Search

Background: The established risk factors for female breast cancer, including nulliparity, age at first birth, age at menarche, and age at menopause, do not adequately explain the occurrence pattern of this cancer. Therefore, additional factors need to be considered to advance our understanding of the causes of breast cancer. Evidence obtained from animals and humans indicates that the perinatal period may be particularly important. Our earlier studies, based on a subsample of the present investigation, suggested that factors thought to be positively associated with estrogen levels during pregnancy are positively associated with breast cancer risk. However, the associated confidence intervals (CIs) were generally wide, indicating considerable variability in the data. Purpose: We studied a large number of incident breast cancer case patients and evaluated several perinatal variables in relation to breast cancer risk. These variables included some not previously studied, such as gestational age and neonatal jaundice. We also investigated twin membership as a predictor variable and death from breast cancer, in addition to occurrence of breast cancer, as an alternative outcome variable. Methods: Birth records for all deliveries at five different hospitals in Sweden during the period from 1874 through 1961 were used to define a large cohort of women. Incident case patients with breast cancer in this cohort were ascertained through the National Cancer Registry or the Uppsala Regional Cancer Registry from 1958 through 1994. In a case-control study nested in the cohort, we abstracted data from birth records on 1068 women with incident breast cancer and on 2727 control subjects individually matched to the case patients on date of birth. We modeled the data through conditional logistic regression. All P values were derived from two-sided statistical tests. Results: We found a markedly reduced risk for breast cancer in women whose mothers had pregnancy toxemia (odds ratio [OR] = 0.41; 95% CI = 0.22-0.79) and an excess risk on women who had neo-natal jaundice (OR = 0.16; 95% CI =1.27-3.67) or who were born before 33 weeks of gestation (OR = 3.96; 95% CI = 1.45-10.81). Compared with the risk for breast cancer in the singleton, the risk for breast cancer in dizygotic twins was increased, although this increase was not statistically significant (OR = 1.72; 95% CI = 0.92-3.20). There was no evidence in these data for a statistically significant or substantial association of breast cancer risk with birth size indicators (i.e., weight and length at birth and placental weight). Conclusion and Implications: Because pregnancy toxemia is associated with low levels of estrogens and neonatal jaundice, severe prematurity, and dizygotic twins with high levels of estrogens, our findings suggest that estrogens and other hormonal factors, known to influence breast cancer risk in the adult, may also play a critical role during the intrauterine period. [J Natl Cancer Inst 1997;88:71-6]

Background: Chronic infection with hepatitis B virus, alcohol consumption, and cirrhosis of the liver are recognized risk factors for primary liver cancer. A few, but not all, studies have suggested that diabetes mellitus also increases risk for this cancer. Purpose: We conducted a population-based cohort study to analyze the risk of developing primary liver cancer and biliary tract (gallbladder, extrahepatic bile ducts, and ampulla of Vater) cancers among patients with diabetes. Methods: A cohort of 153 852 patients with a hospital discharge diagnosis of diabetes in the period from 1965 through 1983 was identified by use of the Swedish In-patient Register. Follow-up for these patients extended from the date of cohort entry through December 31, 1989. Incident cases of cancer during follow-up were identified through the Swedish Cancer Registry. To minimize the impact of selection bias, we excluded from the analysis patients who were diagnosed with liver and biliary tract cancers during the first year of follow-up. Standardized incidence ratios (SIRs) and their 95% confidence intervals (CIs) were computed by use of nationwide rates of liver and biliary tract cancers, adjusted for age, sex, and calendar year, for comparison. Results: During 1-24 years of follow-up, 819 incident cancers in the combined category of primary liver (n = 533) and biliary tract (n = 286) were identified in the cohort, yielding an overall SIR of 2.5 (95% CI = 2.3-2.6). The risk was higher in men (SIR = 3.2; 95% CI = 2-9-3.6) than in women (SIR = 2.0; 95% CI = 1.8-2.2). The incidence of primary liver cancer alone was increased fourfold (SIR = 4.1; 95% CI = 3.8-4.5); again, the risk was higher in men (SIR = 4.7; 95% CI = 4.2-5.2) than in women (SIR = 3.4; 95% CI = 2.9-3.9). Smaller increases in risk were seen for cancers of the gallbladder, the extrahepatic bile ducts, and the ampulla of Vater. After exclusion of diabetic patients with concomitant diseases that predispose to primary liver cancer, such as alcoholism, cirrhosis, and hepatitis, the persistence of an approximately threefold excess risk was observed. Conclusions: Our findings suggest that patients with diabetes are at increased risk of developing primary liver cancer and perhaps cancers of the biliary tract. The mechanisms involved in the association of diabetes and liver cancer remain to be clarified. Additional studies are needed to determine whether patients with insulin-dependent diabetes mellitus and those with non-insulin-dependent diabetes mellitus differ in their risk for primary liver cancer or whether the risk is affected by the type of diabetes treatment. [J Natl Cancer Inst 1996;88:1472-77]

