OBJECTIVE. Barth syndrome, an X-linked disorder that is characterized by cardiomyopathy, neutropenia, skeletal myopathy, and growth delay, is caused by mutations in the taffazin gene at Xq28 that result in cardiolipin deficiency and abnormal mitochondria. The clinical phenotype in Barth syndrome has not been characterized systematically, and the condition may be underrecognized. We sought to evaluate extent of cardioskeletal myopathy, potential for arrhythmia, delays in growth, and biochemical correlates of disease severity in patients with this disorder.
METHODS. We conducted an observational, cross-sectional study of the largest cohort of patients with Barth syndrome to date (n = 34; age range: 1.2–22.6 years). Evaluation included echocardiography, electrocardiography (standard and signal-averaged), microvolt T wave alternans analysis, biochemical and hematologic laboratory analyses, and physical therapy evaluation of skeletal myopathy.
RESULTS. Family history was positive for confirmed or suspected Barth syndrome in 63%. Ninety percent of patients had a clinical history of cardiomyopathy (mean age at diagnosis of cardiomyopathy: 5.5 months; at genetic confirmation of Barth syndrome: 4.6 years). Echocardiography revealed a mean ejection fraction of 50% ± 10%, mean fractional shortening of 28% ± 5%, and mean left ventricular end-diastolic volume z score of 1.9 ± 1.8. Left ventricular morphology demonstrated increased trabeculations or true noncompaction in 53%. Of 16 patients who were evaluated at ≥11 years of age, 7 (43%) had documented ventricular arrhythmia. Growth deficiency was present (mean weight percentile: 15%; mean height percentile: 8%). Laboratory analysis revealed low total white blood cell count (absolute count: <4000 cells per μL) in 25% of those who were not on granulocyte colony-stimulating factor. Hypocholesterolemia was present in 24%, decreased low-density lipoprotein cholesterol in 56%, low prealbumin in 79%, and mildly elevated creatine kinase in 15%.
CONCLUSIONS. Our cohort demonstrated clinical variability, but most had cardiomyopathy and diminished growth velocity, with a propensity toward neutropenia and low cholesterol. There was increased incidence of ventricular arrhythmia, predominantly in adolescents and young adults. Barth syndrome should be considered when boys present with cardiomyopathy, especially when associated with increased left ventricular trabeculations, neutropenia, skeletal muscle weakness, or family history indicating an X-linked pattern of inheritance.
Barth syndrome (BTHS) is an X-linked recessive disorder that is caused by mutations in the taffazin gene (TAZ) at Xq28, leading to severe cardiolipin deficiency in the mitochondrial membrane. First described in 1983,1 BTHS is characterized by cardiomyopathy, neutropenia, skeletal myopathy, growth deficiency, and 3-methylglutaconic aciduria (which is not specific to BTHS). Accurate incidence rates are unknown. The most common presentation is cardiomyopathy in infancy, although skeletal myopathy or neutropenia-related infection may be noted first. Cognitive function usually is normal, although some cognitive difficulties have been reported.2
Clinical features of BTHS have included left ventricular (LV) dilation, hypertrophy, and noncompaction, with varying degrees of congestive heart failure and endocardial fibroelastosis.1,3,4 Some patients have required cardiac transplantation.5 Sudden cardiac death5 and ventricular arrhythmia (VA)6 have been reported. Other features include neutropenia, which may be present at birth and often is cyclic.
Biochemical abnormalities have included reports of low plasma cholesterol.4 Skeletal myopathy, although not yet fully characterized, is a frequent feature of BTHS. Patients frequently have gross motor delay,5,7,8 and Gower's sign may be present.5 Motor and sensory nerve conduction studies tend to be normal.7,8
The clinical phenotype in BTHS has not been shown to correlate with genotype,9 although it has not been characterized systematically. To describe more definitively the spectrum of disease that is seen in this condition, we sought to evaluate the extent of cardiac and skeletal myopathy, potential for arrhythmia, delays in growth, and biochemical correlates of disease severity in a cohort of patients with BTHS.
