Children with cystic fibrosis (CF) diagnosed by neonatal screening have a better nutritional development and other advantages compared with those in a nonscreened group. The two-tier immunoreactive trypsinogen (IRT)/DNA screening protocol has been found superior to the single-tier IRT approach, improving the positive predictive value and thus reducing the false-positive rate. However, variations of the DNA test are required for different populations. In this study we examined CFTR (cystic fibrosis transmembrane conductance regulator) mutations in 331 CF patients attending the centres in Stockholm, Lund and Uppsala, comprising about 75% of the CF population in Sweden. The frequency of deltaF508 among CF alleles was 68.3%. There were two other mutations, 394delTT and 3659delC, found to be fairly frequent, amounting to 8.5 and 7.9%, respectively. Other mutations were comparatively rare. A simple and effective method of analysing the three mutations from Guthrie cards has been developed. Assuming Hardy-Weinberg equilibrium, 90% of our CF patients will be expected to carry at least one deltaF508 allele and 97.6% to carry at least one deltaF508, 394delTT or 3659delC copy. Including the latter two in a screening programme would thus substantially reduce the risk of a false-negative outcome.