Background: The cancer family history is important in identifying individuals with hereditary non-polyposis colorectal cancer (HNPCC). The accuracy of a suspected HNPCC family history reported by patients with colorectal cancer was evaluated.
Methods: This was a prospective population-based study including consecutive patients with colorectal cancer. A questionnaire covering the occurrence of malignancy among relatives was completed.
Results: A total of 1200 patients with colorectal cancer completed the questionnaire. Fulfilment of Amsterdam criteria I or II according to the patients' reports was rejected in three of 14 cases (false-positive rate 21 per cent). Furthermore, seven of 18 probands whose families met the Amsterdam criteria I or II after verification were identified by further exploration in families who, according to the probands, met weaker criteria (false-negative rate 39 per cent).
Conclusion: The present study suggests that family studies on HNPCC are not reliable unless the diagnoses of family members are verified from official sources. If endoscopic screening is offered entirely on the basis of unverified information from patients with colorectal cancer, there is a risk that a large proportion of the families will not be offered relevant surveillance.