PRIMARY BRAIN TUMOURS AS SECOND PRIMARY: A NOVELASSOCIATION BETWEEN MENINGIOMA AND COLORECTAL CANCER Beatrice MALMER*, Bjo ¨ rn TAVELIN, Roger HENRIKSSON and Henrik GRO ¨ NBERG Department of Oncology, Umea ˚ University, Umea ˚ , Sweden In a previous study, a decreased risk for first degree relatives of patients with astrocytoma has been observed for breast and colorectal cancer. The aim of this study was to examine the associations between breast and colorectal cancer as first primary cancer and the risk of developing astrocytoma and meningioma as a second primary cancer. Two cohorts were constructed, one from all cases with breast cancer (1,036,466 person-years) and one from all cases of colorectal cancer (572,422 person-years) in Sweden during the period 1958–1994. The risk of developing astrocytoma and meningioma after breast or colorectal cancer was calcu- lated. A significant increased risk for developing meningioma was seen after colorectal cancer, standardized incidence ratio (SIR) 1.60 confidence interval (CI; 1.32–1.94) and after breast cancer, SIR 1.57 CI (1.36–1.81). The previously observed decreased risk for astrocytoma could not be verified in this study. A novel association between meningioma and colorec- tal cancer, particularly in females, was observed, which justifies further studies to evaluate common aetiological factors. Int. J. Cancer 85:78–81, 2000.  2000 Wiley-Liss, Inc. The aetiology of primary brain tumours (PBT) is a controversial issue, and the only somewhat well-established risk factors for meningioma and astrocytoma are ionising radiations (Karlsson et al., 1998). Family members have an increased risk of meningioma and astrocytoma in some hereditary syndromes, e.g., Turcot syndrome with an association between non-polyoposis colon cancer and astrocytoma, where the mismatch repair gene (MMR) is the disease-causing gene (Paraf et al., 1997). In meningioma, an association with breast cancer has been reported in several studies (Bondy and Ligon, 1996). A relationship with hormonal factors has been proposed since meningiomas frequently express progesterone receptors (Carroll et al., 1993). In contrast with astrocytoma, females suffer more often than males from meningioma. In a previous cohort study, we found a significantly decreased risk in first degree relatives (FDR) of astrocytoma patients to develop colorectal and breast cancer (Malmer et al., 1999). Therefore, a nation-wide cohort study was performed in order to verify this negative association, investigating the risk of developing astrocytoma and meningioma after a primary colorectal or breast cancer. Our study also aimed at raising hypotheses on the aetiology of these PBT. The associations of astrocytoma and meningioma were calculated separately in our study since different PBT have varying gender and age-specific incidences indicating different aetiologies. MATERIAL AND METHODS According to the Swedish legislation, all malignancies must be reported both by the clinician and the pathology department to the National Cancer Registry. Not only all malignant neoplasms are reported, but also all benign tumours in the central nervous system and meninges. The nation-wide Swedish Cancer Registry was used as our study base. Two cohorts were constructed, one from the 99,640 patients with primary colon cancers and 57,232 primary rectal cancers (ICD7 code 153 and 154) (572,422 person-years) and one from the 154,414 patients with female breast cancer (ICD7 code 170) (1,036,466 person-years) recorded from 1958 to 1994 in the Swedish Cancer Registry. A cut-off of 6 months from the primary cancer diagnostic was used to avoid an excess of tumours detected because of careful investigations when diagnosing the first primary cancer. The same cut-off has been used in other studies (Ahsan et al., 1995) and is therefore also suitable for comparisons. These cohorts were also divided according to the median age of diagnosing the primary breast or colorectal cancer, 64 and 72 years, respectively. ICD7 code 193 for PBT was divided, calculating meningioma (pathology code 461) and astrocytoma (pathology code 475 and 476) separately. Statistical methods Person-years were calculated from 6 months after the first primary cancer to the date of death, or to December 31, 1994, whichever came first, using the program PYRS (Coleman, 1989). The expected number of cases was calculated by multiplying the cancer-specific incidence rate for Sweden by calendar and age- specific person-years. The cancer incidence for Sweden for the period 1958–1994 was obtained from the Swedish Cancer Registry and used in the calculations. The standardised incidence ratio (SIR) was defined as the ratio between the observed and the expected number of cases. Exact confidence interval (CI) of the SIR was calculated by the formula of Byar (Breslow and Day, 1980). An exact CI was calculated to test the significance of the SIR. RESULTS A significant increased risk of developing PBT was seen after colorectal cancer SIR 1.28 CI (1.12–1.46) with 224 observed and 174 expected PBT (Table I). The elevated risk was restricted to meningioma, SIR 1.60 CI (1.32–1.94), and not to astrocytoma, SIR 1.07 CI (0.82–1.37). The major part of the risk for developing meningioma was among females with colorectal cancer, SIR 1.85 CI (1.46–2.31), and not among males, SIR 1.20 CI (0.82–1.70). The risk was significantly increased in females both younger and older than the median age 72 years (Table I). The mean time in months between colorectal diagnostic to meningioma diagnostic was 81 months (Fig. 1). A majority of the meningiomas were diagnosed in a 5-year period after the colorectal cancer diagnosis. The mean age of developing meningioma was 77 years in this cohort. The basis of diagnosis of meningioma and astrocytoma after colorectal cancer is summarised in Table II. Meningioma was generally more often diagnosed by autopsy than high-grade astrocytoma. An increased risk of all PBT was also seen in females after a primary breast cancer (Table III), with 367 observed and 295 expected cases, SIR 1.24 CI (1.12–1.38). The risk was restricted to meningioma, SIR 1.57 (1.36–1.81), and not to astrocytoma, SIR 1.00 CI (0.80–1.23). There were no major differences in distribu- tion of the risk ratios of meningioma when comparing the group above and beneath the median age (64 years) of developing the primary breast tumour (Table III). The mean time interval between the diagnosis of breast cancer and meningioma was 99.5 months Grant sponsors: Lion’s Cancer Foundation and the Department of Oncology, Umea ˚ University. *Correspondence to: Department of Oncology, Umea ˚ University Hospi- tal, 901 87 Umea ˚, Sweden. Fax: +46–7852031. E-mail: beatrice.malmer@onkologi.umu.se Received 14 May 1999; Revised 5 July 1999 Int. J. Cancer: 85, 78–81 (2000)  2000 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer