110 Heart 1997;78:1 10-116

REVIEW

The phenotype/genotype relation and the current status of genetic screening in hypertrophic cardiomyopathy, , and the long QT syndrome

J Burn, J Camm, M J Davies, L Peltonen, P J Schwartz, H Watkins

Hypertrophic cardiomyopathy, Marfan syn- encode amino acids, but within the gene there drome, and the long QT syndrome are all are also non-functional sequences (introns). autosomal disorders inherited in a dominant The simplest single gene abnormality is a manner; affected individuals are heterozygous, point mutation in which one base pair is that is they have one normal and one mutant replaced by another. Because of the redun- copy of the gene but not all the gene carriers dancy in the genetic code, point mutations are symptomatic. This failure to express fully may be silent polymorphisms that do not the expected phenotype is traditionally known change the amino acid or they may be mis- as incomplete penetrance. sense mutations that change a single amino acid. Missense mutations can be regarded as analogous to simple spelling errors in one let- Vocabulary and technology ter of a word. For example "the fish is on the In the human genome (consisting of 22 pairs dish" becomes "the fish is on the fish". of autosomes and the X and Y chromosomes) Because each gene exists in a maternal and a there are between 50 000 and 100 000 indi- paternal copy the ultimate product of the gene vidual genes coding for a wide range of prod- is a combination of normal (wild type) and ucts including enzymes and structural abnormal (mutant type). If the subject has proteins. Each chromosome is paired, giving a inherited an abnormal gene from both mother maternal and a paternal contribution to each and father (homozygous) the product is gene. Each gene consists of a series of base entirely of the mutant type. Some ofthe earliest pairs, the sequence of which is used as the recognised genetic disorders, such as sickle template for producing the messenger RNA, cell anaemia, followed this simple inheritance Clinical Genetics, which is used in turn to synthesise proteins on pattern, carriers (heterozygotes) having a mix- Northern Region the ribosomes in the cytoplasm. There are four ture of adult (A) and sickle (S) haemoglobin, Genetic Services, bases: adenine (A), guanine (G), cytosine (C), homozygotes only have haemoglobin S. Newcastle upon Tyne, UK and thymine (T). A occurs opposite T and C Other abnormalities of the gene structure J Burn opposite G in the two strands of the DNA include deletions of one or more base pairs. Cardiological double helix; one, the sense strand, is tran- Such deletions can have a major effect if any- Sciences, St George's scribed; the other (antisense) is not. The thing other than a multiple ofthree base pairs is Hospital Medical School, London SW17, sequences in each strand are complementary removed because the triplet will be read out of UK but run in opposite directions. The two copies frame, which will scramble the code thereafter. J Camm (alleles) of the gene in an individual from each By analogy, loss of one letter from the text M Davies J parent consist of a segment of the double sequence used above while the word length KTL DNA molecules. The base sequence same would make "Thf ishi so nt Mannergerminlia, stranded remained the Helsinki, Finland in the DNA strand is a triplet code in which bed isp". L Peltonen sets of three bases determine an amino acid. In Where an abnormally situated stop codon Instituto di Clin Med the messenger RNA molecule, thymine (T) is results from a point mutation or from a frame Medica Generale e replaced with uridine (U). Of the 64 possible shift translation of the mRNA product of this Terapia Medica, Universita degli Studi coding triplets which can be formed from the gene, a truncated protein product or no prod- di Milano, Milan, Italy four bases, some define gene boundaries: uct is synthesised. More recently yet another P J Schwartz TAG and TGA are termination signals, and class of mutation has been identified: triplet Cardiovascular AUG an initiation signal. The other triplets repeat sequences may be abnormally redupli- Medicine, John code for a specific amino acid, but there is cated, with marked increase in length of the Radcliffe Hospital, Oxford, UK considerable redundancy; there are more gene until it is ultimately inactivated. Such H Watkins triplet permutations than amino acids. repetition may progress from generation to Correspondence to: Genes vary widely in size from a few hun- generation so that the disease phenotype gets Professor M J Davies, BHF Cardiovascular Pathology dred to many thousand base pairs. The genes worse (anticipation). An example is myotonic Unit, Cranmer Terrace, are widely spaced and between them are base dystrophy. London SW17 ORE, UK. sequences with no known function. Within Individuals who are homozygous will have Accepted for publication each gene there are sequences which direct severe disease because either no gene product 6 May 1997 gene expression, and sequences (exons) which is produced or the product is all of the mutant Genetic screening in HCM, Marfan syndrome, and the long QT syndrome ill

