OBJECTIVE. Tuberous sclerosis complex is an autosomal dominant disorder in which hamartomas occur in several organs. Cardiac rhabdomyomas, the most common heart tumors of childhood, are well known to be associated with tuberous sclerosis complex. Our aim for this study was to characterize the incidence, progression, and clinical consequences of tuberous sclerosis complex–associated rhabdomyomas in a large cohort of patients with TSC1 and TSC2 genotypes.
PATIENTS AND METHODS. Patients (154) with tuberous sclerosis complex were evaluated, including clinical assessment, electrocardiography, and echocardiography. Mutations in TSC1 or TSC2 genes were identified in 127 patients.
RESULTS. Cardiac rhabdomyomas were found in 74 (48%) patients. Tumors were most frequent in children younger than 2 years (65%). Tumor regression or disappearance was observed in 37 (68%) of 55 children. However, in 6 (3.9%) of them (aged 10-15 years), cardiac rhabdomyomas were noted to either grow (3 cases) or appear de novo (3 cases), such that the frequency of cardiac rhabdomyomas in adolescents was 6 (54%) of 11. Most (61%) tumors were clinically silent. Clinical manifestations included heart failure (5.4%), arrhythmias (23%), and murmurs (14.9%). One child died as a result of cardiac insufficiency. Cardiac rhabdomyomas were more frequent in theTSC2 (54%) than TSC1 (20%) groups.
CONCLUSIONS. Cardiac rhabdomyomas are seen in the majority of young children with tuberous sclerosis complex. Most produce no clinical consequences and will spontaneously regress. However, during puberty, cardiac rhabdomyomas may enlarge or appear de novo; thus, attention should be paid to potential clinical signs and monitoring by echocardiography should be performed. Cardiac rhabdomyomas were observed more often in the TSC2 group.
Tuberous sclerosis complex (TSC) is an autosomal dominant disorder characterized by the development of distinctive benign tumors in multiple organ systems, including skin, brain, heart, lungs, kidney, and liver. It affects ∼1 in 6000 individuals but is considered to be as frequent as 1 in 6800 in children.1 Cardiac involvement is relatively common in patients with TSC, usually in the form of cardiac rhabdomyomas (CRs). Although often clinically silent, these lesions can cause arrhythmias and cardiac failure2 and, in one series, have been the most frequent cause of death among <10-year-old children with TSC.3
With the advent of echocardiography, cardiac tumors have been diagnosed more frequently. The incidence of primary cardiac tumors among children presenting to pediatric cardiac referral centers is 0.20%4 and is 0.27% among pediatric autopsies.5 Rhabdomyoma is the most common primary cardiac tumor in infancy and childhood, representing 36% to 42% of tumors in autopsy5–7 and 79% in clinical series.4
It is currently thought that the majority of children with CRs have TSC. In a review of all cases of CRs published up until 1990, Harding and Pagon8 found that at least 172 (51%) of 335 cases were associated with TSC. Data regarding TSC manifestations in an additional 117 cases were insufficient to confirm the diagnosis of TSC, but if one includes these 117 possible TSC cases, 86% of CRs were related to TSC. The highest proportion of TSC among CRs was reported by Bosi et al,9 who documented TSC in 30 (91%) of 33 patients with CRs. Multiple rhabdomyomas especially are regarded as characteristic for TSC. Tworetzky et al10 diagnosed TSC in 7 of 23 children with a single ventricular tumor and 61 of 64 patients with multiple lesions. Conversely, CRs are found in 47% to 67% of all patients with TSC.11–13
However, despite the long recognition of a relationship between CRs and tuberous sclerosis, there have been no previous studies that have assessed their incidence and natural history with mutational analysis of the TSC genes in a large number of patients. Such a detailed characterization is our intent for this report.
MATERIALS AND METHODS
From 1986 to 2004, a cohort of 154 children (72 girls and 82 boys) with a diagnosis of TSC participated in this prospective study. The patients were recruited from the departments of neurology, cardiology, and neonatology of the Children's Memorial Health Institute in Warsaw, Poland. All the patients or their legal representatives provided informed consent for participation in the study. The diagnosis of TSC was established in all patients by standardized examination using diagnostic criteria established at the Tuberous Sclerosis Consensus Conference.14
All subjects had a clinical evaluation that included physical examination, echocardiography, electrocardiography, and a chest radiograph. Cardiac ultrasound examinations were performed using a Hewlett-Packard (Andover, MA) machine with spectral Doppler facilities. At study entry, the age of the patients with TSC ranged from 0 to 18 years. In 3 patients, cardiac tumors were found on prenatal sonography, and they were included in this study after birth. Follow-up examinations were performed in 55 patients. The control sonography was proposed every second year to all children with TSC unless the patients required quick evaluation. The follow-up period ranged from 0.5 to 18 years, with a median of 3.5 years.
