Summary Survival from colorectal cancer has been analysed in relation to marital status in a nationwide Danish study of 9596 patients with complete follow-up of 22-26 years. After exclusion of 2294 patients with missing information adjusted five-year survival among married patients was significantly longer (RR=0.85; 95% CI 0.78-0.93).Keywords: Colorectal cancer; marital status; survival; epidemiology Indications of an association between marital status and health were first reported long ago (Ramazzini, 1965;Stavola 1987;Farr, 1859). An extensive literature showed subsequently that married persons have lower mortality rates from a variety of diseases (Blazer, 1982;House et al., 1982;Schoenbach et al., 1986). In addition, marital status has been claimed to be an independent prognostic factor in survival from cancer (Neale et al., 1986;Goodwin et al., 1987). Owing to its high incidence and relatively poor prognosis, colorectal cancer is a major public health problem in most affluent industrial countries, including Denmark (Johansen et al., 1993). We studied the influence of marital status on survival from colorectal cancer in a nationwide cohort with complete long-term follow-up.
Materials and methodsThe present study concerns Danish patients recorded in the Danish Cancer Registry with adenocarcinoma of the colon or rectum diagnosed between 1 April 1968 and 31 March 1972. These patients numbered 9596. Of these, 2182 patients were excluded for at least one of the following reasons: a previous cancer (691 patients, 7.2%); information was missing: on extent of the disease (1112 patients, 6.8%); on the anatomical subsite of the tumour, mainly notified as colon unspecified (657 patients, 6.8%); or the diagnosis was given only on a death record (25 patients, 0.3%). Information on marital status at the time of diagnosis was sought in the Central Population Register (CPR), municipality population registers, probate court records, parish registers and regional administration archives, as reported earlier (Mellemgaard et al., 1989). We also excluded 112 patients (1.2%) whose marital status could not be identified in one of the above-mentioned ways, leaving 7302 persons for the survival study. Updating of the Danish Cancer Register resulted in minor differences in the number of patients including in this and a previous study due to late registration in the register (Mellemgaard et al., 1989). Cases were followed-up by linkage to the CPR before 18 August 1994 for dates of death and emigration and for verification of each personal identification number, which is a unique ten digit identification of every Danish resident.Analyses of survival after colorectal cancer The multivariate analyses included sex, age at diagnosis in the age groups 0-49, 50-59, 60-69, 70-79, 80 -89, and >90 and year of diagnosis (1968, 1969, 1970, 1971, 1972). The anatomical site of the tumour was divided into the categories of right colon, left colon and rectum as previously described (Johansen et al., 1993). The extent of disease at diagnosis was...