type. Homozygotes require both parents to be of how the abnormal protein product produces abnormal gene carriers. This is rare unless the the phenotype. The whole picture is not yet clear parents are related or the frequency of the for hypertrophic cardiomyopathy, Marfan dis- mutant gene is very high in the general popu- ease, or the long QT syndrome, but major lation. With mutations of structural protein advances have been made and the three dimen- genes, disease may be manifest in hetero- sional construct ofphenotype, gene, and product zygotes because the presence of mutant function is proving somewhat different for each. product interferes with the function of the wild type product (termed a dominant Hypertrophic cardiomyopathy negative action). Point mutations in five separate genes are currently known to produce hypertrophic cardiomyopathy.'-3 The phenotype is charac- Tools in gene identification terised by left ventricular wall thickening with a A widely used technique for the initial identifica- characteristic histological appearance known tion of a previously unknown mutated disease as myocyte disarray,4 which separates the con- related gene is linkage analysis. When large fami- dition from left ventricular hypertrophy sec- lies are available in which there are two or more ondary to hypertension or aortic stenosis. generations having clinically affected and non- Disarray occurs at two levels-within the cell, affected individuals, study of DNA extracted myofibrillary structure is disorganised with from peripheral blood cells can lead to identifica- crossing rather than the normal parallel array; tion ofthe chromosome on which the gene is sit- the myocytes are also arranged abnormally in uated. This is based on the utilisation of the relation to each other, running in whorls polymorphic markers in the human genome. around foci of connective tissue. In areas with Over 5000 polymorphic non-functional DNA disarray the myocytes are misshapen with sequences (markers) are evenly spread on indi- incomplete aberrant cell junctions occurring vidual chromosomes and form the basis of the over much of the surface rather than being first generation map of the human genome. A confined to the intercalated discs at the ends known polymorphic marker which systematically of a roughly oblong shaped cell.5 segregates with the disease in the family flags the The current model of hypertrophic cardio- segment of a chromosome where the disease myopathy is that the known genes code for gene lies. The human genome project provides proteins concerned with the myofibril itself.6 rapidly increasing amounts of data on genes These are cardiac heavy chain (chro- residing in that area. However, there are often mosome 14), cardiac T (chromo- many of these; the products which such genes some 1), (chromosome 15), and code for can, however, be considered and a myosin binding protein C (chromosome 11). guess made about the most likely candidate gene The protein coded for by a gene on chromo- for a particular clinical picture. In the final stage some 7 is unknown but a phenotype indistin- DNA sequences are identified in which there is a guishable from the others is produced. The change in base sequence which is always inher- gene on chromosome 7 is tightly linked to pre- ited with the disease, is not present in unaffected excitation in a family. This implies either that individuals, and has an abnormal protein prod- the gene produces both hypertrophic cardio- uct which provides a reasonable explanation for myopathy and pre-excitation, or that there are the phenotype. two immediately adjacent genes which are Linkage analysis, while a powerful tool, rarely separated in meiosis. depends on having large families containing both Virtually all cases of hypertrophic cardiomy- affected and normal individuals. Several factors opathy are heterozygous and the gene is a may disturb what is ostensibly a logical approach dominant negative. This means that the that is certain of success. In the first place, very mutant myofibrillary protein interferes with accurate clinical establishment of whether any the function of the wild (normal) protein in individual is or is not carrying the gene is vital. forming properly aligned myofibrils within the Linkage becomes difficult if some family mem- myocyte. In reports on hypertrophic cardio- bers carry the gene but have no clinical manifes- myopathy wordwide, there are few examples tations. Second, key family members may have of proven new mutations in which neither par- died or moved to other countries, so that blood ent carried the abnormal gene. This means samples are not available; a final problem is that that in the majority of families with an apparent up to 5% of individuals in genetic studies can be sporadic case of hypertrophic cardiomyopathy shown not to have the father named on their there are likely to be other asymptomatic gene birth certificate. carriers. In the past, series which identified