Development of sensitive and specific age- and gender-specific low-density lipoprotein cholesterol cutoffs for diagnosis of first-degree relatives with familial hypercholesterolaemia in cascade testing

Background: The plasma total and low-density lipoprotein-cholesterol (LDL-C) levels that are used as diagnostic criteria for familial hypercholesterolaemia (FH) probands in the general population are too stringent for use in relatives, given the higher prior probability of a first-degree relative being FH (50% vs. 1/500). Our objective was therefore to develop more appropriate LDL-C cutoffs to identify "affected" first-degree relatives found by cascade testing, to test their accuracy and utility in case identification, and to compare them with the published "Make early diagnosis to prevent disease" (MEDPED) cutoffs from the US. Methods: Using a large, anonymised sample of genetically tested first-degree relatives of Netherlands FH probands (mutation carriers/non-carriers, n=825/2469), age- and gender-specific LDL-C diagnostic cutoffs for first-degree relatives were constructed. These were used to test similar data from Denmark (n = 160/161) and Norway (n=374/742). Results: Gender-specific LDL-C diagnostic cutoffs were established for six different age groups, which achieved an overall accuracy (measured as Youden's index) of 0.53 in the Netherlands data, and performed significantly better amongst younger (< 25 years) compared to older first-degree relatives (0.68 vs. 0.42 Youden's index, p < 0.001). Compared with the Netherlands data, age- and gender-adjusted mean LDL-C levels were significantly higher (approximately 0.5 mmol/L) in the Denmark and Norway subjects for both mutation carriers and non-carriers. After adjusting for this difference, the LDL-C cut-offs showed a similar accuracy in identifying mutation carriers from Denmark (81%, range 78%-86%) and Norway (84%, range 82%-86%). Although the MEDPED cutoffs performed significantly worse than these for the Netherlands data (p < 0.001), they performed equally well in overall accuracy for the Norwegian and Danish data, although the LDL-C cutoffs had a significantly higher sensitivity but lower specificity for all three countries. Conclusions: The cutoffs developed here are designed to give the greatest overall accuracy when testing relatives of FH patients in the absence of a genetic diagnosis. They have a more balanced specificity and sensitivity than the MEDPED cutoffs that are designed to achieve higher specificity, which is more appropriate for cascade testing purposes. The data suggest that country-specific LDL-C cutoffs may lead to greater accuracy for identifying FH patients, but should be used with caution and only when a genetic diagnosis (DNA) is not available.