Secondary Logo

Journal Logo

Unexplained sudden death, focussing on genetics and family phenotyping

Raju, Hariharan; Behr, Elijah R.

Current Opinion in Cardiology: January 2013 - Volume 28 - Issue 1 - p 19–25
doi: 10.1097/HCO.0b013e32835b0a9e
ARRHYTHMIAS: Edited by Anthony Tang
Free

Purpose of review Unexplained sudden death and the sudden arrhythmic death syndrome (SADS) affect a small but significant proportion of young and apparently healthy individuals. This review revisits the causes underlying such deaths and the investigational strategies that identify surviving family who may be at risk.

Recent findings Recent epidemiological data is available from case series or government records. The yield from familial cardiological evaluation for inherited conditions has been supported by additional small series. The greatest advance has come with molecular autopsy studies, which have utilized various methodologies and candidate genes to investigate SADS cases and their families.

Summary The latest research replicates and extends the existing knowledge regarding epidemiology and familial evaluation of SADS, whilst genetic studies support a role for the molecular autopsy.

Cardiovascular Sciences Research Centre, St George's University of London, London, UK

Correspondence to Elijah R. Behr, Cardiovascular Sciences Research Centre, St George's University of London, London, SW17 0RE, UK. Tel: +44 2087255939; e-mail: ebehr@sgul.ac.uk

Back to Top | Article Outline

INTRODUCTION

Unexplained sudden death (SUD) refers to an unexpected sudden cardiac death (SCD) that occurs in an apparently healthy and often young individual for no apparent reason. If the death remains unexplained despite comprehensive post-mortem examination, including histopathology and toxicology, the term sudden arrhythmic death syndrome (SADS) is preferred [1]. It is comparable to sudden infant death syndrome (SIDS) as a diagnosis of exclusion after infancy.

Back to Top | Article Outline

EPIDEMIOLOGY OF UNEXPLAINED DEATHS

National studies of SCD have shown variable proportions and calculated incidences of SADS. This is partly because of different populations and criteria for defining causes of death in each study. SADS makes up 13.5% of young SCD (under age 35 years) in the UK, based on death certification, corresponding to an annual incidence of 0.24 per 100 000 population [2]. Recent Danish estimates provide a three-fold higher incidence of up to 0.81 per 100 000 [3]. Irish incidence (0.76 per 100 000) is similar to the Scandinavian experience, although the exclusion of paediatric deaths probably results in an underestimate [4]. Nevertheless, a consistent finding is that national certification underreports SADS in comparison with accurate postmortem interpretation. This is presumably because of the misclassification on certification and the inadequacy of International Classification of Diseases coding of sudden death [5].

Eckart et al.[6▪▪] reported on US military recruits, who provide a uniquely documented multiethnic but biased cohort prescreened by self-reported questionnaire, with exclusion of those affected by known cardiac conditions. This screening process would be expected to reduce the overall incidence of SCD by excluding more readily diagnosable structural conditions, and may increase the proportion of SADS, given that sudden death is often an unheralded event. Nonetheless, they report an annual incidence of SADS of 1.2 per 100 000 in individuals less than 35 years old and 2.0 per 100 000 in older recruits [6▪▪]. Both estimates are higher than European studies, although the authors suggest that the relatively highly trained status of military recruits rather than ethnic composition accounts for the observed difference.

Box 1

Box 1

Young adult men from southeast Asia, in particular northeast Thailand, demonstrate the highest incidence of SADS (38 per 100 000 men per annum) [7], the cause being linked predominantly to the Brugada syndrome [8]. However, a 2011 autopsy-based study of the Han population of south China reports an annual adult incidence of just 1 per 100 000 total population, more akin to European levels [9]. Furthermore, the authors found that only 1 of 74 cases investigated by post-mortem SCN5A sequencing carried a probable Brugada syndrome-associated mutation [10]. This calls into question whether Brugada syndrome is the overriding cause underlying SADS amongst Han Chinese compared with other ethnic southeast Asians.

Back to Top | Article Outline

FAMILIAL EVALUATION FOLLOWING UNEXPLAINED SUDDEN DEATH

Cardiological evaluation of families of SADS deaths determined that both ion channel diseases and cardiomyopathies underlie some cases of SADS [1] and identified surviving blood relatives at risk [1]; the investigation protocol was refined in a later study, improving the diagnostic yield to 53% [11]. These and subsequent studies of clinical diagnoses made following familial SUD or SADS are summarized in Fig. 1[1,11–13,14▪,15].

