Screening for Lynch Syndrome in the General Population–Letter

The cost effectiveness analysis (CEA) by Dinh and colleagues offers the intriguing approach of screening all individuals 25 and older for Lynch syndrome (LS), using PREMM_1,2,6 tool followed by mutation testing (1).

This report must be interpreted with careful attention to certain aspects of the economic evaluation of public health screening programs. As Kastrinos and colleagues noted, patient data for PREMM_1,2,6 were derived from healthcare professionals that had already ordered genetic testing (2). Replication in a primary care office would likely be challenging. Furthermore, PREMM_1,2,6 was developed using apopulation with personal or family history suggestive of LS (2). Even in external validation of PREMM_1,2,6 (and PREMM_1,2), the population studied had personal history of CRC (2, 3). Therefore, generalization of the PREMM_1,2,6 validation study results to the general population may not be appropriate.

In the external validation of PREMM_1,2,6 the cohort size, LS prevalence, sensitivity, specificity, and positive predictive value (PPV) were 1,214, 6.7%, 90%, 67%, and 16.5%, respectively (2). For PREMM_1,2 these were 1,222, 6.6%, 100%, 68%, and 18%, respectively (3). Using the PREMM_1,2,6 numbers in a general population cohort of 100,000 with an LS prevalence of 1 of 440, the PPV will be 0.62%. For PREMM_1,2 the PPV will be 0.7%. This PPV is low, and we propose that the high number of false positives (≈33,000) necessitates that the cost effectiveness analysis include the disutility (negative impact on quality of life) associated with unnecessary genetic testing. We also propose that the cost of administering the PREMM_1,2,6 be considered, given that 440 individuals must be screened for each LS diagnosis.