We conducted an observational, cross-sectional study of 34 patients with BTHS (age range: 1.2–22.6 years) between 2002 and 2004. Participants were recruited from our medical center's cardiomyopathy clinic and through the Barth Syndrome Foundation. Any patient with a diagnosis of BTHS and confirmed TAZ mutation was eligible for the study. Eighteen patients were evaluated at study onset, with 15 returning for 2-year follow-up in addition to 16 new patients who were evaluated in 2004. Initial evaluation included echocardiography and electrocardiography (ECG); subsequently, a more comprehensive analysis was done to include standard and signal-averaged ECG (SA-ECG), microvolt T wave alternans (TWA) analysis, biochemical and hematologic laboratory analyses, and evaluation of skeletal myopathy by a physical therapist. All prospective study patients agreed to participate after informed consent, which was approved by the University of Florida Institutional Review Board.
TAZ was amplified from genomic DNA in 3 segments with polymerase chain reaction (PCR) primer pairs F7 CTTCCCGTTTCCTCCCGTTC/R5 CCAGGGCTCCATGAAAAGGC; F1471 ACTGGGCTGGTTGCAGTGACAG/G1 GCTTGAGTGATCCTCTCACCTC; G2.7 GGAGAAGGGCCTGTTTCATTGAG/3′R AGCTCGGAGAGGGCACTTGAG. The fragments were sequenced with PCR primers and internal primers, using the Big Dye 3.1 system (Applied Biosystems, Foster City, CA) and an ABI Prism 377 sequencer. MacVector version 7.1.1 (Accelrys, San Diego, CA) was used for sequence storage. AssemblyLIGN was used for comparing sequences to the normal reference sequence.
To analyze cardiomyopathy genotype-phenotype correlation, we used ejection fraction (EF) and LV internal diastolic dimension (LVIDd) z scores as measures of the severity of cardiomyopathy. Six pairs of relatives with BTHS and the same mutation were evaluated. We measured variance within genotype using the pairs of a common genotype and estimated between-genotype variance from the unpaired genotypes.
All participants underwent echocardiography in 2002 and 2004 using a protocol that was designed to evaluate LV size, morphology, and systolic and diastolic function. All echocardiograms were recorded digitally and in VHS format, and each study was interpreted and measured by the same cardiologist. Blood pressure, height, and weight were recorded for each study participant. Echocardiograms in 2002 were obtained using Phillips 5500 echocardiography machines and in 2004 using Phillips 7500 machines (Phillips Medical Systems, Bothell, WA). The interpreting cardiologist was blinded to the 2002 results. All measurements and derived indices were expressed as z scores relative to normal values for body surface area or age.10 The z-score value indicates the position of each measurement relative to the normal population expressed as SDs from the population mean. LV dimensions were measured from M-mode tracings. LV mass, volumes, and EF were obtained from apical and short-axis images and calculated using the 5/6 area-length method. LV noncompaction (LVNC) morphology was determined separately by 2 cardiologists who were blinded to the other's interpretation. Both readers graded patients as having morphology consistent with LVNC, no evidence of LVNC, or possible LVNC. The ratio of noncompacted to compacted LV myocardium was measured at end systole at the site of maximal thickness by 1 of the interpreting cardiologists as described previously.11
Evaluation of cardiac conduction and rhythm included 12-lead ECG, SA-ECG, and TWA performed according to standard clinical practice. Eight patients completed a 24-hour Holter monitor. All ECGs, 24-hour Holter monitors, SA-ECGs, and TWAs were supervised and interpreted by a single cardiologist. The SA-ECG and TWA analyses were included in this protocol because these tests have been shown to predict arrhythmic events in adults with dilated cardiomyopathy (DCM).12,13 TWA has not been validated in children, but normal values for pediatric patients have been published.14 SA-ECG as a predictor of arrhythmia has not been well studied in children with cardiomyopathy; it may predict events in children with Duchenne muscular dystrophy15 and surgically repaired heart disease.16 Normal values for children have been published.17
The 12-lead ECG was performed at rest and included the standard-limb and augmented-limb leads and precordial leads V3R, V1, V2, V4, V6, and V7. Corrected QT interval (QTc) was calculated using Bazett's formula. SA-ECG was performed on those who were 6 years or older and evaluated for the presence of late potentials, with an abnormal SA-ECG defined as abnormalities in the duration of the high-frequency, low-amplitude signals (<40 μV), filtered total QRS duration, and the root-mean-squared voltage in the terminal 40 milliseconds.17
TWA was performed according to previously published guidelines14,18 in those who were 10 years or older. Recording of the TWA was done at rest and during a state of increased heart rate, using a stationary bicycle to achieve a heart rate of ∼70% of the predicted maximum or >120 beats per minute (bpm). The electrodes were placed in Frank orthogonal configuration and standard 12-lead placements. The TWA measurements were made with a CH2000 system (Cambridge Heart Inc, Bedford, MA). Results of TWA were evaluated on the basis of the presence or absence of sustained alterations in T wave morphology at rest and with exercise. The criteria for a positive TWA at rest was alternans voltage of >1.0 μV and alternans ratio of >3 and with exercise criteria for a positive test was sustained alternans voltage of >1.9 μV and alternans ratio of >3 for at least 1 minute in any orthogonal lead or 2 adjacent precordial leads at heart rates of 90 to 110 bpm.
Biochemical and Hematologic Analysis
All laboratory analyses were performed according to standard clinical practice. Not all patients consented to have blood drawn. Lipid profile was obtained after study patients fasted overnight. Analyses included total triglycerides and cholesterol; high-density lipoprotein; calculated low-density lipoprotein (LDL); prealbumin; complete blood count; and biochemical markers for myocyte damage, including total creatine kinase (CK), and CK myocardial band (CK-MB) and troponin T. Brain natriuretic peptide (BNP) was measured as a marker of LV dysfunction.19
Evaluation of Skeletal Myopathy
A pediatric physical therapist assessed patients who were 5 to 22 years of age. Hand grip strength was evaluated with a Jamar Hand Dynamometer (Sammons Preston, Bolingbrook, IL), elbow flexors and knee extensors with a MicroFET 2 handheld dynamometer, and ankle dorsiflexors with the MicroFET 2 dynamometer (Hoggan Health Industries Inc, Draper, UT) and an IsoBex muscle strength analyzer (IsoBex 3.0; Hersteller: MDS Medical Device Solutions AG, Burgdorf, Switzerland). Age-appropriate normative data are available for grip strength for ages ≥6 years20; therefore, z scores were generated for each patient's grip strength.
Analysis that compared groups of patients used a 2-sided Wilcoxon test (Kruskal-Wallis test when >2 groups). The Spearman correlation coefficient was used to assess the association between pairs of quantitative variables. For assessment of change in a quantitative variable, the 2-sided Wilcoxon signed-rank test was used. Comparison of proportions was done by Fisher's 2-sided exact test (2 groups) or by the Freeman-Halton exact test (>2 groups). For assessment of the clustering effect within genotypes, variation within the genotype (pairs of like genotypes) was compared with the between-genotype variation (single occurrences) via an F test.21–23 Results are presented as mean ± SD unless otherwise noted; P values of <.05 were considered statistically significant.
Of the original 18 study participants who were seen in 2002, 3 did not return for evaluation in 2004 (2 were alive and well, and 1 had a sudden cardiac death). In 2004, the evaluation was expanded and included the 15 returning patients from 2002 in addition to 16 new patients. Mean age at the time of evaluation in 2002 was 10.2 ± 5 years (range: 1.2–20.5 years) and in 2004 was 10.9 ± 6.2 years (range: 1.7–22.6 years). Medical therapy primarily included heart failure medications and granulocyte colony-stimulating factor (GCSF) (Table 1).
Diagnosis and Family History
Of the 31 patients who were evaluated in 2004, 3 did not have a clinical diagnosis of cardiomyopathy. Of the remaining 28, mean age at diagnosis of cardiomyopathy was 5.5 ± 7.4 months; 1 received a diagnosis of cardiomyopathy by prenatal evaluation. Mean age at genetic confirmation of BTHS was 4.6 ± 4.3 years. Family history was obtained for 30 patients (1 child was adopted). The family history was positive for confirmed BTHS or cardiomyopathy in boys with a clinical picture of BTHS in 19 (63%) of 30, and 15 (50%) had a family history of death related to confirmed BTHS or cardiomyopathy in young boys.
Mutational analysis was available for 30 of 31 patients (1 child had a confirmed mutation at Xq28, but the specific mutation was not provided). This included 11 stop and 7 frameshift mutations, 6 amino acid substitutions, 4 splice variants, 1 amino acid deletion, and 1 complete gene deletion. (The extent of the deletion is not known. However, the presence of several other nearby genes was detected by PCR, including creatine transporter 1, MECP2, and emerin.) Mutations in this group included mutations in exons 1 (n = 2), 2 (n = 9), 3 (n = 2), 4 (n = 1), 6 (n = 3), 8 (n = 2), 9 (n = 2), 10 (n = 3), and 11 (n = 1) and introns 1, 2, 6, and 10 (n = 1 each).
No significant clustering effect within genotype was demonstrated (P = .43 and .87 for EF and LVIDd z score, respectively). Patients also were analyzed with respect to mutation location, because early work on TAZ and its products had suggested the existence of a taffazin form that begins in exon 3, which would be normal in patients with mutations in exons 1 and 2.24 More recent work has failed to find this variant.25 EF and LVIDd z score were not significantly different for those with exon 1 or 2 mutations (n = 11) versus those with mutations in the rest of the gene (P > .75 for both measures). Others have suggested that mutations in exon 8 may cause a more severe phenotype.26 Only 2 patients in this study had exon 8 mutations, both with EF in the normal range.
Most patients manifested growth deficiency, although a wide range of growth percentiles was noted. Of the 26 patients who were ≤18 years of age and were evaluated in 2004, the mean weight equated to the 15th percentile (range: <1%–66%), and 15 (58%) were below the 5th percentile in weight for age. Likewise, evaluation of height in this age group revealed that the mean height equated to the 8th percentile (range: <1%–38%) with 15 (58%) at or below the 5th percentile. BMI was calculated to assess weight for height. Results demonstrated that 44% of patients had a BMI <5th percentile, 48% were within normal limits, and 7% had a BMI >95th percentile. Gastric tubes were used for supplemental feeds in 4 of 31 individuals. Patients who were older than 18 years were plotted on the growth chart at an age of 18 years. The 5 patients who were older than 18 years had a persistently low mean weight of 13th percentile (range: <1%–63%) but a mean height of 50th percentile (range: 8%–90%), suggesting a delayed height growth spurt in the group. A compiled growth chart for 5 patients who were older than 16 years is presented in Fig 1.
Of the 31 individuals who were evaluated with echocardiography in 2004 (Table 2), images were adequate for measurement of LV volumes and EF in 30 patients and M mode for LV dimensions and shortening fraction (SF) for 26 patients. The mean EF was 50% ± 10% (range: 29%–67%) and mean SF of 28% ± 5% (range: 13%–34%). Sixteen (53%) had normal EF (EF ≥51%; z score ≥−1.9), with 10 of 16 being maintained on at least 1 cardiac medication. There were no significant correlations between age and EF, SF, SF z score, LVIDd z score, or LV end-diastolic volume z score (data not shown). The mass/volume ratio z score was within normal limits for 28 (93%). One individual had a mass/volume ratio z score of 4.5 and the clinical appearance of concentric hypertrophic cardiomyopathy. Two had mass/volume ratio z scores at the lower limits of normal (z score = −1.9), and 1 had an abnormally low z score (−2.8). Of the 18 patients who were evaluated by echocardiography in 2002, 14 had a repeat echocardiogram in 2004 that was adequate for comparative quantitative measurements. There were no significant differences in EF (Fig 2), SF, SF z score, or LV end-diastolic volume z score during the 2-year period.
Many patients were found to have qualitatively prominent trabeculations of the LV even when they did not meet published criteria for LVNC11 and were considered to be in the LVNC group by both readers. Thirty patients had echocardiograms that were sufficient to make this determination, and both readers agreed that 16 (53%) had an abnormal morphology of the LV with increased trabeculations. Of the remainder, both readers agreed that 7 (23%) had normal LV morphology but differed in opinion about the LV morphology pattern in 7 (23%), including 1 patient with concentric hypertrophic cardiomyopathy. To determine the extent of noncompaction, we measured the end-systolic ratio of noncompacted to compacted LV myocardium. A ratio of >2 has been used as a diagnostic criterion for LVNC.11 This revealed that the mean ratios were 2.2 ± 0.12, 1.34 ± 0.11, and 1.52 ± 0.24 in the groups with both readers agreeing that abnormal morphology was present, absent, and differed in opinion, respectively. There was no difference in QTc (444 vs 425 milliseconds), EF (48% vs 49%), or LVIDd z score (1.9 vs 2.2) between those with and without LVNC morphology, respectively.
Standard 12-lead ECG was available for review in 29 patients: 17 (58%) had abnormal findings, 7 (24%) had borderline findings, and 5 (17%) were normal. Abnormal findings included left axis deviation (n = 3), LV hypertrophy with strain (n = 4), repolarization abnormalities (n = 17), and right bundle branch block (n = 1). Repolarization abnormalities predominantly included ST flattening or T-wave inversion. Normal sinus rhythm was present in 27 of 29, with 1 patient each having first-degree heart block and ectopic atrial rhythm. Analysis of the QTc interval in 30 patients demonstrated 6 (20%) with prolonged QTc of ≥460 milliseconds and an additional 7 (23%) with borderline QTc prolongation between 450 milliseconds and 459 milliseconds. Of the 8 patients with 24-hour Holter monitors available for review in 2004, 2 were normal; 1 demonstrated significant ventricular ectopy (220 premature ventricular beats per hour with couplets); and the others had findings that included poor heart-rate variability (n = 2), sinus bradycardia with an episode of nonsustained atrial tachycardia (n = 1), and repolarization abnormalities including T-wave inversion and ST segment depression (n = 2). SA-ECG analysis was available in 20 patients, with 15 (75%) being normal, 2 borderline, and 3 abnormal. TWA was performed in 18 patients. Two had an abnormal TWA, both with an implantable cardioverter defibrillator (ICD) and documented VA.
Of 34 patients in the cohort, 7 had documented VA, all older than 11 years (n = 16). These cases included cardiac arrest that resulted in brain death (n = 1), resuscitated sudden death followed by ICD placement (n = 1), VA on locally performed 24-hour Holter monitor (n = 4), and electrophysiology study alone demonstrating inducible VA (n = 1; 5 cases previously reported6). Table 3 demonstrates the method of documentation of arrhythmia and results of recent screening tests that were performed in these individuals. All 6 living patients with previously documented VA currently have an ICD. Table 4 demonstrates echocardiography findings in those with and without documented VA. Although there was a slight trend in those with documented VA toward a larger LV size and more depressed systolic function, this was not statistically significant. In addition, there was no relationship between mutation location and presence of VA.
Table 5 demonstrates pertinent laboratory values. Evidence of neutropenia was found in 4 (25%) of 12 of those in whom neutrophil counts were obtained, and low total white blood cell (WBC) count was present in 5 (20%) of 25, with normal hematocrit and platelets in all individuals. When patients on GCSF are excluded from the analysis, 25% had low total WBC counts. Among patients who were on GCSF, the WBC count tended to be higher than that noted for those who were not on GCSF, but this was not statistically significant (8800 vs 5500 cells per μL; P = .26), and none with an absolute neutrophil count <1000/μL were on GCSF. There was no correlation of total WBC with EF (Spearman r = 0.27, P = .19) and no difference in EF in those who were on GCSF (47% vs 50% respectively; P = .4).
Fasting lipid analysis demonstrated normal triglycerides in all patients and low total cholesterol in 24%. Although LDL was low in >50%, the high-density lipoprotein was more variable, being low in 16% and high (≥75 mg/dL) in 3 (12%). There was no correlation between EF and total cholesterol or LDL. There was no correlation between prealbumin and weight percentile, total cholesterol, or EF. Total CK was mildly elevated in 3 patients, but the range of elevation (192–397 U/L) was very small. Elevated BNP correlated with decreased EF (Spearman r = 0.53, P = .009).
All patients ambulated independently, although 2 older patients used mobility scooters to conserve energy. Many patients had a history of delayed motor milestones, and nearly all reported problems with fatigue in daily life. Grip strength was compared with published age-appropriate normal values,20 with a group mean z score for the right and left hand of −2.9 ± 1.0 and −2.1 ± 0.9, respectively. There was a significant trend toward normalcy with increasing age (Fig 3).
Our study, the largest to date to evaluate systematically a cohort of patients with BTHS, has yielded a detailed analysis of the spectrum of cardiac and clinical characteristics that are associated with the condition, providing a more objective description of this disease. Although the clinical phenotype typically involves cardiomyopathy, cyclic neutropenia, skeletal myopathy, and growth deficiency, these features can vary significantly, likely resulting in underdiagnosis of this disorder. Nearly two thirds of families have a positive family history of confirmed or suspected BTHS, but often this diagnosis is made in retrospect after genetic confirmation of the proband. In particular, our findings could lead to increased awareness and diagnosis of BTHS in patients without a clear family history of the disease.
Disease-causing mutations have been found in all exons of TAZ (G4.5 at Xq28).25 Previous reports indicated the lack of a genotype-phenotype correlation on the basis of qualitative assessment of cardiac symptoms.9 This study evaluated LV size and function (EF and LVIDd z scores) as quantitative measures of the severity of cardiomyopathy to compare individuals within and between genotype. There was no genotype-phenotype correlation with regard to LV size or function. Initial cloning of TAZ suggested multiple alternative spliced products, including “short” TAZ mRNA beginning in intron 2 that could lead to a smaller protein product.24 Recent analysis of TAZ mutations indicated that there are 2 functional gene products and no “short” mRNA.25 Our analysis confirms that those with mutations in exons 1 and 2 do not have a less severe cardiac phenotype. It is likely that important unidentified modifying genetic factors influence the phenotype and the degree of cardiomyopathy in this disorder.
Three of 31 patients (ages 3, 5, and 17 years) who were evaluated in 2004 had no previous individual clinical history of cardiomyopathy by parental report. None of these 3 patients had been on cardiac medications, and all had normal LV dimensions, volume, and EF. None of these patients had the same genotype, and all 3 have relatives with BTHS and clinically significant cardiomyopathy on cardiac medications. Although >50% of the remaining patients in this study had EF and LV diastolic volume z scores within normal limits, most of these individuals are on cardiac medications, and all have had cardiac dysfunction in the past. This suggests that the vast majority of patients with BTHS have clinically significant cardiomyopathy, but determining the true prevalence of cardiomyopathy in this population is difficult. By clinical history, many individuals have systolic function that is variable over time. Several individuals gave a history of cardiomyopathy followed by normalization of EF after initiation of medical therapy. After discontinuation of cardiac medications, a significant decline in EF occurred, and reinitiation of medications resulted in improvement of the LV function.
Of 10 patients who were evaluated at age ≥15 years, 8 received a diagnosis of cardiomyopathy at age ≤2 years, and 1 has never received a diagnosis of cardiomyopathy. In addition, there was no significant change in the indexed LV size or systolic function in 2 years in the 14 patients with adequate serial echocardiograms. In the cohort as a whole, there was not an age-related decline in LV systolic function. These data suggest that in many patients with BTHS, cardiomyopathy often is responsive to standard heart failure medical therapy and may remain stable for many years. However, individuals with BTHS and severe cardiomyopathy that leads to death or requires heart transplantation have been observed frequently. Clinical observations and the detailed cardiac and genetic evaluation in this study have failed to identify specific markers that might predict a risk for severe cardiomyopathy.
The type of cardiomyopathy also varied. Although most individuals had DCM, half had features of prominent LV trabeculations suggesting a form of LVNC, and 1 had hypertrophic cardiomyopathy. In addition, some hearts have been noted to “remodel,” having features of LVNC in infancy that change over time. One patient with marked LVNC has a brother with the same genotype but a hypertrophic cardiomyopathy phenotype. Another patient had a diagnosis of LVNC at birth (EF <20%) and now at 6 years of age has had remodeling of the heart without LVNC morphology and has an EF of 55% (Fig 4). The clinical characteristics of LVNC in children were reviewed recently, and an undulating phenotype was described.27 In addition, LVNC was reported to be “acquired” in patients with mitochondriopathy and Duchenne muscular dystrophy.28,29 Therefore, the percentage of those who have BTHS and have LVNC in a cross-sectional sample may be misleading. A difference in the functional severity of cardiomyopathy between those with and without LVNC morphology could not be demonstrated.
We observed an increased incidence of VA, primarily in older individuals. Specific risk factors for VA could not be identified. All patients with a history of VA had an abnormal ECG, but this did not discriminate between those with and without VA. There also was an increased incidence of prolonged QTc, but this did not predict those with VA. The positive and negative predictive values of SA-ECG and TWA are not relevant with low numbers of positive tests, although both patients with positive TWA had proven arrhythmia. This population is small, longitudinal data do not exist, and very few of these patients have had intracardiac electrophysiologic studies. The true incidence of potential serious arrhythmia therefore is not known, and the predictive value of these screening tests remains to be determined.
The overall incidence of VA in children with idiopathic DCM has not been studied carefully, but ventricular tachycardia has been noted in 11% to 18% of children with idiopathic DCM in retrospective studies,30,31 and the presence of VA has not been proven to be a risk factor for death. In children and young adults with BTHS, potential mechanisms of arrhythmia related to mitochondrial disease include apoptosis32 and absence of cardiolipin affecting mitochondrial lipid-protein interactions and altering mitochondrial ion channels that are important for cardioprotection during ischemia and ischemia-related arrhythmia.33
Similar to previous reports, neutropenia was common but not universally present in this cohort. Hematologic and cardiac phenotypes are not likely linked in severity as evidenced by a lack of correlation between WBC count or GCSF therapy and EF. Other findings included hypocholesterolemia in some patients and low prealbumin in most. Although low prealbumin often is a marker of poor nutrition and many of these patients are growth deficient, a direct relationship between prealbumin and weight percentile could not be demonstrated. Neither hypocholesterolemia nor low prealbumin could be shown to be related to the severity of the cardiomyopathy.
Analysis of CK, CK-MB, and troponin did not reveal biochemical evidence of active muscle cell injury. A few patients had total CK elevations, but the degree of abnormality was small. Although the sample size was small, these laboratory values are not likely useful for diagnosis or evaluation in BTHS. BNP, however, may be useful in monitoring the severity of LV dysfunction, because this was proportional to the decrease in EF. In adults with DCM, BNP is increased in proportion to LV end-diastolic volume and inversely related to EF34 and also has been shown to be useful in evaluating ventricular dysfunction in children.19
Physical therapy assessment suggests that skeletal myopathy generally does not limit ambulation, although a history of delayed motor milestones often is present. There is some evidence that strength may improve with age, although it usually is not normal. Excessive fatigue is an important factor in the symptoms related to skeletal myopathy. Given the role of mitochondria in muscle bioenergetics, this finding is not surprising and supports the idea of impaired energy production in this condition, especially related to decreased endurance.
Results indicate that most patients have growth deficiency, with more than half being below the 5th percentile for height and/or weight. In addition, slightly fewer than one half of the group had BMI less than the 5th percentile, indicating relative poor weight gain in these individuals. However, growth failure is not a uniform finding, and nearly 25% of those who were younger than 18 years had weight above the 20th percentile and more than one half had a normal or increased BMI. Evaluation of patients in late adolescence and early adulthood indicates that some of these individuals may have delayed growth spurts and reach their true height at a later age.
The link between the mitochondrial and cellular functional abnormalities related to cardiolipin deficiency and the resulting morphologic and functional myocardial abnormalities is not well understood. Mitochondrial defects, including mutations in both nuclear and mitochondrial DNA, are associated with DCM and hypertrophic cardiomyopathy, cardiac conduction abnormalities, and sudden death. Cardiolipin is an essential component of the inner mitochondrial membrane, being important for mitochondrial structure, permeability, and function. It is strongly bound to many mitochondrial proteins, including cytochrome C.35 Apoptosis is triggered by release of intramitochondrial proteins, including cytochrome C,33 leading to activation of caspase and initiation of programmed cell death. Decreased cardiolipin synthesis is associated with cytochrome C release and apoptosis.36 Chronic, low-level apoptosis has been shown to cause DCM in a mouse model,37 and cardiomyocyte apoptosis is enhanced in human heart failure.38 Severe cardiolipin deficiency that leads to activation of caspase and increased myocardial cell death may be one mechanism of the development of cardiomyopathy in this disease. There may be some mechanistic similarity with doxorubicin-induced cardiomyopathy, given that doxorubicin binds tightly to cardiolipin and interferes with cardiolipin-dependent proteins.39
A few study limitations should be considered. Many patients with BTHS present in infancy with cardiomyopathy, but few infants were identified for evaluation in this study. The genetic diagnosis usually follows the diagnosis of cardiomyopathy by a few years. Therefore, a survival bias cannot be excluded, and the short- and long-term outcome of those who receive a diagnosis in infancy is not known. Moreover, the natural history of DCM in children is not fully described, and without more complete longitudinal data, it is not possible to know how the prognosis of this genetic cardiomyopathy compares with other forms of DCM.
The diagnosis of BTHS should be considered for boys who present with cardiomyopathy, especially when associated with increased LV trabeculations, neutropenia, skeletal muscle weakness, or a family history suggestive of an X-linked disorder. Our cohort will require long-term follow-up and additional patient enrollment to determine the natural history and long-term prognosis of those who are affected by BTHS.
This work was partially supported by General Clinical Research grant M01 RR00082 from the National Center for Research Resources and through funding from the Barth Syndrome Foundation, the Children's Miracle Network, and the Howard Hughes Medical Institute.
We thank Debra Pruett and Sharon Chapman for performing the echocardiograms; Jon Shuster, PhD, for assistance with statistical analysis; and Melanie Fridl Ross, MSJ, ELS, for editing assistance.
- Accepted March 2, 2006.
- Address correspondence to Carolyn T. Spencer, MD, Congenital Heart Center, University of Florida College of Medicine, PO Box 100296, Gainesville, FL 32610-0296. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Jenni R, Oechslin E, Schneider J, et al. Echocardiographic and pathoanatomical characteristics of isolated left ventricular non-compaction: a step towards classification as a distinct cardiomyopathy. Heart.2001;86 :666– 671
- ↵Hohnloser SH, Klingenheben T, Bloomfield D, et al. Usefulness of microvolt T-wave alternans for prediction of ventricular tachyarrhythmic events in patients with dilated cardiomyopathy: results from a prospective observational study. J Am Coll Cardiol.2003;41 ;2220– 2224
- ↵Stelling JA, Danford DA, Kugler JD, et al. Late potentials and inducible ventricular tachycardia in surgically repaired congenital heart disease. Circulation.1990;82 :1690– 1696
- ↵Freund R, Littell R. SAS System for Regression. New York, NY: John Wiley & Sons; 2000
- ↵Lehmann E, D'Abrera H. Nonparametrics: Statistical Methods Based on Ranks. San Francisco, CA: Holden-Day Inc; 1975
- ↵Barth PG, Valianpour F, Bowen VM, et al. X-linked cardioskeletal myopathy and neutropenia (Barth syndrome): an update. Am J Med Genet A.2004;126 :349– 354
- ↵Pignatelli RH, McMahon CJ, Dreyer WJ, et al. Clinical characterization of left ventricular noncompaction in children: a relatively common form of cardiomyopathy. Circulation.2003;108 :2672– 2678
- ↵Ostrander DB, Sparagna GC, Amoscato AA, et al. Decreased cardiolipin synthesis corresponds with cytochrome c release in palmitate-induced cardiomyocyte apoptosis. J Biol Chem.2001;276 :38061– 38067
- Copyright © 2006 by the American Academy of Pediatrics