carriers by the clinical expression alone failed to detect many gene carriers. [ Heavy chain Genotype-phenotype relations myosin mutations account for around 35% of The understanding of genotype-phenotype rela- cases of hypertrophic cardiomyopathy and tions is proving more complex than originally over 40 separate point mutations in the gene envisaged. A three dimensional approach is are now known. Mutations in the other genes needed. One aspect is the clinical picture, that is, are less common, each accounting for 5-15% the phenotype. This involves the structural of cases. At least eight mutations are now changes observed and the nature and severity of described in cardiac troponin T, three in the clinical disease. The second aspect is the tropomyosin, and three, including two splice gene itself, its DNA sequence, and what protein it site mutations that lead to a frame shift in the codes. The last aspect involves an understanding gene, for myosin binding protein C. 112 Burn, Camm, Davies, Peltonen, Schwartz, Watkins

Outcome pathways for an GENE INHERITED AT BIRTH ening but the most likely explanation for individual carrying a interindividual variation is that other genes hypertrophic cardiomyopathy gene. (ECG abncDrmal) impinge on the phenotypic expression of the CCF, congestive cardiac hypertrophic cardiomyopathy gene. Any gene failure; LV, left that enhanced hypertrophy would be expected ventricular. Phenotype expressed in adolescence to potentiate the phenotypic expression of (echo abnormal) hypertrophic cardiomyopathy. There is considerable selection bias in pub- lished reports concerning hypertrophic car- (Asymptomatic) (Symptomatic) diomyopathy. Families referred to tertiary Die of *0 Die in CCF centres for genetic investigation usually repre- other dilating sent the severe cases where overt disease is cause LV present in many members. The prognosis and severity of disease in cases seen outside spe- ISUDDEN DEATH cialist tertiary referral centres may be better. The potential pathways along which any subject carrying one of the hypertrophic car- diomyopathy genes can progress are well The challenge in the full understanding of known (figure) but what determines which the genotype/phenotype relation in hyper- patient and which path is unknown. Sudden trophic cardiomyopathy is the wide range of death is always a risk at any age, irrespective of the phenotype and clinical affects seen in indi- the presence or absence of symptoms. For vidual subjects. The earliest accounts of the example a study of 150 cases presenting with disease recognised only the phenotype associ- minimal symptoms showed an overall fre- ated with asymmetric septal thickening and quency of sudden death of 1-5% per annum.7 left ventricular outflow obstruction. It is now Another similar study of 202 cases of hyper- apparent that the phenotype is very variable. trophic cardiomyopathy8 put the figure as low The hypertrophy may be symmetric or asym- as 0-6% per annum. Nevertheless the rate of metric, there may or may not be concomitant sudden death is higher in many series from right ventricular involvement, and the degree tertiary referral centres. Factors indicating of increase in left ventricular mass varies higher risk are a history of two or more deaths widely. Some cases have very minimal left ven- in the family, early onset of symptoms, a fall of tricular wall thickening. blood pressure on exercise, and severe left In some degree the variation in phenotypic ventricular wall thickening (> 35 mm). Syn- expression and clinical outcome is dependent cope attacks due to non-sustained ventricular on the specific gene and the mutation. In the tachycardia also indicate high risk, and in this heavy chain myosin gene some specific muta- subgroup amiodorone may reduce the risk.9 tions carry a very high risk of sudden death and have a "malignant" family history. Troponin T mutations produce far less strik- Marfan syndrome ing left ventricular hypertrophy than the aver- Like hypertrophic cardiomyopathy, Marfan age heavy chain myosin mutation but carry a syndrome is an autosomal dominant with a high risk of sudden death.'6 Many families very wide phenotypic range both within with hypertrophic cardiomyopathy, however, affected families and between families.10 Two show the full range of phenotypic appearances genes encode for components of -the even though all the members carry precisely common mutations producing Marfan syn- the same mutation. The cause for this drome involve the fibrillin gene on chromo- intrafamily variation is unknown. The ratio of some 15."l-'4 Fibrillin is a 350 kDa protein wild to mutant protein within the myo- produced by connective tissue cells and cardium, or whether it varies from case to case secreted into the extracellular matrix, where it within a family, is unknown. The reason why forms fibrillar aggregates which serve in adhe- some cases show striking regional ventricular sion of connective tissue structures. The gene involvement with a histologically abnormal is huge, being comprised of many thousands septum and a normal posterior wall is of base pairs encoding numerous repetitive unknown. Local haemodynamic factors might motifs. 