Long QT syndrome
|Long QT syndrome|
|Drawing of a normal ECG tracing (sinus rhythm) with waves, segments, and intervals labeled. The QT interval is marked by the blue line at bottom.|
|Symptoms||Fainting, hearing loss|
|Causes||Genetic, certain medications, low blood potassium, low blood calcium, heart failure|
|Risk factors||Family history of sudden death|
|Diagnostic method||Electrocardiogram (EKG) together with clinical findings|
|Differential diagnosis||Brugada syndrome, arrhythmogenic right ventricular dysplasia|
|Treatment||Avoiding strenuous exercise, getting sufficient potassium, beta blockers, implantable cardiac defibrillator|
|Frequency||~ 1 in 7,000|
|Deaths||~3,500 a year (USA)|
Long QT syndrome (LQTS) is a condition which affects repolarization of the heart after a heartbeat. This results in an increased risk of an irregular heartbeat which can result in palpitations, fainting, drowning, or sudden death. These episodes can be triggered by exercise or stress. Other associated symptoms may include hearing loss.
Long QT syndrome may be present at birth or develop later in life. The inherited form may occur by itself or as part of larger genetic disorder. Onset later in life may result from certain medications, low blood potassium, low blood calcium, or heart failure. Medications that are implicated include certain antiarrhythmic, antibiotics, and antipsychotics. Diagnosis is based on an electrocardiogram (EKG) finding a corrected QT interval of greater than 440 to 500 milliseconds together with clinical findings.
Management may include avoiding strenuous exercise, getting sufficient potassium in the diet, the use of beta blockers, or an implantable cardiac defibrillator. For people with LQTS who survive cardiac arrest and remain untreated, the risk of death within 15 years is greater than 50%. With proper treatment this decreases to less than 1% over 20 years.
Long QT syndrome is estimated to affect 1 in 7,000 people. Females are affected more often than males. Most people with the condition develop symptoms before they are 40 years old. It is a relatively common cause of sudden death along with Brugada syndrome and arrhythmogenic right ventricular dysplasia. In the United States it results in about 3,500 deaths a year. The condition was first clearly described in 1957.
- 1 Signs and symptoms
- 2 Causes
- 3 Pathophysiology
- 4 Diagnosis
- 5 Treatment
- 6 Outcomes
- 7 Epidemiology
- 8 History
- 9 See also
- 10 References
- 11 External links
Signs and symptoms
Many people with long QT syndrome have no signs or symptoms.
Some people may experience the following symptoms:
- Chest pain
- Fainting (or syncope). This may occur when the patient is emotionally or physically stressed. It is unusual in QT syndrome to have any signs before the person actually faints.
- If there is sudden death, and doctors suspect long QT syndrome as the cause, they may recommend that the family members of the deceased get tested for the disease.
Risk factors for long QT syndrome include the following:
- female sex
- increasing age
- liver or renal impairment
- family history of congenital long QT syndrome
- pre-existing cardiovascular disease
- electrolyte imbalance: especially hypokalemia, hypocalcemia, hypomagnesemia
- concurrent administration of interacting drugs
Anorexia nervosa has been associated with sudden death, possibly due to QT prolongation. It can lead a person to have dangerous electrolyte imbalances, leading to acquired long QT syndrome and can in turn result in sudden cardiac death. This can develop over a prolonged period of time, and the risk is further heightened when feeding resumes after a period of abstaining from consumption. Care must be taken under such circumstances to avoid complications of refeeding syndrome.
Mental stress or physical exertion may trigger arrhythmia in some people with LQTS.
LQTS can arise from mutation of one of several genes. These mutations tend to prolong the duration of the ventricular action potential (APD), thus lengthening the QT interval. LQTS can be inherited in an autosomal dominant or a much less common autosomal recessive fashion. The autosomal recessive forms of LQTS tend to have a more severe phenotype, with two variants having associated with other congenital heart disease, autism, immune deficiency and complex syndactyly (LQT8) or congenital neural deafness (LQT1). A number of specific gene loci have been identified to be associated with LQTS. Genetic testing for LQTS is clinically available and may help to direct appropriate therapies . The most common causes of LQTS are mutations in the genes KCNQ1 (LQT1), KCNH2 (LQT2), and SCN5A (LQT3); the following is a list of all known genes associated with LQTS:
|LQT1||192500||alpha subunit of the slow delayed rectifier potassium channel (KvLQT1 or KCNQ1)||The current through the heteromeric channel (KvLQT1 + minK) is known as IKs. These mutations often cause LQT by reducing the amount of repolarizing current, which is required to terminate the action potential, leading to an increase in the action potential duration. These mutations tend to be the most common yet least severe.|
|LQT2||152427||Alpha subunit of the rapid delayed rectifier potassium channel (hERG + MiRP1)||Current through this channel is known as IKr. This phenotype is also probably caused by a reduction in repolarizing current.|
|LQT3||603830||alpha subunit of the sodium channel (SCN5A)||Current through this channel is commonly referred to as INa. Depolarizing current through the channel late in the action potential is thought to prolong APD. The late current is due to the failure of the channel to remain inactivated. As a consequence, it can enter a bursting mode, during which significant current enters abruptly when it should not. These mutations are more lethal but less common.|
|LQT4||600919||anchor protein Ankyrin B||LQT4 is very rare. Ankyrin B anchors the ion channels in the cell.|
|LQT5||176261||beta subunit MinK (or KCNE1), which coassembles with KvLQT1||-|
|LQT6||603796||beta subunit MiRP1 (or KCNE2), which coassembles with hERG||-|
|LQT7||170390||potassium channel KCNJ2 (or Kir2.1)||The current through this channel and KCNJ12 (Kir2.2) is called IK1. LQT7 leads to Andersen-Tawil syndrome.|
|LQT8||601005||alpha subunit of the calcium channel Cav1.2 encoded by the gene CACNA1c.||Leads to Timothy's syndrome.|
LQT1 is the most common type of long QT syndrome, making up about 30 to 35% of all cases. The LQT1 gene is KCNQ1, which has been isolated to chromosome 11p15.5. KCNQ1 codes for the voltage-gated potassium channel KvLQT1 that is highly expressed in the heart. The product of the KCNQ1 gene is thought to produce an alpha subunit that interacts with other proteins (in particular, the minK beta subunit) to create the IKs ion channel, which is responsible for the delayed potassium rectifier current of the cardiac action potential.
Mutations to KCNQ1 can be inherited in an autosomal dominant or an autosomal recessive pattern in the same family. In the autosomal recessive mutation of this gene, homozygous mutations lead to severe prolongation of the QT interval (due to near-complete loss of the IKs ion channel), and are associated with increased risk of ventricular arrhythmias and congenital deafness. This variant of LQT1 is known as the Jervell and Lange-Nielsen syndrome. Furthermore, LQT1 patients also have an endocrine phenotype. During a glucose load, LQT1 patients respond with an exaggerated insulin secretion followed by a temporary insulin resistance. When the resistance diminishes, LQT1 patients are at risk for hypoglycaemia.
Most individuals with LQT1 show paradoxical prolongation of the QT interval with infusion of epinephrine. This can also unmark latent carriers of the LQT1 gene. Many missense mutations of the LQT1 gene have been identified. These are often associated with a high frequency of syncopes, but less sudden death than LQT2.
The LQT2 type is the second-most common gene location in long QT syndrome, making up about 25 to 30% of all cases. This form of long QT syndrome most likely involves mutations of the 'human ether-a-go-go related gene' (hERG) on chromosome 7. The hERG gene (also known as KCNH2) is part of the rapid component of the potassium rectifying current (IKr). (The IKr current is mainly responsible for the termination of the cardiac action potential, and therefore the length of the QT interval.) The normally functioning hERG gene allows protection against early after depolarizations.
Most drugs that cause long QT syndrome do so by blocking the IKr current via the hERG gene. These include erythromycin, terfenadine, and ketoconazole. The hERG channel is very sensitive to unintended drug binding due to two aromatic amino acids, the tyrosine at position 652 and the phenylalanine at position 656. These amino acid residues are poised so a drug binding to them blocks the channel from conducting current. Other potassium channels do not have these residues in these positions, so are, therefore, not as prone to blockage.
The LQT3 type of long QT syndrome involves mutation of the gene that encodes the alpha subunit of the Na+ ion channel. This gene is located on chromosome 3p21-24, and is known as SCN5A (also hH1 and NaV1.5). The mutations involved in LQT3 slow the inactivation of the Na+ channel, resulting in prolongation of the Na+ influx during depolarization. However, the mutant sodium channels inactivate more quickly, and may open repetitively during the action potential.
A large number of mutations have been characterized as leading to or predisposing to LQT3. Calcium has been suggested as a regulator of SCN5A protein, and the effects of calcium on SCN5A may begin to explain the mechanism by which some these mutations cause LQT3. Furthermore, mutations in SCN5A can cause Brugada syndrome, cardiac conduction disease, and dilated cardiomyopathy. In rare situations, some affected individuals can have combinations of these diseases.
LQT5 is an autosomal-recessive, relatively uncommon form of LQTS. It involves mutations in the gene KCNE1, which encodes for the potassium channel beta subunit MinK. In its rare homozygous forms, it can lead to Jervell and Lange-Nielsen syndrome.
LQT6 is an autosomal-dominant, relatively uncommon form of LQTS. It involves mutations in the gene KCNE2, which encodes for the potassium channel beta subunit MiRP1, constituting part of the IKr repolarizing K+ current.
Andersen-Tawil syndrome is an autosomal-dominant form of LQTS associated with skeletal deformities. It involves mutation in the gene KCNJ2, which encodes for the potassium channel protein Kir 2.1. The syndrome is characterized by LQTS with ventricular arrhythmias, periodic paralysis, and skeletal developmental abnormalities such as clinodactyly, low-set ears, and micrognathia. The manifestations are highly variable.
Timothy's syndrome is due to mutations in the calcium channel Cav1.2 encoded by the gene CACNA1c. Since the calcium channel Cav1.2 is abundant in many tissues, patients with Timothy's syndrome have many clinical manifestations, including other congenital heart disease, autism, immune deficiency and complex syndactyly.
This newly discovered variant is caused by mutations in the membrane structural protein, caveolin-3. Caveolins form specific membrane domains called caveolae in which, among others, the NaV1.5 voltage-gated sodium channel sits. Similar to LQT3, these particular mutations increase so-called 'late' sodium current, which impairs cellular repolarization.
This novel susceptibility gene for LQT is SCN4B encoding the protein NaVβ4, an auxiliary subunit to the pore-forming NaV1.5 (gene: SCN5A) subunit of the voltage-gated sodium channel of the heart. The mutation leads to a positive shift in inactivation of the sodium current, thus increasing sodium current. Only one mutation in one patient has so far been found.
GIRK4 is involved in the parasympathetic modulation of the heart. Clinically, the patients are characterized by only modest QT prolongation, but an increased propensity for atrial arrhythmias.
LQT14 is caused by heterozygous mutations in the CALM1 (Calmodulin 1) gene (114180) on chromosome 14q32.
LQT15 is caused by heterozygous mutations in the CALM2 (Calmodulin 2) gene (114182) on chromosome 2p21.
Jervell and Lange-Nielsen syndrome
Jervell and Lange-Nielsen syndrome (JLNS) is an autosomal-recessive form of LQTS with associated congenital deafness. It is caused specifically by mutation of the KCNE1 and KCNQ1 genes.
In untreated individuals with JLNS, about 50% die by the age of 15 years due to ventricular arrhythmias.
Romano-Ward syndrome is an autosomal-dominant form of LQTS not associated with deafness. The diagnosis is clinical and is now less commonly used in centres where genetic testing is available, in favour of the LQT number scheme given above.
LQTS is classified a form of channelopathy.
All forms of LQTS involve an abnormal repolarization of the heart, which causes differences in the refractory period of the heart muscle cells (myocytes). After-depolarizations (which occur more commonly in LQTS) can be propagated to neighboring cells due to the differences in the refractory periods, leading to re-entrant ventricular arrhythmias.
The so-called early after-depolarizations (EADs) seen in LQTS are believed to be due to reopening of L-type calcium channels during the plateau phase of the cardiac action potential. Since adrenergic stimulation can increase the activity of these channels, this is an explanation for why the risk of sudden death in individuals with LQTS is increased during increased adrenergic states (i.e., exercise, excitement), especially since repolarization is impaired. Normally during adrenergic states, repolarizing currents also are enhanced to shorten the action potential. In the absence of this shortening and the presence of increased L-type calcium current, EADs may arise.
The so-called delayed after-depolarizations are thought to be due to an increased Ca2+ filling of the sarcoplasmic reticulum. This overload may cause spontaneous Ca2+ release during repolarization, causing the released Ca2+ to exit the cell through the 3Na+/Ca2+-exchanger, which results in a net depolarizing current.
Drug-induced QT prolongation is usually a result of treatment by antiarrhythmic drugs such as amiodarone and sotalol or a number of other drugs that have been reported to cause this problem (e.g., cisapride). Some antipsychotic drugs, such as haloperidol and ziprasidone, have a prolonged QT interval as a rare side effect. Genetic mutations may make one more susceptible to drug-induced LQT. The antidepressant citalopram is also associated with an increased risk of long QT.
The diagnosis of LQTS is not easy since 2.5% of the healthy population has prolonged QT interval, and 10–15% of LQTS patients have a normal QT interval. A commonly used criterion to diagnose LQTS is the LQTS "diagnostic score", calculated by assigning different points to various criteria (listed below). With four or more points, the probability is high for LQTS; with one point or less, the probability is low. A score of two or three points indicates intermediate probability.
- QTc (Defined as QT interval / square root of RR interval)
- ≥ 480 ms - 3 points
- 460-470 ms - 2 points
- 450 ms and male gender - 1 point
- Torsades de pointes ventricular tachycardia - 2 points
- T wave alternans - 1 point
- Notched T wave in at least 3 leads - 1 point
- Low heart rate for age (children) - 0.5 points
- Syncope (one cannot receive points both for syncope and torsades de pointes)
- With stress - 2 points
- Without stress - 1 point
- Congenital deafness - 0.5 points
- Family history (the same family member cannot be counted for LQTS and sudden death)
- Other family members with definite LQTS - 1 point
- Sudden death in immediate family members (before age 30) - 0.5 points
Those diagnosed with LQTS are usually advised to avoid drugs that would prolong the QT interval further or lower the threshold for TDP. In addition to this, two intervention options are known for individuals with LQTS: arrhythmia prevention and arrhythmia termination.
Arrhythmia suppression involves the use of medications or surgical procedures that attack the underlying cause of the arrhythmias associated with LQTS. Since the cause of arrhythmias in LQTS is EADs, and they are increased in states of adrenergic stimulation, steps can be taken to blunt adrenergic stimulation in these individuals. These include administration of beta receptor blocking agents, which decreases the risk of stress-induced arrhythmias. Beta blockers are an effective treatment for LQTS caused by LQT1 and LQT2.
Genotype and QT interval duration are independent predictors of recurrence of life-threatening events during beta-blocker therapy. To be specific, the presence of QTc >500 ms and LQT2 and LQT3 genotype are associated with the highest incidence of recurrence. In these patients, primary prevention with use of implantable cardioverter-defibrillators can be considered.
- Potassium supplementation: If the potassium content in the blood rises, the action potential shortens, so increasing potassium concentration could minimize the occurrence of arrhythmias. It should work best in LQT2, since the hERG channel is especially sensitive to potassium concentration, but the use is experimental and not evidence-based.
- Mexiletine, a sodium channel blocker: In LQT3, the sodium channel does not close properly. Mexiletine closes these channels and is believed to be usable when other therapies fail. Theoretically, mexiletine could be useful for people with this form of LQTS, but the medication is currently under study for this application and its use is not currently recommended.
- Amputation of the cervical sympathetic chain (left stellectomy). This therapy is typically reserved for LQTS caused by JLNS, but may be used as an add-on therapy to beta blockers in certain cases. In most cases, modern therapy favors ICD implantation if beta blocker therapy fails.
Arrhythmia termination involves stopping a life-threatening arrhythmia once it has already occurred. One effective form of arrhythmia termination in individuals with LQTS is placement of an implantable cardioverter-defibrillator (ICD). Also, external defibrillation can be used to restore sinus rhythm. ICDs are commonly used in patients with fainting episodes despite beta blocker therapy, and in patients having experienced a cardiac arrest.
With better knowledge of the genetics underlying LQTS, more precise treatments hopefully will become available.
For people who experience cardiac arrest or fainting caused by LQTS and who are untreated, the risk of death within 15 years is around 50%. With careful treatment this decreases to less than 1% over 20 years.Those who exhibit symptoms before the age of 18 are more likely to experience a cardiac arrest.
Inherited LQTS is estimated to affect between one in 2,500 and 7,000 people.
The first documented case of LQTS was described in Leipzig by Meissner in 1856, when a deaf girl died after her teacher yelled at her. When the parents were notified of her death, they reported that her older brother, who also was deaf, died after a terrible fright. This was several decades before the ECG was invented, but is likely the first described case of Jervell and Lange-Nielsen syndrome. In 1957, the first case documented by ECG was described by Anton Jervell and Fred Lange-Nielsen, working in Tønsberg, Norway. Italian pediatrician Cesarino Romano, in 1963, and Irish pediatrician Owen Conor Ward, in 1964, separately described the more common variant of LQTS with normal hearing, later called Romano-Ward syndrome. The establishment of the International Long-QT Syndrome Registry in 1979 allowed numerous pedigrees to be evaluated in a comprehensive manner. This helped in detecting many of the numerous genes involved.
- "Long QT syndrome". Genetic and Rare Diseases Information Center (GARD) – an NCATS Program. 2017. Retrieved 14 December 2017.
- Morita H, Wu J, Zipes DP (August 2008). "The QT syndromes: long and short". Lancet. 372 (9640): 750–63. doi:10.1016/S0140-6736(08)61307-0. PMID 18761222.
- Ferri, Fred F. (2016). Ferri's Clinical Advisor 2017 E-Book: 5 Books in 1. Elsevier Health Sciences. p. 736. ISBN 9780323448383.
- Levine E, Rosero SZ, Budzikowski AS, Moss AJ, Zareba W, Daubert JP (August 2008). "Congenital long QT syndrome: considerations for primary care physicians". Cleve Clin J Med. 75 (8): 591–600. PMID 18756841.
- "Long QT Syndrome". NHLBI, NIH. Retrieved 14 December 2017.
- Ackerman, MJ; Priori, SG; Dubin, AM; Kowey, P; Linker, NJ; Slotwiner, D; Triedman, J; Van Hare, GF; Gold, MR (January 2017). "Beta-blocker therapy for long QT syndrome and catecholaminergic polymorphic ventricular tachycardia: Are all beta-blockers equivalent?". Heart rhythm. 14 (1): e41–e44. doi:10.1016/j.hrthm.2016.09.012. PMID 27659101.
Among patients who have experienced a LQTS-triggered cardiac event (arrhythmic syncope, arrhythmic syncope followed by seizures, or aborted cardiac arrest), the untreated natural history is grim, with >50% mortality at 15 years.
- Vincent, Jean-Louis; Abraham, Edward; Kochanek, Patrick; Moore, Frederick A.; Fink, Mitchell P. (2011). Textbook of Critical Care E-Book. Elsevier Health Sciences. p. 578. ISBN 1437715680.
- "Long QT syndrome - Symptoms".
- Thomson, Clare; Wright, Paul (2014-10-15). "Long QT syndrome". The Pharmaceutical Journal. 293 (7833). Retrieved 18 October 2014.
- Jáuregui-Lobera, Ignacio; Jáuregui-Garrido (February 2012). "Sudden death in eating disorders". Vascular Health and Risk Management. 8: 91–8. doi:10.2147/VHRM.S28652. PMC . PMID 22393299.
- Nakajima, T; Kaneko, Y; Kurabayashi, M (2015). "Unveiling specific triggers and precipitating factors for fatal cardiac events in inherited arrhythmia syndromes". Circulation journal : official journal of the Japanese Circulation Society. 79 (6): 1185–92. doi:10.1253/circj.CJ-15-0322. PMID 25925977.
- Hedley PL; Jorgensen P; Schlamowitz S; Wangari, Romilda; et al. (2009). "The genetic basis of long QT and short QT syndromes: a mutation update". Human Mutation. 30 (11): 1486–511. doi:10.1002/humu.21106. PMID 19862833.
- Goldman 2011, pp. 185
- Schwartz, P. J. (2006). "The Jervell and Lange-Nielsen Syndrome: Natural History, Molecular Basis, and Clinical Outcome". Circulation. 113 (6): 783–790. doi:10.1161/CIRCULATIONAHA.105.592899. ISSN 0009-7322.
- Torekov SS, Iepsen E, Christiansen M, et al. (2014). "KCNQ1 Long QT syndrome patients have hyperinsulinemia and symptomatic hypoglycemia". Diabetes. 63 (4): 1315–25. doi:10.2337/db13-1454. PMID 24357532.
- Shinnawi R, Gepstein L; Gepstein (September 2014). "iPS Cell Modeling of Inherited Cardiac Arrhythmias". Curr Treat Options Cardiovasc Med. 16 (9): 331. doi:10.1007/s11936-014-0331-4. PMID 25080030.
- Tristani-Firouzi M, Jensen JL, Donaldson MR, Sansone V, Meola G, Hahn A, Bendahhou S, Kwiecinski H, Fidzianska A, Plaster N, Fu YH, Ptacek LJ, Tawil R (2002). "Functional and clinical characterization of KCNJ2 mutations associated with LQT7 (Andersen syndrome)". J. Clin. Invest. 110 (3): 381–8. doi:10.1172/JCI15183. PMC . PMID 12163457.
- Shah, Dipak P.; Baez-Escudero, Jose L.; Weisberg, Ian L.; Beshai, John F.; Burke, Martin C. (2012). "Ranolazine Safely Decreases Ventricular and Atrial Fibrillation in Timothy Syndrome (LQT8)". Pacing and Clinical Electrophysiology. 35 (3): e62–e64. doi:10.1111/j.1540-8159.2010.02913.x. ISSN 0147-8389.
- Wang F, Liu J, Hong L, et al. (2013). "The phenotype characteristics of type 13 long QT syndrome with mutation in KCNJ5 (Kir3.4-G387R)". Heart Rhythm. 10 (10): 1500–6. doi:10.1016/j.hrthm.2013.07.022. PMID 23872692.
- Moric-Janiszewska E, Markiewicz-Łoskot G, Loskot M, Weglarz L, Hollek A, Szydlowski L (2007). "Challenges of diagnosis of long-QT syndrome in children". Pacing Clin Electrophysiol. 30 (9): 1168–1170. doi:10.1111/j.1540-8159.2007.00832.x. PMID 17725765.
- Schwartz PJ, Moss AJ, Vincent GM, Crampton RS (1993). "Diagnostic criteria for the long QT syndrome. An update". Circulation. 88 (2): 782–4. doi:10.1161/01.CIR.88.2.782. PMID 8339437.
- "QT Drug List by Risk Groups". Arizona Center for Education and Research on Therapeutics. Archived from the original on 2010-12-24. Retrieved 2010-07-04.
- Compton SJ, Lux RL, Ramsey MR, et al. (1996). "Genetically defined therapy of inherited long-QT syndrome. Correction of abnormal repolarization by potassium". Circulation. 94 (5): 1018–22. doi:10.1161/01.CIR.94.5.1018. PMID 8790040.
- "Genotype- and Phenotype-Guided Management of Congenital Long QT Syndrome". Current Problems in Cardiology. 38 (10): 417–455. 2013-10-01. doi:10.1016/j.cpcardiol.2013.08.001. ISSN 0146-2806.
- https://ars.els-cdn.com/content/image/1-s2.0-S0146280613001217-gr6.jpg. Missing or empty
- Tranebjaerg L, Bathen J, Tyson J, et al. (1999). "Jervell and Lange-Nielsen syndrome: a Norwegian perspective". American Journal of Medical Genetics. 89 (3): 137–46. doi:10.1002/(SICI)1096-8628(19990924)89:3<137::AID-AJMG4>3.0.CO;2-C. PMID 10704188.
- Jervell A, Lange-Nielsen F; Lange-Nielsen (July 1957). "Congenital deaf-mutism, functional heart disease with prolongation of the Q-T interval and sudden death". Am. Heart J. 54 (1): 59–68. doi:10.1016/0002-8703(57)90079-0. PMID 13435203.
- Romano C, Gemme G, Pongiglione R (September 1963). "Rare cardiac arrythmias of the pediatric age. II. Syncopal attacks due to paroxysmal ventricular fibrillation". Clin Pediatr (Bologna) (in Italian). 45: 656–83. PMID 14158288.
- Ward OC (April 1964). "A new familial cardiac syndrome in children". J Ir Med Assoc. 54: 103–6. PMID 14136838.
- Moss AJ; Schwartz PJ (2005). "25th Anniversary of the International Long-QT Syndrome Registry". Circulation. 111 (9): 1199–1201. doi:10.1161/01.CIR.0000157069.91834.DA. PMID 15753228.
- Goldman, Lee (2011). Goldman's Cecil Medicine (24th ed.). Philadelphia: Elsevier Saunders. p. 1196. ISBN 1437727883.