In 127 patients, TSC gene-mutation identification was performed as described.15
For statistical analyses, Fisher's exact (for cells with size ≤5) and χ2 tests were used. The level of significance was set at P < .05.
Cardiac tumors were found on initial echocardiographic examination in 74 (48%) of 154 patients with TSC. The incidence of cardiac tumors, however, depended on the patients' age (Table 1). CRs were seen at the highest frequency in those <2 years of age (66%) and were significantly less common in those between 2 and 11 years of age (26% [P < .0001]). CR incidence then seemed to rise during puberty, because those between 12 and 15 years of age had a 54% incidence, but not significantly different in comparison to those between 2 and 11 years old (P = .057). It is interesting to note that the difference in incidence was more evident in girls, because in this subgroup the difference between adolescents and children between 2 and 11 years old was significant (P = .015).
Multiple CRs were found in 33 (45%) patients with cardiac tumors. They were more frequent in children under 2 years of age (57%) than in older patients (41%).
Serial echocardiographic examinations were performed in 55 cases, including 38 patients in whom the first examination showed ≥1 CR. Follow-up ranged from 0.5 to 18 years (median: 3.5 years). Of these 38 patients, tumors disappeared in 7 (18%) and decreased in 19 (50%) patients (Fig 1). In 3 (8%) cases tumors were larger at the follow-up examinations, whereas in 9 patients (24%) they did not change in size. Of the 17 patients without CRs on the first echocardiographic examination, tumors appeared in 3 (18%) of the cases (Fig 2).
The ages of the patients with TSC in whom new tumors were identified were 10, 12, and 14 years. In addition, the 3 children in whom CR tumor growth was seen in serial follow-up studies were aged 11, 12, and 15 years. Of the 6 patients with new or growing tumors, 5 were girls (Table 2). None of them received hormonal contraception. All patients with TSC in whom CR growth was evident by echocardiography were in the pubertal age. The overall rate of new tumor appearance plus CR growth was 4 (36%) of 11 children aged 12 to 15 years. In contrast, CR growth was not seen at all in the youngest children (0 of 64 children aged <2 years; P < .006) and was rarely seen in children older than 2 to 11 years (2 of 69 [2.9%]; P = .028). CRs were identified throughout the heart, including every cavity. The tumors were most commonly found in the right ventricle (RV) (35%), the left ventricle (LV) (22%), the interventricular septum (33%), the left atrium (5%), and the right atrium (5%). Of those found in the RV, 1 tumor produced LV outflow tract obstruction.
Most CRs (45 of 74 cases [61%]) had no apparent clinical manifestations. Clinical features that were seen included cardiac failure (4 patients in the group <2 years of age, 5.4% of all patients with CRs), arrhythmias (17 patients [23%]), and murmurs (11 patients [14.9%]). Arrhythmias were both atrial (12 patients) and ventricular (5 patients). Four patients required antiarrhythmic therapy. Of those, 1 patient died, and in other 3 children the CRs regressed and the antiarrhythmic agents might have been discontinued. In 1 patient with normal echocardiographic examination results, electrocardiographic evidence of Wolff-Parkinson-White syndrome was seen. This finding disappeared during follow-up. Only 1 child in this series died as a result of rhabdomyoma. This child had multiple tumors in all heart cavities and presented with ventricular dysfunction from birth. She died at the age of 3 months. Chest radiograph anomalies (cardiomegaly) were noted in 4 children with symptomatic cardiac failure.
Genetic analysis to identify mutations in TSC type 1 (TSC1) or 2 (TSC2) was performed for 127 patients. TSC1 mutations were found in 15 (12%) and TSC2 mutations in 93 (73%) patients, whereas in 19 (15%) patients no mutation could be identified. Cardiac tumors were seen in 50 (53.8%) of 93 patients with TSC2 mutations, in comparison to 3 (20%) of 15 of those with TSC1 mutations (P < .0001) (Table 3). In the group with no mutation identified, CRs were seen in 7 (36.8%) of 19.