14 15 Common repeats include an epider- account for the preponderance of septal thick- mal growth factor (EGF) like and transform- ing growth factor P (TGFP) binding protein like motifs. The phenotypic range is very Phenotypic spectrum infibrillin gene defects wide-abnormal fibrillin is responsible for all Chromosome 15, fibriflin-1 gene the phenotypes listed in the table. Mutations Marfan syndrome Eye dominant in exons 24-32, containing the longest contin- Skeleton dominant Cardiovascular system dominant uous stretch of EGF repeats, are particularly Neonatal Marfan syndrome liable to cause a severe defect, often fatal in the Isolated familial aortic root dilatation/aortic dissection neonatal period.16 MASS phenotype Mitral prolapse Congenital contractual arachnodactyly is Aortic root dilatation Skeletal abnormality due to mutations in the second fibrillin gene Skin involvement on chromosome 5.17 A phenotype very close to Marfanoid craniosynostosis Marfan syndrome has also been reported in Chronmosome 5, fibrdlin-2 gene single families due to a mutation in the gene Congenital contractual arachnodactyly coding for ax2(1) collagen and an unknown Genetic screening in HCM, Marfan syndrome, and the long QT syndrome 113

gene on chromosome 3. Some families breed heart is totally normal by all morphological true, that is, have all members confined to one and structural criteria. In such cases the cause phenotypic group, while other families show of death at necropsy, unless there is prior ECG wide variation.'4 15 Cardiac involvement can data, cannot be ascertained. In these cases occur in all the subtypes. screening by ECG of surviving relatives is the In large part the severity of the phenotype only way to establish the diagnosis. depends on the alterations in the amount of Linkage has been established on four chro- wild type versus mutant fibrillin that is pro- mosomes: 11, 7, 3, and 422-24 but a significant duced. Some mutations have a classic domi- number of families appear linked to none of nant negative effect in that mutant fibrillin these chromosomes.25 Between 1995 and 1996 interferes with the function of the normal pro- the genes for LQTI (long QT syndrome linked tein. This can be tested in vitro-fibroblasts to chromosome 11), for LQT2 (long QT syn- grown in culture from a skin biopsy produce a drome linked to chromosome 7), and for regular network of that can be LQT3 (long QT syndrome linked to chromo- seen as bundles by electron microscopy.819 some 3) have been identified.2630 The gene for Fibroblasts from subjects carrying abnormal LQT is KVLQT, which, when coassembled fibrillin genes produce microfibrils which form with min K, encodes for the IECS current.28 29 thinner fragmented strands-very severe cases The gene for LQT2 is HERO, a potassium produce no microfibrils. The phenotype thus channel that carries the IK, current. The gene depends on the ratio of wild type to mutant for LQT3 is SCN5A, the cardiac sodium fibrillin which in turn depends on the nature of channel gene and the three mutations the mutation. Over 80 single point mutations described so far3l affect a region thought to be deletions and insertions are known-so far important for sodium inactivation. It is tempt- every Marfan family has proved unique, a ing to speculate that a fifth locus for the long reflection of the huge size of the gene. QT syndrome is a mutation in the min K gene Relatively severe disease seems to occur with on chromosome 21. proportions of mutant fibrillin exceeding the The mutated genes for LQT2 and LQT3 threshold of approximately 10%. Many mis- have been expressed in Xenopus oocyte32-34; sense deletions, premature stop codons, or this has represented a major step forward whole segment deletions produce either no because it has allowed the first demonstration mutant fibrillin or relatively small proportions of an alteration in function due to a specific of mutant type, and the clinical phenotype is mutation. then mild. For example in the mitral aortic Expression of the mutant SCN5A genes has skeletal skin (MASS) syndrome the patients shown that they result in a small sustained lack the full clinical criteria for Marfan syn- inward current which is likely to be sufficient drome but have an insertion into the fibrillin to disrupt the normal balance between inward gene leading to a premature stop codon. The and outward currents during the plateau level of mutant fibrillin is as low as 6% and phase, and hence prolong cardiac action symptoms mild. potentials. It has practical importance in that In Marfan syndrome, therefore, while many this persistent inward current is blocked by cases are dominant negatives with mutant fib- mexiletine, a sodium channel blocker.34 rillin interfering with normal fibrillin, others Similarly, in a cellular model mimicking the are null alleles, that is, there is a simple defi- SCNSA and HERG defects present in LQTS, ciency of fibrillin. it has been observed