Background: Although cigarette smoking has consistently been shown to be positively related to the risk of adenomatous polyp development (benign neoplastic growth of epithelial tissue in the colon)? most studies of cigarette smoking and the risk of colorectal cancer have been negative. However, in two large prospective studies in women and men, a statistically significant association between cigarette smoking and an increased risk of colorectal cancer was found, but only after more than 35 years of smoking. Purpose: To shed further light on the alleged relationship between long-term smoking and colorectal cancer risk, we performed a retrospective cohort study among Swedish construction workers, with many long-term smokers, complete long-term follow-up, and a large number of observed cases. Methods: We analyzed the association of smoking with colon cancer and with rectal cancer, using data on a cohort of approximately 135 000 male construction workers. High-quality exposure information was collected with the use of a comprehensive questionnaire filled out at the time of enrollment in the cohort, from 1971 through 1975. Complete follow-up was achieved through 1991 and the subjects were observed for an average of 17.6 years, thereby contributing approximately 2 375 000 person-years of follow-up. We calculated age-adjusted rate ratios (RRs) with the use of Poisson-based multiplicative multivariate models followed by further multivariate modeling that adjusted for other variables. Results: A total of 713 incident colon cancers and 505 rectal cancers were observed. There was no statistically significant association between current smoking status, number of cigarettes smoked or number of years smoking, and risk of colorectal cancer. The age-adjusted RRs were 0.98 (95% confidence interval [CI] = 0.82-1.17) and 1.16 (95% CI = 0.94-1.44) for colon and rectal cancers, respectively, among current smokers, and 1.07 (95% CI = 0.63-1.82) and 1.08 (95% CI = 0.58-2.03) among smokers of 25 or more cigarettes per day, relative to nonsmokers. Among smokers for more than 30 years at the start of follow-up, the age-adjusted RRs were 1.03 (95% CI = 0.85-1.25) and 1.21 (95% CI = 0.96-1.53) for colon and rectal cancers, respectively, relative to nonsmokers. Heavy smokers of cigars and pipes had a statistically nonsignificant tendency toward excess risk for colon cancer, but there was no clear dose-risk trend. Conclusion: Our large cohort study did not indicate any excess risk of colon cancer in males who were long-term heavy smokers and provided only weak support for an association with rectal cancer. Our data are thus consistent with the majority of previous reports. The reasons for the discrepancies in comparison with recent U.S. data have yet to be identified. [J Natl Cancer Inst 1996;88:1302-7]

Background: Testicular cancer incidence peaks during the third decade of life, and an increasing trend in the occurrence of this disease has been seen in many countries. Few risk factors have been established for testicular cancer, but evidence suggests that causal factors operate early in life, perhaps even in utero. Purpose: We performed a case-control study to evaluate specific perinatal characteristics as risk factors for the two main histopathologic types of testicular cancer: seminomas and nonseminomas. Methods: Two hundred thirty-two case patients with invasive testicular cancer and 904 individually matched control subjects (all singleton births), nested in a cohort of men born at 10 hospitals in Sweden from 1920 through 1978, were included in the study. Perinatal information about the study subjects and their parents was obtained from birth records. For the mothers, information was recorded concerning age at menarche, marital status, parity, age at delivery, maternal weight at delivery and at first visit to a maternal-care center, education and/or profession, and any medical problems encountered during pregnancy or after delivery. For the fathers, information was obtained concerning age at delivery and education and/or profession. For the study subjects, information was recorded about the following: gestational age, birth length and weight, placental weight, method of delivery (normal, cesarean section, or with forceps or vacuum extractor), medical problems during hospital stay, presence of jaundice, method of feeding at discharge, and duration of hospital stay after birth. Individuals diagnosed with testicular cancer were identified through the Swedish National Cancer Registry and the Uppsala Regional Cancer Registry. The data were analyzed by use of conditional logistic regression modeling. Results: When testicular cancer was modeled as a single entity, significantly elevated risks were associated with neonatal jaundice (adjusted odds ratio [OR] = 2.30; 95% confidence interval [CI] = 1.07-4.94) and with either low (