The tumors in patients with TSC2 mutation produced more severe manifestations than in those with TSC1 mutation. Heart insufficiency was seen in 4 patients with TSC2 mutation and in none with TSC1 mutation (P < .001). One child with TSC2 mutation died. Arrythmias seemed to be more frequent in the TSC2 than the TSC1 group (15 [16.1%] and 2 [13.3%], respectively), but the difference was not significant. One patient with TSC1 mutation (6.7%) and 10 patients with TSC2 mutation (10.7%) presented with murmurs; the difference was also not significant.
Analysis of the frequency of CRs by age according to genotype is difficult because of the small number of patients in this series with TSC1 mutations. However, similar to observations on the entire set of patients, patients with TSC2 mutation aged 0 to 2 years had the highest frequency of CRs (29 of 36 [80%]), which dropped significantly in the group of 2- to 11-year-olds (15 of 43 [35%]; P < .0001).
Five patients with TSC1 mutation underwent follow-up echocardiography. Tumor regression was observed in 1 patient, complete resolution was observed in another patient, and in 3 other patients without the CRs on first echocardiography, the results of the examination remained the same. In the TSC2 group, follow-up examination was performed for 36 patients. New tumor appearance was noted in 2 children, and in 2 other patients the existing tumors had enlarged. Tumor regression or resolution was observed in 14 and 3 cases in the TSC1 and TSC2 groups, respectively. In 15 patients, the result of echocardiography remained unchanged.
Here we have documented the incidence, natural history, and outcome of cardiac tumors in patients with TSC in the largest series yet reported and provided a comparison of these features with TSC1 versus TSC2 mutation.
We confirmed that these tumors are seen at their highest frequency in very young children (66% in children <2 years of age). In the majority of patients, there is partial (50%) or complete (18%) resolution of CRs during serial follow-up assessment by echocardiography.
In contrast to previous reports, we have demonstrated that cardiac tumors may grow and/or appear de novo in adolescents with TSC. Because of the significance of this observation, we carefully reviewed the echocardiograms of the 6 patients in whom this was seen. All echocardiographic examinations were performed by the same physician under the same conditions and using the same standards. Moreover, in the 3 patients in whom CRs appeared when none had been seen previously, the same echocardiographic device was used for the initial and follow-up examination. On the basis of these findings, it seems reasonable to perform echocardiography on all adolescents with TSC. However, because of variable dynamics of tumor evolution, it is difficult to establish universal timing of follow-up echocardiography in patients with TSC.
The natural history of CRs is intriguing and quite distinct from that of any other manifestation of TSC. However, another TSC lesion, pulmonary lymphangioleiomyomatosis, is seen exclusively in females, implicating estrogen as a critical factor in stimulating the growth of the smooth muscle–like cells that comprise lymphangioleiomyomatosis. Similarly, we hypothesize that maternal transplacental estrogen stimulation accounts for the observation that CRs are seen at their highest frequency and severity in the first months of life.12 The proliferating aberrant myocytes that make up rhabdomyomas are distinctive in their massive accumulation of glycogen and show evidence for complete loss of either TSC1 or TSC2 with high expression of phospho-S6.16 This molecular signature of activation of mammalian target of rapamycin (mTOR) is similar to that seen in many other TSC lesions, consistent with the function of the TSC1/TSC2 proteins in regulating the activation of mTOR.17 Extending this reasoning further, we postulate that the growth/appearance of rhabdomyomas in the small fraction of patients with TSC during puberty reflects the effects of pubertal hormonal changes.
CRs have important diagnostic implications. When discovered prenatally or at birth, CRs are rarely associated with other manifestations of TSC.18,19 In infants, the skin lesions of TSC may not be found even with a Wood light examination, and brain imaging studies (computed tomography or MRI) in early infancy have limited resolution and require general anesthesia. Echocardiography is perhaps the most useful single diagnostic test for TSC in this age group.18 The absence of visible cardiac lesions does not exclude the presence of the disease. One should be always aware that because of echocardiographic resolution limitations, very small tumors may be missed. Recently, MRI was proven to provide valuable information about cardiac tumors in fetuses and children.7,20 It should be noted that early detection of cardiac tumors in fetuses allows consideration of further pregnancy management.19,21
CRs may be detected in any of the cardiac chambers. In the majority of patients with TSC, the tumors are 5 to 15 mm in diameter. They appear with similar incidence or slightly more frequently on the left side of the heart than on the right12,23,24 and more commonly in the ventricles than in the atria.12,23 In a large study of 47 patients with TSC and CRs, Nir et al25 noted tumors in the LV in 32 patients (68%) (in the LV septum in 38% and the LV apex in 38%) and the RV in 31 patients (66%) (in the RV septum in 38% and RV apex in 32%). Eighteen patients (38%) had tumors located in the LV only, whereas 14 patients (30%) had tumors in the RV only, and 15 patients (32%) had tumors in both the LV and RV. Right atrial tumors were present in 3 patients (6%). In this study, CRs localized predominantly in the RV (35.3%), and the right atrium was the least likely area affected.
In patients with TSC, myocardial involvement and replacement of the ventricular muscle by noncontractile tumor tissue may mimic a cardiomyopathy. In our retrospective studies of children with TSC, we had 6 patients in whom the diagnosis of cardiomyopathy was established in the first months of life.12 After subsequent echocardiographic studies the diagnosis was abandoned, probably because of the regression of the tumor. Fortunately, intramural lesions rarely produce clinical signs. The prognosis for this group of patients is better than for patients with intracavitary tumors. In our study, most CRs were asymptomatic.
Congestive heart failure develops in 2% to 4% of children with CRs.12,25 Only 1 of 47 patients reported by Nir et al25 had the symptoms of heart failure. It is interesting to note that a cardiac murmur may be heard in a relatively small proportion of patients. We detected it in 11 (10%) of 108 children with cardiac masses.
The majority of descriptions in the literature of patients with CRs leading to heart failure are related to masses in the left heart. One of our patients with multiple rhabdomyomas located mainly in the LV outflow tract died at the age of 3 months. Five of 6 infants with TSC and symptomatic CRs reported by Webb et al26 died as a result of cardiac failure; all 6 had large LV lesions.
There is a common opinion that the rhythm abnormalities in patients with TSC are caused by intramural rhabdomyomas interrupting the conduction pathways and leading to ectopic electrical foci or an accessory electrical circuit producing preexcitation (Wolff-Parkinson-White syndrome). Nir et al25 documented Wolff-Parkinson-White syndrome in at least 2 (9%) of 23 patients with rhabdomyomas who had electrocardiograms. In our patients, the frequency of Wolff-Parkinson-White syndrome was not as high (0.65%) but still higher than in the general population (0.15%).
This is also the first report on the phenotype-genotype correlations in cardiac tumors in TSC. We found that CRs are significantly more frequent among patients carrying TSC2 mutation than those carrying TSC1 mutation. Moreover, in patients with TSC2 mutation, the cardiac tumors produced more severe symptoms. Heart insufficiency was seen in patients with TSC2 mutation only, and 1 of them resulted in a patient's death.
CRs are seen in most young children with TSC. They are more frequent and produce more severe symptoms among patients with TSC2 mutation in comparison to those with TSC1 mutation. These cardiac tumors tend to regress with age but can grow or appear during puberty. In young children, CRs may be the sole clinical symptom of TSC but do not by themselves make a diagnosis of TSC. Thus, mutational analysis of the TSC genes should be considered. Prenatal echocardiography will often detect rhabdomyomas in fetuses with TSC and provide important information for parents and care providers. Because of the growth of CRs in some patients with TSC at puberty, echocardiography should be performed for detection.
- Accepted May 4, 2006.
- Address correspondence to Sergiusz Jóźwiak, MD, PhD, Department of Neurology and Epileptology, Children's Memorial Health Institute, Aleja Dzieci Polskich 20, 04-730 Warsaw, Poland. E-mail: or
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Smith HC, Watson GH, Patel RG, Super M. Cardiac rhabdomyomata in tuberous sclerosis: their course and diagnostic value. Arch Dis Child.1989;64 :196– 200
- ↵Muhler EG, Turniski-Harder V, Engelhardt W, von Bernuth G. Cardiac involvement in tuberous sclerosis. Br Heart J.1994;72 :584– 590
- ↵Roach ES, Gomez MR, Northrup H. Tuberous Sclerosis Complex Consensus Conference: revised clinical diagnostic criteria. J Child Neurol.1998;13 :624– 628
- ↵Meikle L, McMullen JR, Sherwood MC, et al. A mouse model of cardiac rhabdomyoma generated by loss of TSC1 in ventricular myocytes. Hum Mol Genet.2005;14 :429– 435
- ↵Kwiatkowski DJ, Manning BD. Tuberous sclerosis: a GAP at the crossroads of multiple signaling pathways. Hum Mol Genet.2005;14(spec No. 2) :R251– R258
- ↵Webb DW, Thomas RD, Osborne JP. Cardiac rhabdomyomas and their association with tuberous sclerosis. Arch Dis Child.1993;68 :367– 370
- Copyright © 2006 by the American Academy of Pediatrics