that mexiletine corrects the action potential prolongation induced by anthopleurin, a blocker of sodium channel The long QT syndrome inactivation used to mimic LQT3, but not that At present the most straightforward geno- induced by dofetilide, an I,, blocker used to type/phenotype link occurs in the long QT mimic LQT2.'5 syndrome.20 Even here, however, the pheno- Expression of the mutant HERG genes has type is broader than initially envisaged. The led to the identification of two main conse- proband in families usually presents with syn- quences.'4 Some of the mutant proteins do not cope due to ventricular tachyarrhythmias or form functional channels and do interact with cardiac arrest on exercise or emotional stress, normal HERG channel when expressed in often under the age of 15. Within the family Xenopus oocytes. This implies that patients asymptomatic gene carriers may also have a with these mutations will probably express half long QT interval; of these up to 20% ulti- the normal number of channels carrying I,. mately become symptomatic. In some fami- Other mutant channels do not express lies, however, a few carriers may have a normal detectable currents but cause a dominant neg- QT interval-these individuals have a lower ative suppression of the normal HERG func- but still definite risk of syncope. The mortality tion; thus patients with these mutations will in untreated symptomatic cases exceeds 60% have a major reduction in IK, with a large effect in 15 years. Effective treatments are already on ventricular repolarisation. available and mortality has been reduced to The expression studies that have succeeded approximately 3-4% within 10 years from the in comparing different mutations3334 indicate first episode. The mainstays of treatment are P that the magnitude of the effect on depolarisa- blockers, left cardiac sympathetic denerva- tion varies according to the specific mutation. It tion,21 and pacing whenever there is clear evi- will be interesting to determine if there is a dence of pause induced or bradycardia correlate in the severity of the clinical manifes- dependent syncope.18 In subjects who die, the tations and the spectrum of either HERG or 114 Burn, Camm, Davies, Peltonen, Schwartz, Watkins

SCNSA dysfunction in patients with LQT2 dren in particular, phenotypic expression and LQT3. Patients with LQT1, LQT2, and detectable by echocardiography may be absent LQT3 have a different electrocardiographic and only appear later in life. phenotype'6 with a different shape of the T Genetic screening is at the moment expen- wave. Attempts are being made to correlate sive and its purpose must always be ques- the various mutations with the clinical tioned. One purpose is clear-the extension of responses.'7 scientific knowledge, which is likely in the The sodium channel blocker mexiletine future to benefit all individuals with the abnor- produced a considerable shortening of the QT mal gene. Unknown genes can only be identi- interval in LQT3 patients, but not in LQT2 fied by linkage analysis within large families and LQT1 patients. Even though a QT short- which have, therefore, become an extremely ening is by no means a guarantee of protection valuable research resource. Science has from life threatening arrhythmias, these data become competitive and there are pressures to suggest that it may be appropriate to test the study families so as to be the first to describe a potential value of mexiletine specifically in the new gene or mutation. The question is patients with the SCNSA mutations. whether this knowledge is beneficial, neutral, Heart rate increase produced a rather or harmful to both the family as a whole or to marked shortening of the QT interval among individuals within it. LQT3 patients. This effect was much less evi- One school of thought discourages the dent among LQT2 and LQT1 patients. The screening of asymptomatic children to detect inference could be that LQT3 patients may be gene carriers. The advice is heavily biased by at lesser risk of syncope during physical exer- the question of gene carriage in diseases such cise, when the progressive heart rate increase as Huntington's chorea, in which the disease is may allow appropriate QT shortening. These not manifest until middle age and nothing can patients may also be those less likely to be pro- be offered in terms of treatment. It is felt that tected by [B blockers, which would maintain an parents do not have the right to decide for excessively low heart rate. their children whether they wish to be tested. Finally, in these initial studies it has been The three cardiological genes discussed here possible to observe an association between do not, however, fit well into this scenario, and physical or emotional stress and syncope for gene testing has potential benefits. LQT1 patients, whereas LQT3 patients seem The situation is perhaps clearest for the to be more at risk during rest or sleep. LQT2 long QT syndrome. The different long QT patients are between these two extremes. genes differ in the phenotype they produce, These observations