FIGURE 1

FIGURE 1

A tertiary centre referral-based Dutch series of 140 cases yielded 33% diagnoses following SADS with familial cardiological and targeted genetic evaluation, with inherited conditions comprising 97% of these [14▪]. They present a protocol-driven approach to familial evaluation with investigations beyond resting ECG performed on clinical suspicion only. This may have reduced the overall yield, but increased the sensitivity of additional investigations. Relatives’ ECGs contributed to familial diagnosis in 14% of all SADS cases, whereas drug-provocation testing and cardiac magnetic resonance (CMR) imaging showed higher yields, albeit when utilized in a minority of cases directed by clinical suspicion. Furthermore, the absolute number of diagnostic CMR scans was marginally higher than echocardiography. Whilst primary arrhythmia syndromes represented the majority of causes, irrespective of age, the proportion related to cardiomyopathy increases with age [14▪], reflecting age-related penetrance. If replicated by other studies, this may lead to increasing importance of structural imaging, including CMR, in evaluation of families in which the index SADS case dies in adulthood.

A population-based study in a different region of the Netherlands utilizing cardiogenetic evaluation following aborted or premature SCD (including four SADS cases) showed a clinical diagnostic yield of just 14% [13]. Post-mortem protocols and familial evaluation were performed at the clinician's discretion. The authors explained that this represented a ‘real-world’ experience, but concluded that such families should be seen at regional and national centres of excellence [16].

Following SADS, surviving family members affected by an inherited arrhythmia syndrome require risk stratification for their risk of sudden death. Brugada syndrome poses a particular problem, given that the only universally accepted treatment is an implantable cardiovertor defibrillator (ICD). A retrospective study identified that the majority of SADS cases attributable to familial Brugada syndrome would not have been considered high risk [17▪], as only a minority reported prior syncope. Asymptomatic and ostensibly low-risk individuals may therefore account for the majority of premature deaths by virtue of low event rates.

Indeed, only a minority of evaluated relatives are at high risk from an inherited cardiac condition sufficient to require ICD implantation [11]. A recent British study confirmed that ICD implantation was indicated for only 2% of relatives of SADS cases [12]. This may reflect that the SADS case was the highest risk individual in each family but was undetected prior to death. Nevertheless, 30 of 146 (21%) evaluated relatives required risk modification, medical therapy, and on-going follow-up and advice [12].

The potential diagnostic utility of investigating the surviving paediatric relatives following SADS has been suggested in a small series of just 16 cases. Three (19%) families received a diagnosis [18]. Although this is less than the yield in predominantly adult familial evaluation, the potential for diagnosis of treatable disease in children is compelling [18]. Another approach is to integrate adult and paediatric evaluation, offering additional benefits: better communication of familial results; more confident assessment of the significance of borderline results; and possibly increased diagnostic yield [16].

Back to Top | Article Outline

MOLECULAR AUTOPSY AND GENETIC TESTING

The molecular autopsy has progressed since early work by Tester and Ackerman suggested a 31% yield of mutations in genes associated with ion channel diseases [19,20▪▪]. Most further research has focussed on screening candidate genes that correspond to diagnoses made at familial evaluation (Fig. 1). These include catecholaminergic polymorphic tachycardia (CPVT) – RyR2; long QT syndrome (LQTS) – KCNQ1, KCNH2, SCN5A, KCNE1, and KCNE2; and Brugada syndrome – SCN5A. These are summarized in Fig. 2[19,20▪▪,21▪,22,23,24▪▪,25].

FIGURE 2

FIGURE 2

Recent data from Skinner et al.[24▪▪] reinforces a 15% yield from mutation analysis of LQTS-risk genes in a population-based New Zealand study. The molecular autopsy also complemented familial evaluation. In three out of the five families in which a SADS case was diagnosed with LQTS on genetic testing, clinical findings in relatives were insufficient to meet diagnostic criteria [24▪▪]. More recently, Tester et al.[20▪▪] reported on an expanded series of 173 SADS cases. Reassuringly, this shows a similar yield of 14% from LQTS-targeted sequencing. Nonetheless, their yield of LQTS and CPVT risk mutations has fallen in comparison with their original report [19]. It is possible that, as the threshold for referral to a post-mortem cardiogenetics service has decreased, the yield of an underlying genetic cause of death has fallen concurrently. This new data also supports a higher prevalence of a positive molecular diagnosis in SADS cases with a family history of possible cardiac events [20▪▪].

Another recent but small, single-centre, retrospective, molecular autopsy series from Denmark reveals a yield of 8.3% (3 of 36) following selective RyR2 sequencing [23]. This compares reasonably with the 12% (20 of 173) yield from Tester et al.[20▪▪]. All post-mortem studies thus far have used a targeted approach, with up to 64 of the 105 exons of RyR2 being sequenced; the differences in yield may therefore partly be explained by variation in the coding sequence evaluated.

The earliest molecular autopsy case series were studied retrospectively, using archived, formalin-fixed, paraffin-embedded (FFPE) tissue as a DNA source. Unfortunately, formalin fixation causes alterations in DNA sequence. Residual dried blood spot samples from national newborn screening programmes have therefore been assessed as a novel source. Gladding et al.[25] extracted DNA successfully from New Zealand blood spots and used whole-genome amplification prior to sequencing. They were able to perform diagnostic screening of LQTS-risk genes in all 19 cases of their cohort of SIDS and SADS cases, despite some blood spots being up to 39 years old, and showed that six disease-causing mutations identified were all present in at least one first-degree blood relative [25]. Blood-spot-derived DNA has been replicated in a 2012 Danish retrospective population-based study, which determined a yield of 11% of LQTS-associated mutations in 44 cases [21▪]. Neonatal screening cards are therefore a better clinical source of DNA than FFPE tissue, although fresh or frozen blood and tissue remain the gold standard.

In the absence of suitable post-mortem DNA, Wisten et al.[26] screened DNA from one or more first-degree blood relatives of SADS cases for the evidence of mutations in the five major LQTS genes, irrespective of familial phenotype. They identified LQTS-associated mutations in family members as responsible for 3 of 25 (12%) cases [26], a yield not dissimilar to molecular autopsy series. De-novo mutations may therefore cause only a minority of unheralded LQTS deaths, and screening DNA in relatives (particularly both parents) can be an effective approach in SADS. However, the same investigators screened for RyR2 mutations in families without LQTS-associated mutations and found no additional yield [26]. This may represent the higher prevalence of de-novo mutations amongst those affected by CPVT [27]. A similar family genetic screening protocol reported by Allegue et al.[28] found four LQTS mutations (11%) in a series of 35 SADS cases.

Alternative technologies have also been utilized to reduce costs: the MassARRAY and SnapShot systems that identify specific pathogenic mutations allowing the screening of multiple samples concurrently. They lack the capacity, however, for the detection of novel mutations that are private to a family and are likely to account for a number of cases. Nonetheless, after using a MassARRAY panel of 433 known mutations, Allegue et al.[28] detected five mutations amongst SADS cases or relatives and then only one further novel LQTS-associated mutation by direct sequencing. This may be a cost-effective first step in a tiered approach to molecular autopsy in an already well-characterized population. In contrast, the SnapShot technique described by Edelmann et al.[29] was designed to identify just 58 known mutations; it failed to demonstrate any known LQTS mutations amongst 35 SADS cases.

Familial evaluation has also recognized that cardiomyopathies underlie some SADS cases with apparently normal hearts [11]. Zhang et al.[22] identified three (12%) possible or probable disease-causing genetic variants in PKP2, the most common arrhythmogenic right ventricular cardiomyopathy (ARVC)-risk gene, in 25 SADS cases. One (4%) mutation was a ‘radical’ frame-shift mutation strongly associated with disease [30], strengthening the view that ARVC is responsible for a small, but clinically significant, minority of SADS cases [11]. Hypertrophic cardiomyopathy appears to constitute a smaller proportion of SADS cases on the basis of familial evaluation [11] and appears to have a negligible definitive molecular yield on the basis of the recent MassARRAY study [28].

Overall, the yield from familial evaluation still exceeds that of molecular autopsy, but it is likely that both approaches will be useful and complementary; Fig. 3 illustrates a combined protocol summarizing an appropriate diagnostic pathway [11,14▪].

FIGURE 3

FIGURE 3

Back to Top | Article Outline

OTHER SUSPICIOUS SUDDEN DEATHS

SADS is ‘diagnosed’ at postmortem by exclusion of a definitive cause of death [11]. However, patient history and activity at time of death can lead to incorrect certification. For example, otherwise unexplained drowning has been investigated using molecular autopsy by Tester et al.[31]. They found 23% (8 of 35) of cases with either KCNQ1 or RYR2 mutations; the yield increased to nearly 50% in those drowning victims with a personal or family history suspicious for undiagnosed channelopathies [31]. The authors conclude very reasonably that unexplained drownings should be considered as potential SADS cases.

Sudden unexpected death in epilepsy (SUDEP) is a certifiable cause of death, and post-mortem evaluation is identical to SADS, albeit in the context of a prior history of epilepsy. The pathophysiology of such deaths is contentious, but likely includes arrhythmias either secondary to hypoventilation or primary [32]. A study of SUDEP cases from Sydney revealed two (4.2%) of 48 cases had probable LQTS-risk mutations [33]. This is lower than the published series of true SADS cases and reinforces animal studies that support hypoventilation as the main cause. Nevertheless, according to mouse models, KCNQ1, one of the main LQTS genes, may also be expressed in the brain [34]. This suggested that some ion channelopathies may have both epileptic and cardiological phenotypes. It is likely, however, that SUDEP cases are a heterogeneous mix, with true primary cardiac causes representing a minority.

Back to Top | Article Outline

NOVEL CAUSES OF FATAL ARRHYTHMIAS

The early repolarization syndrome (ERS) is part of the spectrum of idiopathic ventricular fibrillation (IVF); those affected show a stereotypical ECG appearance of inferior J-point elevation in the context of prior cardiac arrest and no other phenotype. Whilst the ECG appearance appears to be heritable in the general population, it also lies within the spectrum of normal [35]. A recent single-centre study of familial evaluation suggests that the ECG appearance was more prevalent than expected in the blood relatives of SADS cases [36]. However, the prevalence was unchanged amongst those who were diagnosed with another inherited cardiac condition compared with those with no diagnosis. This questions the clinical conclusions that may be drawn from a finding of early repolarization in the context of family history of SADS: it neither predicts the likelihood of an alternative underlying familial condition nor indicates ERS as the cause of SADS.

Mutations in KCND3, responsible for the cardiac transient outward potassium current, were initially associated with IVF; mutations in this gene have now been found at low frequency (<2%) amongst SADS cases [37]. This supports the premise that IVF and SADS share some common underlying causes, and potentially allows cascade familial evaluation in situations where no phenotype would be apparent in surviving family, even amongst carriers. However, the clinical value of making these diagnoses at present is somewhat limited, given the absence of suitable risk stratification for mutation carriers without prior arrhythmias or symptoms.

Back to Top | Article Outline

CONCLUSION

Clinical and genetic research continues to define the variety of cardiac conditions that underlie SADS deaths and the role for the molecular autopsy. Future research will investigate novel clinical diagnostic methods, the full range of rare and common genetic variation underlying SADS, and the clinical utility in families.

Back to Top | Article Outline

Acknowledgements

Dr Raju receives funding from the British Heart Foundation (charitable organization) and has previously received funding from Cardiac Risk in the Young (charitable organization). Dr Behr is funded by the Higher Education Funding Council for England (HEFCE) and has received research funding from Biotronik, Boston Scientific, Cardiac Risk in the Young and the British Heart Foundation.

Back to Top | Article Outline

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

  • ▪ of special interest
  • ▪▪ of outstanding interest

Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 81).

Back to Top | Article Outline

REFERENCES

1. Behr E, Wood DA, Wright M, et al. Cardiological assessment of first-degree relatives in sudden arrhythmic death syndrome. Lancet 2003; 362:1457–1459.
2. Papadakis M, Sharma S, Cox S, et al. The magnitude of sudden cardiac death in the young: a death certificate-based review in England and Wales. Europace 2009; 11:1353–1358.
3. Winkel BG, Holst AG, Theilade J, et al. Nationwide study of sudden cardiac death in persons aged 1–35 years. Eur Heart J 2011; 32:983–990.
4. Margey R, Roy A, Tobin S, et al. Sudden cardiac death in 14- to 35-year olds in Ireland from 2005 to 2007: a retrospective registry. Europace 2011; 13:1411–1418.
5. Behr ER, Casey A, Sheppard M, et al. Sudden arrhythmic death syndrome: a national survey of sudden unexplained cardiac death. Heart 2007; 93:601–605.
6▪▪. Eckart RE, Shry EA, Burke AP, et al. Sudden death in young adults: an autopsy-based series of a population undergoing active surveillance. J Am Coll Cardiol 2011; 58:1254–1261.

This is an epidemiological study of sudden death of US military recruits in a well-characterized population, which used US registry data and standardized postmortem investigations to determine the underlying cardiac causes of death.

7. Tungsanga K, Sriboonlue P. Sudden unexplained death syndrome in north-east Thailand. Int J Epidemiol 1993; 22:81–87.
8. Sangwatanaroj S, Prechawat S, Sunsaneewitayakul B, et al. New electrocardiographic leads and the procainamide test for the detection of the Brugada sign in sudden unexplained death syndrome survivors and their relatives. Eur Heart J 2001; 22:2290–2296.
9. Cheng J, Makielski JC, Yuan P, et al. Sudden unexplained nocturnal death syndrome in Southern China: an epidemiological survey and SCN5A gene screening. Am J Forensic Med Pathol 2011; 32:359–363.
10. Nakajima T, Kaneko Y, Saito A, et al. Identification of six novel SCN5A mutations in Japanese patients with Brugada syndrome. Int Heart J 2011; 52:27–31.
11. Behr ER, Dalageorgou C, Christiansen M, et al. Sudden arrhythmic death syndrome: familial evaluation identifies inheritable heart disease in the majority of families. Eur Heart J 2008; 29:1670–1680.
12. Caldwell J, Moreton N, Khan N, et al. The clinical management of relatives of young sudden unexplained death victims: implantable defibrillators are rarely indicated. Heart 2012; 98:631–636.
13. Hendrix A, Borleffs CJW, Vink A, et al. Cardiogenetic screening of first-degree relatives after sudden cardiac death in the young: a population-based approach. Europace 2011; 13:716–722.
14▪. Van der Werf C, Hofman N, Tan HL, et al. Diagnostic yield in sudden unexplained death and aborted cardiac arrest in the young: the experience of a tertiary referral center in The Netherlands. Heart Rhythm 2010; 7:1383–1389.

This is the largest series of familial evaluation following SUD yet reported, involving 140 families, which confirmed the value of this evaluation modality in a protocol-driven study.

15. Tan HL, Hofman N, van Langen IM, et al. Sudden unexplained death: heritability and diagnostic yield of cardiological and genetic examination in surviving relatives. Circulation 2005; 112:207–213.
16. Raju H, Behr ER. Paediatric evaluation for inherited conditions: how do we investigate? Europace 2011; 13:304–305.
17▪. Raju H, Papadakis M, Govindan M, et al. Low prevalence of risk markers in cases of sudden death due to Brugada syndrome: relevance to risk stratification in Brugada syndrome. J Am Coll Cardiol 2011; 57:2340–2345.

This study underlines that asymptomatic, and apparently low-risk, individuals constitute the majority of SADS cases because of Brugada syndrome, thereby emphasizing the need for improvement of risk stratification.

18. Tomaske M, Keller DI, Bauersfeld U. Sudden cardiac death: clinical evaluation of paediatric family members. Europace 2011; 13:421–426.
19. Tester DJ, Ackerman MJ. Postmortem long QT syndrome genetic testing for sudden unexplained death in the young. J Am Coll Cardiol 2007; 49:240–246.
20▪▪. Tester DJ, Medeiros-Domingo A, Will ML, et al. Cardiac channel molecular autopsy: insights from 173 consecutive cases of autopsy-negative sudden unexplained death referred for postmortem genetic testing. Mayo Clin Proc 2012; 87:524–539.

This is the largest series of referred cases for molecular autopsy following SUD/SADS yet reported, involving 173 cases, which replicated findings from their smaller series and added analysis of clinical and demographic data that may help with the prioritization of molecular investigation if performed in a stepwise fashion.

21▪. Winkel BG, Larsen MK, Berge KE, et al. The prevalence of mutations in KCNQ1, KCNH2, and SCN5A in an unselected national cohort of young sudden unexplained death cases. J Cardiovasc Electrophysiol 2012. doi: 10.1111/j.1540-8167.2012.02371.x. [Epub ahead of print]

This is a population-based study of molecular autopsy following SADS, suggesting a lower molecular yield of LQTS than seen in referred population series.

22. Zhang M, Tavora F, Oliveira JB, et al. PKP2 mutations in sudden death from arrhythmogenic right ventricular cardiomyopathy (ARVC) and sudden unexpected death with negative autopsy (SUDNA). Circ J 2012; 76:189–194.
23. Larsen MK, Berge KE, Leren TP, et al. Postmortem genetic testing of the ryanodine receptor 2 (RYR2) gene in a cohort of sudden unexplained death cases. Int J Legal Med 2012. doi: 10.1007/s00414-011-0658-2. [Epub ahead of print]
24▪▪. Skinner JR, Crawford J, Smith W, et al. Prospective, population-based long QT molecular autopsy study of postmortem negative sudden death in 1 to 40 year olds. Heart Rhythm 2011; 8:412–419.

This is the first study that quantifies the yields of both molecular and clinical evaluation of SADS in a comparative fashion, thus highlighting their complementary qualities.

25. Gladding PA, Evans CA, Crawford J, et al. Posthumous diagnosis of long QT syndrome from neonatal screening cards. Heart Rhythm 2010; 7:481–486.
26. Wisten A, Boström IM, Mörner S, Stattin E-L. Mutation analysis of cases of sudden unexplained death, fifteen years after death: prompt genetic evaluation after resuscitation can save future lives. Resuscitation 2012. doi: 10.1016/j.resuscitation.2012.05.015. [Epub ahead of print]
27. Priori SG. Clinical and molecular characterization of patients with catecholaminergic polymorphic ventricular tachycardia. Circulation 2002; 106:69–74.
28. Allegue C, Gil R, Blanco-Verea A, et al. Prevalence of HCM and long QT syndrome mutations in young sudden cardiac death-related cases. Int J Legal Med 2011; 125:565–572.
29. Edelmann J, Schumann S, Nastainczyk M, et al. Long QT syndrome mutation detection by SNaPshot technique. Int J Legal Med 2012; 126:969–973.
30. Kapplinger JD, Landstrom AP, Salisbury BA, et al. Distinguishing arrhythmogenic right ventricular cardiomyopathy/dysplasia-associated mutations from background genetic noise. J Am Coll Cardiol 2011; 57:2317–2327.
31. Tester DJ, Medeiros-Domingo A, Will ML, Ackerman MJ. Unexplained drownings and the cardiac channelopathies: a molecular autopsy series. Mayo Clin Proc 2011; 86:941–947.
32. Shorvon S, Tomson T. Sudden unexpected death in epilepsy. Lancet 2011; 378:2028–2038.
33. Tu E, Bagnall RD, Duflou J, Semsarian C. Postmortem review and genetic analysis of sudden unexpected death in epilepsy (SUDEP) cases. Brain Pathol 2011; 21:201–208.
34. Goldman AM, Glasscock E, Yoo J, et al. Arrhythmia in heart and brain: KCNQ1 mutations link epilepsy and sudden unexplained death. Sci Transl Med 2009; 1:2ra6.
35. Noseworthy PA, Tikkanen JT, Porthan K, et al. The early repolarization pattern in the general population: clinical correlates and heritability. J Am Coll Cardiol 2011; 57:2284–2289.
36. Nunn LM, Bhar-Amato J, Lowe MD, et al. Prevalence of J-point elevation in sudden arrhythmic death syndrome families. J Am Coll Cardiol 2011; 58:286–290.
37. Giudicessi JR, Ye D, Kritzberger CJ, et al. Novel mutations in the KCND3-encoded Kv4.3 K+ channel associated with autopsy-negative sudden unexplained death. Hum Mutat 2012; 33:989–997.
Keywords:

inherited cardiac conditions; molecular autopsy; sudden arrhythmic death syndrome; sudden unexplained death; sudden unexplained death in epilepsy

© 2013 Lippincott Williams & Wilkins, Inc.