In this study, we have determined the mutation panorama in a Swedish cohort referred for genetic LQTS testing as part of ordinary health care. Between March 2006 and October 2009, the department of Clinical Genetics in Umeå was to our knowledge the only laboratory in Sweden screening the LQTS genes. Among the 200 index patients, 64 different mutations were identified in 103 patients (52%); of which 58% occurred in KCNQ1, 24% in KCNH2, 13% in SCN5A, 3% in RYR2, 1% in KCNE1, and 1% in KCNE2. Thirteen of the mutations were found in more than one family, whereas 51 occurred only once. Among these mutations, 28% were novel at the time of detection, and had thus never been reported previously.
LQTS founder mutations
Two of the recurring mutations, KCNQ1 p.Y111C and KCNQ1 p.R518*, were identified in 26 of the 103 cases, thus accounting for approximately 25% of the mutations in the Swedish LQTS population. We have recently shown that family members carrying these mutations share a common haplotype that is specific for each mutation
 [Abstract number 154:Winbo A. Stattin E.L. Nordin C. Persson J. Diamant U.B. Jensen S.M. Rydberg A. Origin, genotype and clinical phenotype of the Long QT Syndrome R518X/KCNQ1 mutation in Sweden. Presented at the 46th Annual Meeting of the Association for European Paediatric and Congenital Cardiology (AEPC), May 23–26 2012 in Istanbul]. The mutation KCNQ1 p.Y111C was introduced and enriched in the Ångerman River valley approximately 600 years ago
. Functional in-vitro studies have demonstrated that it is a “malignant” mutation with a strong dominant-negative effect, causing disturbed function of the wild-type ion channel
[27, 28]. In contrast to these findings, we recently showed that the KCNQ1 p.Y111C mutation presents with a low incidence of life threatening events in a Swedish Y111C-positive LQTS population
. Furthermore, we showed that p.Y111C is a founder mutation in this population
, a finding which also contrasts to in vitro-data indicating it to be a malignant mutation. One explanation for these discrepancies could be the presence of population-specific modifiers, genetic or other, such as the recently described polymorphisms in the 3’-UTR of KCNQ1, mitigating the effect of the mutated allele by reduced expression
. Possibly, one or several of these polymorphisms could exert its attenuating effect through the creation of a novel miRNA-binding site, a theory that has been proposed for other disorders where large differences in phenotypic expression occur
The mutation KCNQ1 p.R518* has previously been reported in several populations
[32, 33], as well as a founder mutation in Sweden
 and Norway
, although Berge et al. did not report any founder mutations in the more recent Norwegian LQTS population survey
. A strong founder effect has been described in the Finnish population
. In our study, we identified three of the Finnish founder mutations (KCNQ1 p.G589D, KCNH2 p.R176W, and KCNH2 p.L552S), as well as the two common Norwegian mutations (KCNQ1 p.R518* mentioned above, and KCNQ1 p.Q530*) in several of the patients
LQTS genotype-negative index cases
According to published studies, approximately 25% of index cases with the clinical phenotype of LQTS remain genotype-negative after comprehensive assessment of the three most common LQTS genes (KCNQ1, KCNH2, and SCN5A)
[12, 13, 15]. In this study, 102 index patients (51%) referred for LQTS testing were negative after sequencing of the KCNQ1, KCNH2, KCNE1, KCNE2 and SCN5A genes. As in any molecular genetic study of disease, there is a possibility that these individuals have mutations missed due to technical limitations (e.g. DHPLC), or located in regions not included in the analysis; such as in the gene promotors or introns of the genes chosen for study, or in another LQTS-associated gene. However, these patients had a significantly shorter QTc, and reported family history than the mutation carriers, and some of them might therefore be suspected of not having LQTS (Figure
1). Several publications indicate that patients suspected of having LQTS may actually have CPVT,
[8, 9, 36] and that can be confirmed also in the present study. Among 36 genotype-negative index patients selected for RYR2 screening based on a history of arrhythmia, aborted cardiac arrest and/or syncope and/or a family history of SCD, we identified a disease-causing mutation in 8%. Tester et al. evaluated the prevalence of RYR2 mutations in a cohort of patients referred for screening of LQTS genes, identifying mutations in RYR2 among 6% of the 269 genotype-negative patients
. Berge et al. identified mutations in RYR2 in 17% of the 41 genotype-negative index patients referred for LQTS testing
. Thus, it is critical to recognise CPVT as an important differential diagnosis to LQTS, and to consider mutation screening of the RYR2 gene in patients who do not have a mutation in one of the LQTS-associated genes.