should be accepted with which in turn may determine the class of drug caution because the small number of patients to use. There is clear evidence that drug treat- studied so far does not allow extrapolation to ment will reduce the risk of sudden death and the entire population affected by long QT that asymptomatic gene carriers are also at syndrome and overlap certainly exists. risk. The benefit to individuals or families of Nonetheless, this study has provided the first knowing who has the gene are therefore well demonstration of differential responses of established. patients with long QT syndrome to interven- The situation is also clear in Marfan dis- tions targeted to their specific genetic defect. ease. The gene for fibrillin is so large that The long QT syndrome represents a model every family is unique with regard to their own of correlation between specific mutations and mutation. To sequence the gene in a Marfan specific alterations in function, thus represent- family is a major research undertaking, not a ing the best example available so far of routine clinical tool. Marfan families have to genotype-phenotype correlations. Further be managed by simple clinical techniques such development is likely to have an impact on the as echocardiography to detect developing aortic management of these patients by allowing root dilatation. Carriers have to be detected by novel and highly specific interventions to be physical measurements of hands, feet, aortic tested.38 root, and so on. There is a case to be made for centres developing specific clinical expertise in managing the cardiovascular complications of Genetic screening-when is it indicated? Marfan disease but not for a central genetic The cornerstone of managing a patient with screening service at the moment. This situa- hypertrophic cardiomyopathy, Marfan disease, tion may, however, change when automated or the long QT syndrome is a full family his- methods of screening the whole fibrillin gene tory of the ages and causes of death. These become established and widely available. In histories are a major indicator for assessing the vitro testing of fibroblasts in culture shows risk of sudden death in the individual. abnormal fibrillin structure, amount, or The taking of such histories will, however, metabolism. This can help establish the diag- inevitably alert the patient to the fact the dis- nosis in clinically ambiguous cases and, if car- ease is familial. If there have been sudden ried out in children, give some indication of deaths or aortic dissection at a young age in the expected severity of disease without having the family there is likely to be anxiety and to identify the precise mutation involved. requests for advice on who else is at risk, the Hypertrophic cardiomyopathy is currently magnitude of the risk, and whether anything the most uncertain with regard to the need to can be done. Echocardiography and electro- detect gene carriers. It is technically feasible to cardiography will detect only a proportion of screen all of the known contractile protein dis- gene carriers in all three conditions. In chil- ease genes to detect point mutations. With Genetic screening in HCM, Marfan syndrome, and the long QT syndrome 115

current technology this is laborious; in practise ever, or there may be anger if negative deci- such screening is largely limited to families sions over life style or having a family prove to with several affected individuals such that the have been needless. Those who find they do relevant disease can be first identified by link- carry the gene may be converted to chronic age analysis. invalidism or anxiety. It is reasonable to carry There is a risk of sudden death in asympto- out a full genetic study when the family asks matic gene carriers, even when echocardio- for it and unreasonable to deny it to them. graphic abnormalities are absent or minor. This view is supported by a lay body (Genetic The magnitude of risk of sudden death, even Interest Group)39 representing families. Two in asymptomatic gene carriers, is higher with provisos prevail: screening must be linked to some specific mutations of the heavy chain expert counselling, and family members who myosin gene than others and is also high with do not wish to take part must not be put under the troponin T gene. These are arguments for pressure. identifying the responsible gene and detecting carriers. Such high risk families will often, however, have already declared themselves by Genetic screening-who should do it? more than one sudden death. In symptomatic Genetic screening for hypertrophic cardiomy- individuals who are already having syncopal opathy and the long QT syndrome is at the attacks, investigation to identify the different moment largely carried out by tertiary referral mechanisms which produce syncope and treat centres in cardiology, with a specific research accordingly is entirely justified. What is far less interest in molecular genetics. This research clear at the moment is whether the asympto- service will ultimately have to be taken over