ORs indicates odds ratio; 95% CIs, 95% confidence intervals; NA, not available.

  • *

    Adjusted for age, sex, drinking history, body mass index, exercise habit, and referral pattern to our hospital.

  • Interactions were obtained from the model in which the family history and smoking status were classified as No/Yes (0/1) and never/pack-years, ≤30/pack-years, and >30 pack-years (0/1/2), respectively. Interactions were modeled as a product of family history in score and smoking status in score.

  • NA indicates the OR and interaction-P value were not available because of insufficient data.

  • Note that ovary and bladder cancers were not shown because of insufficient data.

Head and neckNo1.00(Reference)1.35(1.11–1.65)2.04(1.67–2.49) 
Yes0.85(0.34–2.12)3.01(1.37–6.64)1.70(0.90–3.21).71
EsophagusNo1.00(Reference)2.16(1.59–2.95)3.82(2.83–5.16) 
Yes3.14(0.96–10.30)5.34(2.29–12.46)5.70(2.96–10.95).20
StomachNo1.00(Reference)1.54(1.34–1.77)1.73(1.50–2.00) 
Yes1.66(1.42–1.95)2.60(2.14–3.15)2.49(2.07–2.99).22
ColorectumNo1.00(Reference)0.98(0.84–1.14)0.92(0.78–1.08) 
Yes1.49(1.19–1.86)1.95(1.41–2.71)1.57(1.14–2.15).36
LiverNo1.00(Reference)1.32(0.99–1.77)1.22(0.91–1.63) 
Yes1.91(1.06–3.46)2.22(1.18–4.16)3.42(2.09–5.61).27
PancreasNo1.00(Reference)1.09(0.78–1.53)1.16(0.82–1.63) 
Yes1.93(0.88–4.24)1.03(0.25–4.34)3.07(1.40–6.72).58
LungNo1.00(Reference)1.67(1.42–1.98)4.48(3.80–5.28) 
Yes1.55(1.16–2.08)2.59(1.75–3.84)6.26(4.82–8.12).56
BreastNo1.00(Reference)0.98(0.87–1.10)0.97(0.72–1.31) 
Yes1.44(1.21–1.71)1.95(1.36–2.81)4.33(1.65–11.40).01
UterusNo1.00(Reference)1.42(1.22–1.65)1.36(0.93–2.01) 
Yes1.36(0.98–1.88)1.25(0.60–2.61)2.50(0.74–8.41).73
ProstateNo1.00(Reference)0.91(0.62–1.33)0.62(0.43–0.90) 
Yes4.52(1.60–2.80)8.66(2.97–25.27)7.65(3.31–17.71).14
ThyroidNo1.00(Reference)0.82(0.60–1.13)0.51(0.30–0.85) 
Yes6.42(2.82–14.61)NA 13.59(2.72–67.88)NA
LymphomaNo1.00(Reference)1.03(0.81–1.30)0.76(0.58–0.99) 
Yes1.29(0.31–5.39)1.88(0.43–8.24)3.28(0.90–1.96).22

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The current case–control study, using a population derived from hospital outpatients and adjusted for various lifestyle factors, provided clear evidence for familial aggregation in the majority of 14 common sites of cancer among Japanese. Having a family history of a disease can imply shared genetic factors, shared environment, or both. In the current study, we adjusted for potential confounding factors to minimize the influence of environment factors. To our knowledge, only 1 previous cohort study has systematically examined the risk of family history for women across all organs in this way.2 Furthermore, the strengths of the current study are that age confounding was able to be completely controlled by exact age matching.

A positive family history is a significant risk factor at most cancer sites,1–3 as confirmed herein. The risk was particularly high for prostate and thyroid cancer, a finding that is consistent with other studies.1, 3 Similar to previous studies, the ORs for breast cancer were not significantly high in the current study,1–3 although family history is a well-established risk factor for breast cancer.9

Although systematic studies of across-site associations of familial cancer risk have to our knowledge been limited,1–3, 10 many articles have been published to date regarding individual cancer sites. A reciprocal association between breast and prostate cancers was observed in the current study. Breast and prostate cancers clearly cluster in families,11–13 and in many cases a genetic basis has been established.14–16 An increased risk for pancreatic cancer among those with a family history of ovarian and prostate cancer was found in the current study. Pancreatic cancer also has been observed as a feature of the BRCA2 spectrum in several other studies.17, 18 However, the 95% CIs of these risks were rather wide in the current study because of the small number of subjects. Therefore, careful interpretation is required.

