1. The use of nasal preparations was strongly associated with cancer of the nasal sinuses (RR = 5-2; 95% CI= 19-13-4) (ICD-0 site codes 160.2-160.7; n=23 cases). We found a relative risk of 1-7 (95% CI = 0-5-5-5) for tumours of the nasal cavity (ICD-0 site code 160.0, n=24 cases);0-884 149 4) 0 6-7;and; for tumours of other sinonasal sites (ICD-0 site codes 160.1 and 160.8, n = 4 cases). An examination of the variation in relative risk by histologic type revealed the strongest association with adenocarcinoma (RR = 8&0; 95% CI = 1 -2-50-2); the relative risks were smaller for squamous cell,2-3,
2. An examination of total weeks of nasal preparation use (frequency x duration) revealed an almost ninefold increase in risk for subjects who had 520 or more total weeks of use. The magnitude of this risk reflects the close association between frequency and duration of use; subjects who reported using nasal preparations for many years also tended to use them for many weeks per year. When next-of-kin interviews were excluded from analysis, the RR associated with 520 or more total weeks of use was 7-0
3. Some aspects of this study require that caution be used in the interpretation of the results. We interviewed 32 sinonasal cancer cases, 18 next-of-kin of deceased cases, and three next-of-kin of living cases who were unable to participate. By contrast, all control interviews were conducted with living controls. To the extent that recall of nasal preparation use differed between cases and controls and between living cases and next-of-kin, the results we obtained may be biased. However, the exclusion of next-of-kin respondents from the analysis had little impact on the risk associated with 520 or more weeks of nasal preparation use
4. An association between smoking and sinonasal cancer has been previously reported. 12 14 Hayes et al13 noted a moderate increase in risk associated with smoking but did not report the impact of smoking on specific histologic types of tumours. Elwood14 observed an excess risk for all histologic types of sinonasal cancer, finding the strongest associations with transitional cell and squamous cell tumours
5. An association between the use of nasal The occurrence of sinonasal tumours has been preparations and an increased risk of sinonasal cancer related to direct inhalation of known or suspected was previously reported by Hayes et al,4 who found a carcinogens such as nickel, wood dust, and aromatic five-fold increased risk among men with a history of hydrocarbons.'6 17 Delayed mucociliary transport nasal spray use but no increase in risk among women may also permit carcinogens to remain in contact with with this exposure. By contrast, Brinton et al'2 found the sinonasal epithelium for an extended period of no association between regular use of nose drops and time, thus increasing the likelihood of malignant sinonasal cancer risk after controlling for a history of transformation