by matic gene carrying relatives of such patients the National Health Service in some form. can have their risk of sudden death reduced by All three diseases discussed here have the either drug therapy or restriction of physical potential to devastate families by sudden activity. expected deaths, and families have the right to At a practical level this means that genetic the best available advice. Few cardiologists study of the family of an apparently isolated have genetic expertise and equally few geneti- living case of hypertrophic cardiomyopathy cists have the clinical experience either to may not be justified. The management of a manage cases or to identify carriers by clinical family in whom there is an unexpected sudden criteria. The future must lie in partnership and death under 50 years of age as the first mani- joint management. A recent confidential festation of hypertrophic cardiomyopathy is inquiry into counselling for genetic disorders40 not clearly defined. In the United Kingdom, has found that in all the non-cardiac clinical many index cases will be seen by pathologists specialties looked at so far there is imperfect working for the coroner and their responsibili- understanding of, or action on, the require- ties need defining. If nothing is done and ments of patients with genetic problems. It is another sudden death occurs, the family will unlikely that cardiovascular related genetic rightly feel aggrieved. It is therefore advisable disorders will prove to be different. to inform the family through their general practitioner, which will start the sequence of a Supported by an educational grant from Bayer plc. detailed family history and lead almost inevitably to referral of siblings and offspring to a cardiologist for screening. It is unresolved 1 Watkins H, Sideman JG, Seidman CE. Familial hyper- at the moment whether such screening should trophic cardiomyopathy: a genetic model of cardiac go beyond echocardiography and ECG. hypertrophy. Hum Mol Genet 1995;4:1721-7. 2 Watkins H, Conner D, Thierfelder L, Jarcho JA, MacRae C, McKenna WI, et al. Mutations in the cardiac myosin binding protein-C gene on chromosome 11 cause familial hypertrophic cardiomyopathy. Nature Genet 1995;11: Ethical aspects of genetic screening 434-7. One deterrent to the development of genetic 3 Watkins H, McKenna WI, Thierfelder L, Suk HJ, Anan R, O'Donaghue A, et al. Mutations in the genes for cardiac registers and predictive testing is the potential troponin T and alphatropomyosin in hypertrophic car- for eugenic abuse. In 1995 China passed a law diomyopathy. NEnglJMed 1995;332:1058-64. 4 Davies MJ, McKenna WJ. Hypertrophic cardiomyopathy prohibiting marriage if a couple carried genetic pathology and pathogenesis. Histopathology 1995;26: disorders. It is therefore mandatory that 493-500. 5 Sepp R, Servers NJ, Gourdie RG. Altered patterns of car- genetic services are developed for the benefit diac intercellular junction distribution in hypertrophic of individuals and their families rather than cardiomyopathy. Heart 1996;76:412-17. 6 Thierfelder L, Watkins H, MacRae C, Lamas R, McKenna improving society or making savings in costs W, Vosberg HP, et al. Alpha-tropomyosin and cardiac for the State. troponin T mutations cause familial hypertrophic car- diomyopathy: a disease of the sarcomere. Cell 1994;77: The detection of gene carriers can cause 701-12. considerable detriment to family members. 7 Spirito P, Rapezzi C, Autore C, Bruzzi P, Bellone P, Ortolani P, et al. Prognosis of asymptomatic patients with Insurance cover may become difficult to hypertrophic cardiomyopathy and nonsustained ventricu- obtain and companies may load premiums, lar tachycardia. Circulation 1994;90:2743-7. 8 Cecchi F, Olivotto I, Montereggi A, Santoro G, Dolara A, even if the individual is shown not to be a car- Maron BJ. Hypertrophic cardiomyopathy in Tuscany: rier. The parent who has passed on the gene to clinical course and outcome in an unselected regional population. Am Coll Cardiol 1995;26 1529-36. children may develop feelings of guilt and may 9 McKenna WI, Oakley CM, Krikler DM, Goodwin JF. be made to feel guilty by the other parent. Improved survival with amiodarone in patients with hypertrophic cardiomyopathy and ventricular tachycar- Siblings who do not carry the gene are spared dia. BrHeartJ_ 1985;53:412-16. any fiuther medical surveillance and will usu- 10 Pyeritz RE. The Marfan syndrome. In: Royce PM, Steinmann B, eds. Connective tissue and its heritable disor- ally be delighted. Guilt occurs over personal ders. New York: Wiley-Liss, 1994:961-8. survival when a much loved sibling dies, how- 11 Pereira L, Levran 0, Ramirez F, Lynch JR, Sykes B, 116 Burn, Camm, Davies, Peltonen, Schwartz, Watkins

