CLINICAL PATHOLOGY: AN UPDATE

A few facts about the basic biology of mitochondria are helpful for the clinician.² Mitochondria are small compartments within every nucleated cell and are the principal source of adenosine triphosphate (ATP). ATP is essential for all active cellular processes, so it is not surprising that a relative deficiency of ATP can lead to the dysfunction of many different organs, and ultimately cause cell death if it is severe and prolonged. Neurons, muscle, and endocrine organs are particularly dependent on ATP, explaining why the brain, heart, and pancreas are regularly involved in mitochondrial disorders, along with skeletal muscle. One of the most puzzling features of mitochondrial disease is that some patients present with a slowly progressive course clinically similar to neurodegenerative disorders; others with a relapsing, remitting encephalopathy, yet others with abrupt onset stroke-like episodes. At present we do not have a clear explanation for this, and different patterns can be seen in the same individual at different times.

ATP is produced by the mitochondrial respiratory chain linked to oxidative phosphorylation, which is a collection of about 100 proteins clustered on the inner mitochondrial membrane into discrete complexes. Reducing equivalents (electrons) are passed from NADH and FADH₂ to complex I and II, and are passed on to complex III (cytochrome b) and IV (cytochrome c oxidase), pumping protons (H+) from the mitochondrial matrix into the intermembrane space in the process. This generates an electrochemical gradient which is harnessed by complex V (ATP synthase) to synthesise ATP from adenosine diphosphate (ADP).

A negative genetic test must be interpreted with great caution because the causative mutation may only be present in some tissues and not others

Most mitochondrial proteins are synthesised from genes in the cell nucleus, but 13 essential respiratory chain peptides are synthesised within the mitochondria themselves from small circles of DNA—the 16.5 Kb mitochondrial genome, or mtDNA. MtDNA also codes for 24 ribonucleic acids which form part of the protein synthetic machinery within the mitochondrial matrix. Each cell contains between a few hundred and many thousands of copies of mtDNA, which are inherited exclusively down the maternal line. The nuclear encoded proteins include most of the respiratory chain subunits (including all subunits of complex II, or succinate dehydrogenase), the mitochondrial DNA polymerase (polγ, encoded by the nuclear gene POLG1) and numerous other proteins required for the maintenance of mtDNA, mtDNA expression, the assembly of the respiratory chain, and the translocation of nuclear encoded proteins into the mitochondria. This means that mitochondrial disorders can be autosomal recessive, autosomal dominant, X-linked, or maternally inherited.³

Pathogenic mutations of mtDNA fall into two groups: deletions (where regions of the molecule are absent, often involving multiple genes) and point mutations (affecting a single base pair). The point mutations either directly affect the genes coding for the respiratory chain polypeptides, or the RNA genes themselves.⁴ All mutations are thought to cause disease primarily because they impair oxidative phosphorylation and thus lead to relative ATP deficiency. Some mtDNA mutations affect all the molecules in a cell or individual, and are thus called homoplasmic—for example, the mutations that cause Leber hereditary optic neuropathy, or maternally inherited non-syndromic deafness. By contrast, many patients with mtDNA mutations have a mixture of normal (wild-type) and mutated mtDNA within each mitochondrion, and are called heteroplasmic—for example, deletions of mtDNA found in chronic progressive external ophthalmoplegia, the Kearns-Sayre syndrome, the common 3243A>G mtDNA tRNA LeuUUR mutation first described in a patient with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS), the 8344A>G mtDNA tRNA Lys mutation found in myoclonic epilepsy with ragged red fibres (MERRF), or the 8993T>C/G ATPase 6 gene mutation found in neurogenic weakness with retinitis pigmentosa (NARP).

