Paediatric single mitochondrial DNA deletion disorders: an overlapping spectrum of disease

While Kearns and Sayre’s original clinical description was published in 1958 and Pearson’s was published in 1979 (Kearns and Sayre 1958; Pearson et al. 1979), knowledge that mtDNA deletions were responsible was only established in 1988 (Holt et al. 1988). Subsequently, many individual case reports and case series have been published that increasingly suggest a clinical overlap between these historically different phenotypes (Manea et al. 2009; Yamashita et al. 2008; Pitceathly et al. 2012; Grady et al. 2014; McShane et al. 1991).

All 34 patients described here had single mtDNA deletions: 11 (A–J and AG) presented with varying degrees of anaemia and haematological dysfunction and fulfilled diagnostic criteria for the Pearson marrow–pancreas syndrome. However, pancreatic exocrine dysfunction was only documented in four: A, B, G and I (Table S2). Even in those with pancreatic dysfunction, the dysfunction was first documented at least 2 years after the initial presentation with anaemia, providing further evidence that the presence of pancreatic exocrine dysfunction should not be considered as an essential diagnostic criterion for Pearson syndrome. Our findings are in agreement with previous observations that one third of Pearson syndrome patients present solely with anaemia (Manea et al. 2009), with pancreatic dysfunction at presentation in only 12.7 % of cases, although this increases to 18 % by the age of 4 years (Lee et al. 2007). Low birth weight has previously been reported to be the most common nonhaematological finding in Pearson syndrome, affecting up to 63 % of patients (Manea et al. 2009). While at first glance our cohort might appear to reinforce these observations, with four of ten being <3rd centile at birth, on closer examination, two of these patients (D and G) were born prematurely at 34 weeks’ gestation, while patients D and P were second twins.

A novel finding of our study is the extent of multisystem disease in patients with Pearson syndrome (Fig. 3). Historically, <20 % of patients were reported to have symptoms unrelated to bone marrow or gastrointestinal tract (Manea et al. 2009), but in our cohort, all Pearson syndrome patients had involvement of other systems. The renal tract was most commonly involved, with evidence of renal dysfunction in five Pearson syndrome cases, including Fanconi-type tubulopathy in two cases (G and H) and profound glomerular impairment in two (A and C). Renal involvement was not confined to those children with haematological problems, and overall, 17 of 20 (85 %) patients for whom results were available had abnormal glomerular and/or tubular dysfunction. The majority of patients had proximal tubulopathy, but other renal manifestations included glomerular compromise (A, K, L and W), nephrocalcinosis (diagnosed on autopsy in E and by ultrasound in W). End-stage renal failure occurred in one patient (K).

Our study demonstrates the relatively high frequency of renal disease in patients with mtDNA deletions. This is particularly interesting in light of previous observations of progressive renal failure in a murine model with mtDNA deletions (Inoue et al. 2000) and validates this model for preclinical trials of novel therapies for mtDNA deletion disorders. Furthermore, the utility of highly sensitive urinary markers of tubular damage, NAG and RBP (Herget-Rosenthal et al. 2004; Vaidya et al. 2008), allows presymptomatic detection of renal involvement and, consequently, early intervention. The presence of renal tubulopathy as detected by NAG/creatinine and RBP/creatinine ratios can also be a useful diagnostic clue and serve to increase clinical suspicion of an underlying mitochondrial disorder.

Ten of the 23 patients whose presentation was not secondary to anaemia fulfilled clinical diagnostic criteria for KSS. Patients U, X and AD were classified as having PEO and patients P, Q, Y, AA, AB, AE and AH as PEO+, since ptosis is recognised to precede ophthalmoplegia by several years (Jackson et al. 1995). In 15 of the 16 patients with ptosis, it was the only clinical symptom at initial presentation. Since SLSMDs can be detected noninvasively in children (on blood or urine analysis), we recommend that all children >1 year with undiagnosed new-onset ptosis should be screened for SLSMDs. It is to be noted, however, that ptosis in patient F was present at birth, and thus SLSMDs should be considered even in patients with apparent congenital ptosis, particularly if other systems subsequently become involved.

Cardiac involvement is well recognised in KSS and reported to affect almost 60 % of cases (Berenberg et al. 1977). Cardiac manifestations are less well known in Pearson syndrome (Rahman and Leonard 2000; Krauch et al. 2002; Akaike et al. 1997), but we observed cardiac involvement in three of 11 (27 %) children: two had rhythm disturbances: patient I, and patient B previously reported by Rahman et al. (Rahman and Leonard 2000); one (H) had RVH. Cardiac manifestations in other phenotypic subgroups were two KSS and three PEO+ patients (all had conduction defects), and three in the unclassified group (K and R had LVH; L had atrial fibrillation).

The spectrum of neurological presentations in childhood SLSMDs is constantly increasing (Morel et al. 2009; Lee et al. 2007). In the study reported here, the major documented neurological findings were generalised hypotonia and muscle weakness, with resultant delay in gross motor development in five patients. Neuromuscular symptoms were more common in patients without haematological impairment, possibly reflecting the high mortality rate in patients with Pearson syndrome, i.e. patients died before onset of neuromuscular symptoms. Early mortality of Pearson patients likely also explains the predominance of brain imaging abnormalities in patients without bone marrow manifestations. The most frequently observed brain MRI abnormalities were basal ganglia and white matter lesions. A notable exception was patient R, who appeared to show signs of impaired neuronal migration. To our knowledge, this is the first time an mtDNA deletion has been associated with imaging changes suggestive of a neuronal migration disorder, although other defects of mitochondrial oxidative phosphorylation (OXPHOS) function have been seen in patients with neuronal migration defects (van Straaten et al. 2005). As previously reported (McShane et al. 1991), we observed progression from Pearson syndrome to KSS in the two survivors who had long-term follow up.

Three patients, K, L and R, had no obvious external ophthalmoplegia or ptosis and, since they also did not have haematological compromise, fell outside any of the classical SLSMD phenotypes. They were diagnosed during investigation for other problems: faltering growth (K), Fanconi syndrome (L) and persistent neonatal lactic acidosis (R). Since patients with multisystem involvement are increasingly being investigated for mitochondrial disorders, it is likely that in the future a greater number of children will fall into this ill-defined group and new phenotypes of SLSMDs will emerge.

The range of clinical phenotypes was reflected in the varied need for symptomatic support. The need for careful endocrinology monitoring is evidenced by the requirement for hormone replacement in 12 of our cases: one required thyroxine, four cortisol replacement therapy, three vitamin D for hypoparathyroidism and three growth hormone. However, as might be expected, the most common endocrine abnormality was diabetes mellitus; five of the nine patients with abnormal glucose tolerance tests required insulin. Patients with impaired pancreatic exocrine function required pancreatic enzyme replacement. The other commonly compromised systems were cardiac (pacing was required in five patients) and renal: eight patients needed medical therapy (electrolyte replacement, in very high doses in some cases), one had renal replacement therapy while being considered for a transplant and one had lithotripsy to treat renal calculi. Finally, the importance of regular audiometry is emphasised by the finding of impaired hearing in ten patients and the need for cochlear implantation in one. This degree of multidisciplinary input, with contribution to management by specialist audiological physicians, cardiologists, endocrinologists, gastroenterologists, haematologists, nephrologists, neurologists, ophthalmologists and palliative care physicians, again emphasises the need for coordinated care of children with SLSMDs at a tertiary specialist centre.

Overall, our data provides further support for a clinical continuum of syndromes associated with SLSMDs manifesting in childhood. However, subdivision into Pearson syndrome and KSS may be useful for prognosis, as demonstrated by survival analysis of this cohort (Fig. 2). Five-year survival in patients manifesting with a Pearson phenotype was <50 % from time of initial presentation compared with 100 % for other phenotypic subgroups (P < 0.001). Although this mortality rate is high, it is considerably lower than a historical report of 76 % 5-year mortality (Rotig et al. 1995). Interestingly, survival at 4 years was 55 % in a more recent cohort (Manea et al. 2009), suggesting the possibility that earlier recognition of Pearson syndrome and intense management of known complications may be improving survival rates. Furthermore, all patients with Pearson syndrome who survived to 8 years were still alive at 18 years, which may be a useful prognostic feature when counselling parents and important information in planning transition to adult services.

Genotype–phenotype correlation for SLSMDs remains controversial. Traditionally, there was thought to be no relationship between the length of mtDNA deletion and clinical phenotype (Lopez-Gallardo et al. 2009; Rotig et al. 1995; Aure et al. 2007). However more recent work suggests that location (Lopez-Gallardo et al. 2009; Yamashita et al. 2008) and number of deleted tRNA molecules (Yamashita et al. 2008) may possibly influence phenotype, and a study of 87 patients (only five of whom presented younger than 20 years) indicated that mtDNA deletion size was correlated with both age at onset and progression rate. However, the impact of deletion size was mediated by the degree of heteroplasmy seen in muscle biopsy (Grady et al. 2014). It also suggested that deletion of the MT-CYB gene was associated with a more severe phenotype. In contrast, in our cohort, we found no correlation between patient age and deletion size or MT-CYB deletion. There was a significant, if minor, statistical correlation between degree of heteroplasmy and age at presentation. However, since quantitation was performed using visual inspection, this finding should be interpreted with caution.