Background: For unknown reasons, the age-standardized incidence of testicular cancer has shown a rapid increase in virtually all countries (mostly Western) studied. For populations with a sufficiently long period of cancer registration, this development can be traced back to the first half of this century. Purpose: By evaluating data from six countries with long periods of cancer registration (Denmark, Norway, Sweden, the former German Democratic Republic [East Germany], Finland, and Poland), we sought to determine whether the increase in testicular cancer risk follows a birth cohort pattern and, if so, to quantify and compare any birth cohort effects. Methods: A total of 30 908 incident cases of testicular cancer, diagnosed from 1945 through 1989 in men who were 20-84 years of age, were identified in population-based cancer registries in the six countries. In addition to performing simple trend analyses, we fitted several Poisson regression models (with the explanatory variables age, time period [calendar time], and birth cohort) to the data. Individual models were estimated by the maximum likelihood method. Results: The age-standardized incidence of testicular cancer was found to vary among the six populations and, on the basis of total registration data, increased annually at rates ranging from 2.3% (in Sweden) to 5.2% (in East Germany). A comparison of several regression models indicated that birth cohort was a stronger determinant of testicular cancer risk than was calendar time for all six populations. Within each population, little variation in testicular cancer risk was observed for men born between 1880 and 1920; thereafter, the risk began to increase. Among men born in Denmark, Norway, and Sweden between 1930 and 1945 (the period encompassing the Second World War), the increasing trend in risk was interrupted (i.e., a leveling in risk occurred). After 1945, an uninterrupted increase in risk was observed for all six populations. With men born around 1905 as the reference group, the relative risk of testicular cancer for those born around 1965 varied from 3.9 (95% confidence interval [CI] = 2.7-5.6) in Sweden to 11.4 (95% CI = 8.3-15.5) in East Germany. Conclusions and Implications: The increasing trend in testicular cancer risk observed for these six populations follows a birth cohort pattern. This distinct risk pattern provides a framework for the identification of specific etiologic factors.

Background.- Because the incidence of childhood acute lymphatic leukemia peaks between 2 and 4 years of age, the risk factors may exert their influence during the prenatal and/or the neonatal periods. Results of previous studies of perinatal risk factors have been contradictory, perhaps because most studies either have been hospital based or have been restricted to limited geographical areas. Purpose: A nationwide case-control study was carried out to identify maternal and perinatal risk factors for this disease. Methods: The case-control study was nested in cohorts defined by all live births in Sweden recorded in the nationwide Medical Birth Register. Since 1973, this register has routinely collected information on all hospital births in regard to maternal demographic data, reproductive history, pregnancy, delivery, and the neonatal period. From the Swedish National Cancer Register, 613 case subjects were identified in successive birth cohorts from 1973 through 1989. Five control subjects per case subject were randomly selected from the pool of children matched by sex and month and year of birth. Conditional logistic regression was used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs) for potential risk factors and to estimate their effects after adjustment for possible confounders. Results: Risk of childhood lymphatic leukemia at all ages increased with Down's syndrome (OR = 20.0; 95% CI = 4.2-94.2), maternal renal disease (OR = 4.4; 95% CI = 1.6-12.1), use of supplementary oxygen (OR = 2.3; 95% CI = 1.5-3.6), postpartum asphyxia (OR = 1.8; 95% CI = 1.2-2.6), birth weight of more than 4500 g (OR = 1.7; 95% CI = 1.1-2.7), and hypertensive disease during pregnancy (OR = 1.4; 95% CI = 1.0-1.9). Down's syndrome affected risk mostly in children younger than 5 years, whereas other factors affected those children 5 years old or older. Being one of a multiple birth also increased risk among older children (OR = 2.5; 95% CI = 1.0-6.0). Use of supplementary oxygen may act as a causal intermediate (surrogate) for postpartum asphyxia and its causes, as would high birth weight for its causes. Conclusions: Several maternal and perinatal risk factors were found to be associated with childhood lymphatic leukemia, but they showed age-specific differences. Overall, only a few risk factors were identified, and these accounted for a small proportion of cases. We concluded that most risk factors for childhood lymphatic leukemia remain unidentified in very young children.

Background: Joint replacement with metal implants has been the standard procedure for surgical treatment of irreversible degeneration of hip and knee joints for more than two decades. However, reports of local malignancy after joint replacement and experimental studies that suggest a carcinogenic action of metal ions and polymethylmethacry-late (an acrylic compound used to stabilize the implant in the host) have raised concern about the possible long-term risks associated with metal implants. Purpose: Our aim was to study cancer risk in a Swedish cohort of patients who had hip replacement surgery during the period 1965 through 1983. Methods: We studied the risk of cancer in a cohort of 39 154 patients (14 869 men and 24 285 women), identified in the nationwide Swedish Inpatient Register with at least one hip replacement during the period 1965 through 1983. The patients were followed through 1989 by means of record linkage to the Swedish Cancer Register. The cohort contributed a total of 327 922 person-years at risk. Standardized incidence ratios (SIRs) were computed using age-, sex-, and period-specific incidence rates derived from the entire Swedish population. Results: The overall relative risk of cancer was increased by only 3%. Bone cancer--the focus of previous concerns--occurred in six cases versus 4.3 expected, and connective tissue cancer occurred in 28 cases versus 25.9 expected. Increased risks were observed for kidney cancer (SIR = 1.31; 95% confidence interval [CI] = 1.13-1.51), prostate cancer (SIR = 1.13; 95% CI = 1.04-1.22), and melanoma (SIR = 1.23; 95% CI = 1.00-1.50). The relative risk of gastric cancer steadily declined with increasing follow-up time, in both men and women (SIR = 0.58; 95% CI = 0.39-0.84 more than 10 years after hip replacement). Conclusion: In this study, the largest study to date to evaluate hip replacement and subsequent cancer risk, the overall cancer risk appears to be negligible from a public health perspective, and our results have not produced any strong evidence against the continued use of these devices. Nevertheless, the small but statistically significant increases in kidney and prostate cancers and the decrease in gastric cancer deserve further study. [J Natl Cancer Inst 87:28-33, 1995]