In general, human cells can tolerate high levels of mutated mtDNA before there is a biochemical defect of the respiratory chain, probably because the remaining wild-type molecules work overtime, compensating for the mutated molecules. However, once a critical threshold is exceeded, oxidative phosphorylation becomes defective. High proportions of mutated mtDNA are associated with a more severe clinical phenotype, but the relationship is complex. The proportion of mutated mtDNA can vary between different organs in the same individual, and also between cells in the same organ. This contributes to the clinical variability, and changing levels of mutated mtDNA probably contributes to clinical progression. It is not totally clear whether the biochemical defect arises solely from the high proportion of mutated mtDNA, the absolute level of mutated mtDNA, or a deficiency of wild-type mtDNA. Both random and non-random processes change the level of mutated and wild-type mtDNA over time, and there is some evidence that this is regulated by nuclear genes that have yet to be identified.⁵ Heteroplasmy poses a particular problem when investigating a suspected mitochondrial disorder if the proportion of mutated mtDNA is so low that it is undetectable in some tissues, particularly blood. A negative genetic test must be interpreted with great caution because the causative mutation may only be present in some tissues and not others.

Some nuclear genetic defects cause secondary abnormalities of mtDNA which can be either qualitative (secondary deletions or secondary point mutations) or quantitative (loss of mtDNA—called mtDNA depletion).³ These secondary defects are thought to accumulate with time, contributing to the progressive phenotype of these disorders. Since the nuclear defect causes an mtDNA abnormality, the phenotype can be very similar to primary mtDNA disorders, and includes the classical mitochondrial phenotype of chronic progressive external ophthalmoplegia which can be transmitted as a recessive or dominant trait.

CLINICAL PRESENTATIONS

Some patients with mitochondrial disease present with a classical syndrome (table 1). These syndromes are well known and described in most textbooks and review articles.^3,4,6 Unfortunately the clinical diagnosis is often not clear cut. Three groups pose a particular challenge:

• patients presenting with one or two components of a more common syndrome (so called oligosymptomatic subjects, such as the young individual with an occipital ischaemic stroke, or isolated fatigue and muscle cramps developing in middle age, or mild bilateral ptosis without ophthalmoplegia in an elderly individual)

• patients with multiple organ involvement that do not slot neatly into a recognised category

• patients who probably do not have a mitochondrial disorder, but it is difficult to be sure.

As neurologists, we tend to see patients with a primarily neurological phenotype. This can be highly variable, and include central neurological features (fluctuating or progressive encephalopathy, stroke-like episodes, epilepsy or myoclonus) or peripheral neuromuscular features (such as ptosis, ophthalmoplegia, proximal myopathy, or an axonal sensorimotor neuropathy). Common extra-neurological features include cardiomyopathy and cardiac conduction defects, endocrine dysfunction (particularly diabetes mellitus), and gastrointestinal complications which have been underrecognised in the past (recurrent vomiting and pseudo-obstruction can be presenting features, and dysphagia is common later in the disease course). The combination of unexplained neurological features with one or more extra-neurological features should raise the question of a mitochondrial disorder. But given the high prevalence of cerebrovascular disease, diabetes, and cardiac dysfunction, this question is regularly raised when the diagnosis is not a mitochondrial disorder at all.

AN EXPANDING PHENOTYPE OF NUCLEAR GENE DEFECTS IN ADULTS

The phenotype associated with POLG1 mutations is currently being described. Dominant and recessive mutations were first described in families with autosomal dominant or recessive chronic progressive external ophthalmoplegia,⁷ some with profound sensory ataxic neuropathy.⁸ Some of the dominant families have additional features including parkinsonism and primary gonadal failure.⁹ And children, with recessive mutations in POLG1, may present with an encephalopathy and liver failure due to depletion of mtDNA (the Alper-Huttenlocher syndrome).¹⁰ These individuals appear to be particularly sensitive to sodium valproate, which may precipitate the hepatic dysfunction—hence the importance of reaching a diagnosis in children with unexplained encephalopathy. Recessive POLG1 mutations have also been found in adults with late onset cerebellar ataxia,¹¹ particularly in Scandinavia, probably due to a founder effect in a geographically isolated population.¹² In some of these patients, epilepsy and headache are the presenting features, often persisting for many years before the more complex phenotype evolves.¹² This poses a particular diagnostic challenge because in the early stages patients with POLG1 mutations may have the kind of common neurological problem seen every day in the general neurology clinic with no family history, only developing a more complex phenotype after 10 or more years.