The presence of copy number variants (CNVs) within the LQTS disease genes have also been suggested as an explanation for the lack of identified mutations
[37–39]. Therefore, we performed MLPA analysis of all 200 index patients, identifying 2 CNVs in the KCNH2 gene. Thus, the yield of CNVs was 2.0% among the 100 genotype-negative index patients without an identified mutation in any of the LQTS genes or RYR2. In the study of Tester et al. CNVs were found in 4.8% of 42 patients with QTc duration ≥ 480 ms and/or a Schwartz score ≥ 4 who were negative for mutations in 12 of the LQTS-associated genes
. Eddy et al. identified CNVs in 11.5% of 26 patients with Schwartz score ≥ 4 who were negative for mutations in the KCNQ1, KCNH2, and SCN5A genes
. Barc et al. identified CNVs in 3.2% of 93 patients with Schwartz score ≥ 3 who were negative for mutations in the KCNQ1, KCNH2, and SCN5A genes
. These findings suggest that CNVs might be a more frequent cause of LQTS than mutations in all of the less common LQTS-associated genes (ANKB, KCNE1, KCNE2, KCNJ2, CACNA1C, CAV3, SCN4B, AKAP9, and SNTA1) together
[38–40]. Thus, it is important to consider MLPA analysis in patients who do not possess a mutation in one of the most common LQTS-associated genes.
Genetic cascade screening in family members
A total of 481 relatives in the 103 families with an identified mutation have participated in genetic cascade screening, of which 41% were found to be mutation carriers, and 59% were not carriers. Thus, 2.9 (302/103) individuals per family carried a mutation and were thereby at risk for LQTS-associated symptoms and SCD. This finding is lower than the result in Norway, where, 4.7 (305/66) patients per family carried a heterozygote mutation
. In contrast to Imboden et al.
 no female predominance among mutation carriers and no non-random inheritance, with a significant greater number of affected than expected, could be observed in this cohort.
This study compared with other population surveys
In the five largest LQTS population surveys that have been published to date, involving the five most common LQTS-causing genes (KCNQ1, KCNH2, SCN5A, KCNE1 and KCNE2), the mutation yield was 72%, 51%, 50%, 36%, and 32%, respectively (Table
[8, 12–15]. In two of the studies, when using more stringent criteria (i.e. Schwartz score ≥ 4), the mutation detection rate was raised from 50% to 72%, and from 32% to 71%, respectively
[8, 13]. In this study, we obtained a mutation detection rate of 52%, which lies in the range of the other population studies. Among individuals with a definite prolonged QTc, 77% carried a mutation, which is in line with the two studies using more stringent criteria. We were not able to categorise all the index patients, since phenotypic information was not available for all of the patients.
The largest survey, including 2,500 consecutive, unrelated LQTS patients, presented one of the lowest mutation yields of 36%. However, the degree of diagnostic relevance in the referred patients of that study could not be evaluated, also due to lack of phenotypic information.
The mutations in the KCNQ1, KCNH2 and SCN5A genes were distributed over the entire coding regions and adjacent splice sites. The vast majority were heterozygous missense mutations. The distribution of mutations between the different genes and the type of mutation concur with findings of the other population surveys. However, the rate of KCNQ1 mutations is higher in our study (58%), since both of the Swedish founder mutations p.Y111C and p.R518* are located in this gene. Most of the mutations (≈60 %) in the published surveys have not been reported previously, whereas we only identified 28% novel mutations in this study. It is possible that this lower yield is due to the more than 10 years of publications of several large LQTS studies, which suggests that the increase in new LQTS mutations is beginning to be saturated.