Background: To evaluate hypotheses about the relationship between immune alterations and cancer, several investigators have determined cancer incidence in groups of patients with rheumatoid arthritis (RA), a chronic autoimmune disease. The primary finding has been an increased risk of hematopoietic cancers. Purpose: In this study, we have attempted to refine estimates of the association between RA and subsequent development of specific cancers. Methods: We investigated site-specific cancer risk associated with RA in a population-based cohort study of 11683 Swedish men and women with a hospital (inpatient) diagnosis of RA. These case patients were identified from 1965 to 1983 and had follow-up through 1984 by computer linkage of the Swedish Hospital Inpatient Register to the National Swedish Cancer Registry (840 case patients with cancer) and the Swedish Registry of Causes of Death. Cancer risk was estimated by standardized incidence ratios (SIRs) for specific cancers. Results: For men and women overall, there were decreased risks for cancers of the colon (SIR = 0.63; 95% confidence interval [CI] = 0.5-0.9), rectum (SIR = 0.72; 95% CI = 0.5-1.1), and stomach (SIR = 0.63; 95% CI = 0.5-0.9) and an increased risk for lymphomas (SIR = 1.98; 95% CI = 1.5-2.6). Conclusions: The reduced risk for colorectal cancer in patients with RA is consistent with previous studies of RA patients and with reports which state that use of nonsteroidal anti-inflammatory drugs may protect against the development of large bowel cancers. The excess of lymphomas also confirms a number of earlier investigations of RA patients. [J Natl Cancer Inst 85:307-311, 1993]

Background: Little is known about the etiology of cancer of the exocrine portion of the pancreas, which produces a variety of digestive enzymes. Smoking, certain dietary factors, and diabetes mellitus are considered to be risk factors, although the risk estimates are modest in most instances. A recent cohort study of patients with chronic pancreatitis indicated a ninefold to 16-fold increased risk for pancreatic cancer. Purpose: Our purpose was to evaluate the relationship between various clinical types of pancreatitis and pancreatic cancer. Methods: Data for this study were collected from all inpatient medical institutions in Sweden from 1965 until 1983 by the Swedish National Board of Health and Welfare. Data were recorded on individual hospital admissions and discharges in the Inpatient Register. All patients with records in the Inpatient Register coded for acute, chronic, or unspecified pancreatitis were considered for inclusion in the study. A population-based cohort of 7956 patients with at least one discharge diagnosis of pancreatitis was monitored (up to 19 years of follow-up) for the occurrence of pancreatic cancer by record linkages to the Swedish Cancer Registry and Registry of Causes of Death. Results: A total of 46 pancreatic cancers were diagnosed during follow-up compared with 21 expected (standardized incidence ratio [SIR] of 2.2; 95% confidence interval [CI] 1.6-2.9) for the Uppsala Health Care Region. The excess risk for women and men was similar--most pronounced during the first period of follow-up (2-4 years) after discharge and close to unity after more than 10 years of follow-up. Patients with chronic pancreatitis and patients with more than one discharge diagnosis of either acute or unspecified pancreatitis were at higher risk (SIR = 3.8; 95% CI 1.4-8.2 and SIR = 4.8; 95% CI 1.9-9.9, respectively) compared with those with only one discharge of acute (SIR = 1.6; 95% CI 0.9-2.7) or unspecified (SIR = 2.1; 95% CI 1.2-3.2) pancreatitis. Conclusions: Our finding of a moderate excess of pancreatic cancer among patients with pancreatitis, especially the chronic or recurrent forms, supports some earlier clinical and case-control studies, but it is not consistent with the ninefold to 16-fold risk reported in a recent cohort study. The absence of an increased risk 10 years or more after first discharge for pancreatitis argues against a straightforward causal relationship. Because of the relatively short interval between diagnosis of pancreatitis and pancreatic cancer, it is possible that some forms of pancreatitis are a precursor to pancreatic cancer or that shared risk factors for both diseases (e.g., cigarette smoking) may also be involved. [J Natl Cancer Inst 86:625-627, 1994]