Pyeritz RE, et al. A molecular approach to the stratification 27 Curran ME, Splawski I, Timothy KW, Vincent GM, Green of cardiovascular risk in families with Marfan's syn- ED, Keating MT. A molecular basis for cardiac arrhyth- drome. N EnglJ Med 1994;331:148-53. mia: HERG mutations cause long QT syndrome. Cell 12 Lee B, Godfrey M, Vitale E. Linkage of Marfan syndrome 1995;80:795-803. and a phenotypically related disorder to two fibrillin 28 Barhanin J, Lesage F, Guillemare E, Fink M, Lazdunski genes. Nature 1991;352:330-4. M, Romey G. KVLQT1 and IsK (mink) proteins associ- 13 Kainulainen K, Sakai LY, Child A, Pope FM, Puhakka L, ate to form the IKS cardiac potassium current. Nature Ryhanen L, et al. Two mutations in Marfan syndrome 1996;384:78-80. resulting in truncated fibrillin polypeptides. Proc Nadl 29 Sanguinetti MC, Curran ME, Zou A, Shen J, Spector PS, Acad Sci USA 1992;89:17-21. Atkinson DL, et al. Coassembly of KVLQT1 and mink 14 Dietz HC, Pyeritz RE. Mutations in the human gene for (IsK) proteins to form cardiac IKS potassium channel. fibrillin-1 (FBN-1) in the Marfan syndrome and related Nature 1996;384:80-3. disorders. Hum Mol Genet 1995;4:1799-809. 30 Wang Q, Curran ME, Splawski I, Burn TC, Millholland 15 Tsipouras P, Devereux RB. Marfan syndrome: genetic JM, VanRaay TJ, et al. Positional cloning of a novel basis and clinical manifestations. Semin Dermatol potassium channel gene. KVLQT 1 mutations cause car- 1993;12:219-28. diac arrhythmias. Nat Genet 1996;12:17-23. 16 Kainulainen K, Karttunen L, Puhakka L, Sakai L, 31 Wang Q, Shen J, Li Z, Timothy K, Vincent GM, Priori Peltonen L. Mutations in the fibrillin gene responsible for SG, et al. Cardiac sodium channel mutations in patients dominant ectopia lentis and neonatal Marfan syndrome. with long QT syndrome, an inherited cardiac arrhythmia. Nature Genet 1994;6:64-9. Hum Mol Genet 1995;4:1603-7. 17 Putnam EA, Zhang H, Ramirez F, Milewicz DM. Fibrillin- 32 Bennett PB, Yazawa K, Makita N, George AL. Molecular 2 (FBN2) mutations result in the Marfan-like disorder, mechanism for an inherited cardiac arrhythmia. Nature congenital contractural arachnodactyly. Nature Genet 1995;376:683-5. 1995;11:456-8. 33 Dumaine R, Wang Q, Keating MT, Hartmann HA, 18 Kielty CM, Shuttleworth CA. Abnormal fibrillin assembly Schwartz PJ, Brown AM, et al. Multiple mechanisms of by dermal fibroblasts from two patients with Marfan syn- Na+ channel linked long-QT syndrome. Circ Res 1996; drome. J Cell Biol 1994;124:997-1004. 78:916-24. 19 Kielty CM, Phillips JE, Child AH, Pope FM, Shuttleworth 34 Sanguinetti C, Curran ME, Spector PS, Keating MT. CA. Fibrillin secretion and microfibril assembly by Spectrum of HERG K+-channel dysfunction in an inher- Marfan dermal fibroblasts. Matrix Biol 1994;14:191-9. ited cardiac arrhythmia. Proc Nail Acad Sci USA 1996; 20 Schwartz PJ, Locati EH, Napolitano C, Priori SG. The 93:2208-12. long QT syndrome. In: Zipes DP, Jalife J, eds. Cardiac 35 Priori SG, Napolitano C, Cantu F, Brown AM, Schwartz electrophysiology from cell to bedside, 2nd ed. Philadelphia: PJ. Differential response to Na+ channel blockage, /fi WB Saunders, 1995:788-811. adrenergic stimulation, and rapid pacing in a cellular 21 Schwartz PJ, Locati EH, Moss AJ, Crampton RS, Trazzi R, model mimicking the SCNSA and HERG defects Ruberti U. Left cardiac sympathetic denervation in the present in the long QT syndrome. Circ Res 1996;78: therapy of congenital long QT syndrome: a world-wide 1009-15. report. Circulation 1991;84:503-1 1. 36 Moss AJ, Zareba W, Benhorin J, Locati EH, Hall WJ, 22 Keating MT, Atkinson D, Dunn C, Timothy K, Vincent Robinson JL, et al. ECG T-wave pattern in genetically GM, Leppert M. Linkage of a cardiac arrhythmia, the distinct forms of the hereditary long QT syndrome. long QT syndrome and the Harvey ras-l gene. Science Circulation 1995;92:2929-34. 1991;252:704-6. 37 Schwartz PJ, Priori SG, Locati EH, Napolitano C, Cantu 23 Jiang C, Atkinson D, Towbin JA, Splawski I, Lehmann F, Towbin AJ, et al. Long QT syndrome patients with MH, Li H, et al. Two long QT syndrome loci map to mutations on the SCNSA and HERG genes have differ- chromosomes 3 and 7 with evidence for further hetero- ential responses to Na+ channel blockade and to increase geneity. Nature Genet 1994;8:141-7. in heart rate. Implications for gene-specific therapy. 24 Schott nI, Charpantier F, Pettier S, Foley P, Drouin E, Circulation 1995;92:3381-96. Bouhour JB, et al. Mapping of a gene for long QT syn- 38 Roden DM, Lazzara R, Rosen MR, Schwartz PJ, Towbin dromes to chromosome 4q25-27. Am Y Hum Genet 1995; JA, Vincent GM, for the SADS Foundation Task Force 57:1114-22. on LQTS. Multiple mechanisms in the long-QT syn- 25 Towbin JA, Li H, Taggart RT, Lehmann MH, Schwartz drome. Current knowledge, gaps, and future directions. PJ, Satler CA, et al. Evidence of genetic heterogeneity in Circulation 1996;94:1196-2012. Romano-Ward long QT syndrome. Analysis of 23 fami- 39 Genetic Interest Group. The present organisation of genetic lies. Circulation 1994;90:2635-44. services in the United Kingdom. London: Genetic Interest 26 Wang Q, Shen J, Splawski I, Atkinson D, Li Z, Robinson Group, 1995. JL, et al. SCNSA mutations associated with an inherited 40 Harris R, Williamson P. Confidential enquiry into coun- cardiac arrhythmia, long QT syndrome. Cell 1995;80: selling for genetic disorders. J R Coll Physicians 1996; 805-11. 30:316-7.