Probands carrying multiple mutations
Patients carrying multiple mutations have been shown to present with a more severe phenotype compared to patients carrying only one mutation
. In one study, the compound mutation carriers had longer QTc intervals and a younger age-at-onset compared to patients with only one mutation
. In this study, four of the 103 (4%) genotype-positive patients carried more than one definitely pathogenic mutation. Two of them were homozygous for the mutation and two compound heterozygote. All of them are female and have a family history of LQTS. The (homozygous) carrier of the KCNQ1 c.1893dup mutation, had a QTc prolongation of 512 ms; however, there is no information about any history of syncope or current treatment with beta-blockers. The parents of the c.1893dup mutation carrier both carried the mutation in heterozygous form; the mother had a QTc of 435 ms while the father has a QTc denoted as normal (data not shown). The KCNE1 p.R32H (homozygous) mutation carrier, had a QTc of 447 ms, is being treated with beta-blockers and has not experienced syncope. The parents of the p.R32H mutation carrier were not available for testing, but hemizygosity of the mutation in the proband was excluded by MLPA (data not shown). The carrier of the (compound heterozygote) mutations KCNQ1 p.P73T and SCN5A p.V411M, had a history of suspected seizures and syncope, and she was not treated with beta-blockers but with Phenytoin. The carrier of the (compound heterozygote) mutations KCNQ1 p.Y111C and SCN5A p.A29V, had a history of presyncope and a QTc of 490 ms.
In the Norwegian survey, no patients had more than one definitely pathogenic mutation, whereas Kapplinger et al. reported 9%, and Tester et al. reported 11% patients with multiple mutations among the genotype-positive patients
[8, 13, 14]. In the study of Westenskow et al. compound mutations were reported in 12% of the genotype-positive LQTS probands. However, of the 20 probands in their study assigned as having multiple mutations, over half possessed either the KCNE1 p.D85N or KCNQ1 p.P448R common polymorphism as the “second hit”
. Similarly, Tester et al. reported SCN5A p.A572D as a mutation in 3 of their patients
. In our study, SCN5A p.A572D was identified in 3 patients, KCNQ1 p.P448R in 3 patients, and KCNE1 p.D85N in 12 patients, all of which were determined to be non-pathogenic (
Additional file 2). In four of the patients, these variants occurred together with a definitely pathogenic mutation. If we had regarded these as pathogenic mutations, our yield of multiple mutations among the genotype-positive patients would have been 8% (8/103) instead of 4%.
Sequence variants of unknown significance - polymorphisms
Missense mutations are the most frequent form of mutation in the LQTS genes, accounting for about 65-73% of the mutations in the large LQTS population surveys. Careful interpretation of identified genetic variants is important, because a missense variant may or may not cause an altered/distorted protein and a disease phenotype
. In this study, several of the index patients carried rare variants, such as KCNE1 p.D85N, KCNQ1 p.P448R, SCN5A p.A572D, SCN5A p.S1103Y, and SCN5A p.R1193Q. The possible effect of these missense variants is difficult to interpret and they are referred to in the literature as both mutations and functional polymorphisms
[44, 45]. Although these variants might contribute to the phenotype, we did not consider these as disease-causing mutations by themselves, since careful interpretation of genetic test results is critical in clinical practice
Limitations of the study
The 200 index cases were almost consecutively included; we have excluded 27 index cases with other diagnoses such as Jervell and Lange-Nielsen syndrome, Brugada Syndrome, short QT syndrome, and healthy individuals sent for LQTS-screening due to a history of first-degree relative with SUD. No selection of the patients was performed, resulting in a cohort that ranges from low suspicion of LQTS to high. Since the patients were referred for LQTS screening in ordinary health care, the clinical data were collected retrospectively and is thus not complete for all families. Only 23 of 105 exons (8–15, 44–50, 83, 88–105) of the gene RYR2 were analysed. DNA was not available for screening of all five genes in some of the individuals, and therefore it is possible that some double mutations might have been missed. Due to the lack of DNA, there were incomplete analyses of the KCNQ1 gene in three individuals, in the KCNH2 gene in two, KCNE1 in five, KCNE2 gene in four, and in the SCN5A gene in nine individuals. For the same reason MLPA was not performed in two cases.