Although the results from the across-sites analysis must be viewed with caution due to reduced numbers of observations in the subgroups, among the other sites, the 3 across-sites (esophagus–liver, stomach–bladder, and colorectum–lymphoma) exhibited positive associations in the current study. In an earlier report, no association with stomach–bladder and colorectum–lymphoma was reported.10 To our knowledge, the observation of esophageal–liver cancers has not been reported in other studies. However, interpretation of these associations may be difficult; lymphomas especially feature a diversity of subtypes whose etiologies differ. In addition, among the well-known across-site associations, no significantly increased risks between breast–ovary and breast–colorectum were observed in the current study. Therefore, replication of the current findings in a future study is warranted.

In the current study, no clear difference between parents and sibling family history for cancer risk were observed except for esophageal, pancreatic, and lung cancer, and similar ORs were observed in stomach, colorectal, liver, and breast cancers. Considering the age difference, perhaps the similar risks noted between siblings and parents may be interpreted as differences. Previous investigations consistently revealed greater associations with siblings than parents.3, 19 The exact reason for these findings is unknown. It may be due to recessive genetic factors or a shared environment in the same family during the same period.

For breast cancer, a significant interaction between family history and smoking was observed in the current study. The data linking cigarette smoking with breast cancer are conflicting in the general population20–22 and those in high–risk groups.23, 24 A possible explanation for the result is that in the absence of functional protein from the DNA repair genes, BRCA1 and BRCA2, and perhaps other predisposition genes, multiple carcinogens in cigarette smoke may induce DNA damage. In addition, a higher risk among smokers with a family history compared with nonsmokers with a family history was herein demonstrated in other sites. Genetic predispositions may reflect common genetic polymorphisms in genes that are involved in the biotransformation of procarcinogens in tobacco such as cytochrome P450,25 glutathione S-transferase,26 and N-acetyltransferase.27

In the current study, a decreased risk for prostate and thyroid cancer was found only for smokers without a family history. For prostate cancer, earlier studies demonstrated no association,28 except in a case–control study in which a reduction was noted in heavy smokers.29 A previous meta–analysis reported that a reduced risk in thyroid cancer was associated with smoking.30 The results of the current study also demonstrated a smaller effect of smoking on the risk of lung cancer compared with studies in Western countries, a finding that is consistent with previous study in performed in Japan.31

Before drawing conclusions from the current study, certain potential limitations should be considered. One possibility is recall bias, in which cases are apt to recall more cancers among family members than are control subjects. This may lead to the appearance of an increased risk among those with a family history of cancer. Because it was not possible to evaluate the accuracy of self-reported family histories, it therefore cannot be excluded that some family histories may have been misclassified. However, the questionnaires were conducted before the diagnosis of any disease at our hospital, and we did not focus on any specific cancer when collecting the data. Therefore, it is unlikely that the recall bias substantially distorted the results. Moreover, to minimize such misclassifications, we limited subjects to those ages 20 to 79 years and focused on parents and siblings only because their history is likely to be reported with greater accuracy compared with that of more distant relatives.32, 33 Another possibility that deserves consideration is detection bias,34 meaning that subjects with a family history were possibly more likely to visit our hospital. If this was the case, then the incidence of a family history of cancer among control subjects would be higher than among the general population, which could lead to an underestimation of links. The internal validity of this hospital-based study is also a potential threat to causal inference in this population. We used noncancer patients at our hospital as controls, given the likelihood that our cases arose within this population base. However, earlier comparisons of lifestyle characteristics between these outpatients in our hospital and individuals randomly selected from the general population confirmed that they were not substantially different.8 The medical background of controls is another potential source of bias; however, our previous study, which focused on a female population, demonstrated only a limited impact for this variable,35 and little difference would be expected for males. To account for differences between cases and controls, we adjusted for the pattern of referral to our hospital. Given this background and methodology, we conclude that the present use of noncancer outpatients as referents in our HERPACC-type epidemiologic study is reasonable. Furthermore, such use is associated with higher compliance and accuracy of information in questionnaire surveys than with the general population. In the current study, we obtained a 98.8% response rate. In addition, the majority of the subjects were from Aichi prefecture, giving comparable residential areas for cases and for the referent group. The relatively small sample size of subjects is a potential limitation of the current study with regard to exploring risks with across-site family history. Therefore, careful interpretation is required. We were not able to evaluate only the genetic predispositions and completely exclude the environment factors because we could not obtain information regarding lifestyle factors of the subjects' relatives.

In summary, the results of the current study demonstrated excess risks with a family history for several cancers. Cancer risk counseling, the management of high–risk patients, and decisions regarding the appropriateness of predictive genetic testing also require information concerning familial cancer history. Therefore, further systematic epidemiologic studies of family history in relation to the development of cancer are warranted.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, Culture, and Technology of Japan; and by a Grant-in-Aid for the Third Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health, Labour, and Welfare of Japan.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES