Diagnostic criteria for respiratory chain disorders in adults and children

Adult RC diagnostic criteria.

The adult RC diagnostic criteria5 classify patients into a definite, probable, possible or unlikely RC disorder based on the presence of major and minor criteria for clinical features and laboratory markers of increased mitochondrial content, decreased RC function, molecular data and metabolic indicators of RC dysfunction (tables 1 and 2⇓). A definite diagnosis is defined as the identification of either two major criteria or one major plus two minor criteria. A probable diagnosis is defined as either one major plus one minor criterion or at least three minor criteria. A possible diagnosis is defined as either a single major criterion or two minor criteria, one of which must be clinical. Evidence from at least two relatively independent types of investigation (i.e., clinical, histologic, biochemical, or molecular) is required to establish a definitive diagnosis. This principle was also supported in a previous review of biochemical diagnosis of RC disorders.2

Patients.

We performed a retrospective review of the hospital, genetics, and laboratory records of all patients from the Royal Children’s Hospital in Melbourne referred to the mitochondrial diagnostic laboratory at the Murdoch Institute between January 1994 and January 1997. Of a total of 146 patients, 28 were excluded because of inadequate clinical records (11) or the subsequent confirmation of a nonmitochondrial diagnosis (17). The male/female ratio in the resulting study cohort of 118 patients was 1.4, and the median age at presentation was 10 months (range newborn to 16 years). At the time of the study, 48 patients (40%) had died; the median age at death being 29 months. Each patient’s available data were re-evaluated using the adult criteria. Agreement or disagreement with the assignment of individual criteria was monitored, and the overall certainty of diagnosis based on the adult criteria system was compared with the patient’s previously assigned assessment. Modifications were then proposed to correct systematic errors in the assignment of individual criteria and to improve the overall performance.

Tissues: histology, enzyme, functional, and molecular analysis.

Tissue samples were processed as described previously.6,7⇓ Histology reports from biopsies of skeletal muscle (58 patients), liver (33 patients), and heart (37 patients) were reviewed. RC studies were performed on 173 samples (76 skeletal muscles, 45 livers, 25 fibroblast cell lines, 14 lymphoblast cell lines, and 13 heart muscles) taken from 97 patients, 57 of whom had more than one tissue assayed. Complex I, II, II and III, III, IV, and citrate synthase (CS) activities were assayed in tissues and cell lines as described previously.6,7⇓ Enzyme complex activities were expressed as ratios relative to the CS activity and to the complex II activity for that sample. RC function was assessed by measuring rates of adenosine triphosphate synthesis in permeabilized skin fibroblasts and by determining whether patient fibroblasts could grow in a culture medium containing the slowly metabolized sugar galactose instead of glucose.7 DNA from patient tissues and cell lines was assessed for common mitochondrial DNA (mtDNA) point mutations (A3243G, A8344G, T8993G, T8993C) and mtDNA rearrangements, as described previously.6 DNA was extracted from 137 samples (35 skeletal muscles, 20 livers, 1 heart, 22 fibroblast cultures, and 59 lymphocyte or lymphoblast pellets) from 101 patients.

Clinical features.

Eleven (9.2%) of the patients met the major clinical criterion: eight had Leigh disease, one had Alpers disease, and two had mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) syndrome. Of the remaining 107 patients, 88 met the minor clinical criterion. The scoring of clinical features differed from our original interpretation in eight patients (7%). Details are provided in the supplementary table related to this article, which can be found on the Neurology Web site (go to www.neurology.org). Patient 1 presented at 4 weeks of age with vomiting and liver disease, which was followed by progressive neurodegeneration with choreoathetosis, neuropathy, optic atrophy, external ophthalmoplegia, and growth retardation leading to her demise at the age of 4 years. This patient received the same score with the adult criteria as a patient with only a single clinical feature (e.g., cardiomyopathy), highlighting a limitation in the ability of a system based on only major or minor criteria to weigh quantitative and continuous data.

Patients 2 through 8 did not receive any clinical score with the adult criteria but had one or more features consistent with a possible mitochondrial disorder with onset in utero,8 the neonatal period,9 or childhood1 (see details in the online table on the Neurology Web site). The following additional features were identified: unexplained hydrops fetalis or neonatal/infant collapse/death, stillbirth associated with a paucity of intrauterine movement and correlated with thin long bones with multiple fractures, isolated non-neurologic involvement such as liver failure, significant failure to thrive, movement disorder, hypotonia or hypertonia, and significant developmental delay.

Histology.

None of the 67 skeletal muscle biopsies showed true ragged red fibers. Four patients had marked subsarcolemmal accumulation of mitochondria (SSAM; see Patients 9 through 12 in the online table) on electron microscopy (EM), but these findings do not warrant any score in the adult criteria. In children with RC defects, true ragged red fibers are a rare finding, and SSAM is much more common,7 so the scoring of histology differed from our original interpretation in these four patients.

Light microscopy showed abnormal liver pathology in 20 of the 33 patients biopsied; 10 of whom had microvesicular steatosis, whereas the remainder had nonspecific pathologic abnormalities. Only three patients (all with microvesicular steatosis) had significant abnormalities on EM, including an increase in mitochondrial number, abnormal stacking of the cristae, pallor of the matrix, and loss of mitochondrial granules, as previously described.10 These three patients met the minor histology criterion in the adult criteria. Of the 37 heart biopsies reviewed, only one patient with Barth syndrome had widespread EM abnormalities of the mitochondria meriting a minor histology criterion.

Enzyme activity.

Of the 59 patients with skeletal muscle submitted for histochemistry, four had significantly abnormal results on COX histochemistry (one with absent COX activity and three with a mosaic staining pattern). All four met the major enzyme criterion.

An additional 14 patients met the major enzyme criterion of less than 20% residual enzyme activity and 10 patients met the minor enzyme criterion of 20 to 30% residual activity. A number of concerns with this criterion were identified, and the scoring differed from our interpretation for six patients (Patients 13 through 18 in the online table on the Neurology Web site). The results of RC studies for the following two patients exemplify these concerns. Patient 13 was a 10-month-old boy seen for a history of failure to thrive, hypotonia, nystagmus, and lactic acidemia. His residual complex I activity was 30 to 40% in skeletal muscle and fibroblasts and 40% to 50% in liver and lymphoblasts (figure 1). None of the individual results meet the criteria for a minor enzyme code (i.e., 20% to 30% activity), but collectively they appear to warrant at least a minor enzyme code. Patient 14 presented in infancy with cyclical vomiting, cerebellar ataxia, and lactic acidosis. Residual activity of complex IV in fibroblasts and skeletal muscle was just over 20% (figure 2), meriting a minor enzyme code. However, given the consistency of the results, we would interpret them as warranting a major enzyme code.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Respiratory chain enzyme profiles for skeletal muscle (black), liver (gray), lymphoblasts (angled stripes), and fibroblasts (dotted) from Patient 13. Bars represent residual enzyme activities expressed as percentages of control means relative to citrate synthase (CS) activity. Complex III activity was not measured in the liver sample. Complex I CS ratios are all greater than 30%, so the data do not score either a major or minor enzyme criterion, despite complex I activity being less than all other RC enzymes in all four tissues.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Respiratory chain enzyme profiles for skeletal muscle (black) and fibroblasts (dotted) from Patient 14. Bars represent residual enzyme activities expressed as percentages of control means relative to citrate synthase (CS) activity. Complex IV CS ratios are marginally greater than 20% in both tissues so the data score a minor enzyme criterion, despite the complex IV activity being consistently low in both tissues.

Significant levels of residual activity are common in RC disorders, and the enzyme activities are prone to secondary effects,4 making the delineation of circumscribed boundaries difficult. For example, the results of the complex I to CS ratio for all the skeletal muscle biopsies reveal no clear cut-off between normal and abnormal activities (figure 3), and therefore the choice of less than 20% and 20 to 30% as major and minor criteria is arbitrary. Altogether, the scoring of enzyme data by the adult criteria differed from our interpretation in a total of six patients since it did not allow for additivity of results from more than one tissue or for the more robust nature of data from patient cell lines.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Residual activity of complex I CS ratios in the 66 skeletal muscle biopsies analyzed in this patient series. The values are expressed as a percentage of the normal control mean and arranged in order of increasing activity. Note that the data appear continuous without any obvious cut-off between normal and abnormal values.

Molecular analysis.

The A3243G mutation was identified in blood of two adolescents with typical features of MELAS. This mutation merits a major molecular criterion because it meets the six indicators of pathogenicity outlined by the authors of the adult criteria.5 Since completing this review, we have identified pathogenic mtDNA mutations in two of the complex I–deficient patients (C3303T in Patient 12 and G14459A in a patient with Leigh disease) and pathogenic nuclear gene mutations in one complex IV–deficient patient (SURF1 mutations in Patient 14, see figure 2) and in the patient with Barth syndrome (BTHS gene). The relatively low yield of mtDNA mutation analysis in our pediatric population is in agreement with previous reports.11-13⇓⇓

Metabolic evaluation.

Twenty-six patients received a minor metabolic code for elevated lactate in blood or CSF. Our evaluation of this criterion did not identify any specific problems. However, marginal elevations received the same weight as significant and persistent elevations. The sensitivity and specificity of metabolic investigations could potentially be improved by more comprehensive metabolic screening such as a daily lactate and pyruvate profiles,14 exercise testing,15 or functional imaging studies such as magnetic resonance spectroscopy or PET.5

Functional studies.

Functional studies do not receive any score in the adult criteria, resulting in our disagreement with the scoring in two patients. Fibroblasts from Patient 19 had only 15% of the control mean value for complex I–linked ATP synthesis.7 Fibroblasts from patients 19 and 20 also failed to grow in galactose medium. Relevant functional studies used in other laboratories include oxygen consumption,16,17⇓ fibroblast lactate/pyruvate ratio,18 and oxidation of radiolabeled substrates.19