CLINICAL INVESTIGATION OF SUSPECTED MITOCHONDRIAL DISORDERS

The investigation of suspected mitochondrial disease has been discussed extensively in the past, and readers are referred to recent reviews on the topic.¹ But are there any “rules of thumb” that can help? After taking a detailed personal and family history (paying particular attention to “common” phenotypes such as premature cardiac death, diabetes, and relatives who apparently died from “multiple sclerosis”), followed by a neurological and general clinical examination, the next step is to look for subclinical multisystem involvement to accurately and better describe the phenotype. An electrocardiogram to show evidence of a cardiomyopathy or a conduction defect (fig 1), an echocardiogram to assess myocardial thickness and function, and a fasting blood glucose to diagnose diabetes are mandatory, and a formal retinal assessment can be revealing, identifying a pigmentary retinopathy that is not immediately apparent on direct ophthalmoscopy. Blood and CSF lactate measurement are often carried out, but rarely help because further investigation is usually pursued whatever their values. Arguably lactate levels should be dropped from the standard panel of tests performed on adults with suspected mitochondrial disease. Biochemical and molecular confirmation is ultimately required for optimal management, but the descriptive clinical tests add weight to the subsequent laboratory approach, particularly if the first batch of molecular tests is negative.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

12-lead electrocardiogram showing heart block in a patient with the Kearns-Sayre syndrome (courtesy of Dr Andrew Scheafer)

LABORATORY INVESTIGATION OF SUSPECTED MITOCHONDRIAL DISORDERS: WHEN AND HOW SHOULD YOU START?

Diagnosing a mitochondrial disorder can be quick and easy, but it can also be protracted and expensive. It is useful to come off the fence before ordering the first test. Do you really think it is likely to be mitochondrial disease a priori? This is important because it will influence how thorough the diagnostic approach should be. Both disease prevalence and the clinical picture are fundamental to this issue, and guide further investigations.

The current best estimate of disease prevalence is about one in 5000 of the population,¹³ and a simple consideration of the prevalence of the different phenotypes and molecular defects can help structure the investigation algorithm:

• In adults, the most common mitochondrial disease is Leber hereditary optic neuropathy, affecting about one in 14,000 adult males.¹⁴ This has a characteristic clinical presentation with subacute bilateral painless visual failure predominantly affecting men in young adult life (fig 2). This is not, of course, always the case. We have recently diagnosed a man presenting in his 70s, and up to 30% present without a clearcut family history. Diagnosis is usually straightforward and 95% of cases can be confirmed with a blood test sent from the outpatient clinic and carried out by many molecular genetics laboratories throughout the United Kingdom. The phenotype is striking and usually clearcut—the main diagnostic challenge lies with other mitochondrial diseases.

• The next most common clinical group are patients with ptosis and chronic progressive external ophthalmoplegia (present in approximately one fifth of adult onset mitochondrial disorders). This may be isolated and subtle, or associated with additional neurological or systemic features as in the Kearns-Sayre syndrome. Most cases do not have a family history and the age at presentation is wide, from childhood to late adult life. The previous dogma that diplopia is uncommon in this group is not true. Transient diplopia affects about one third of cases, although—unlike myasthenia—this does not usually fluctuate on a daily basis.¹⁵ Dysphagia is relatively common, but is generally not an early feature nor as severe as that seen in oculopharyngeal muscular dystrophy. Patients usually need a muscle biopsy because the molecular defect cannot reliably be detected in blood.

• The next most common group are patients with the 3243A>G tRNA LeuUUR mutation. Although first described in a Japanese patient with MELAS, the phenotype is extremely diverse, including isolated diabetes and deafness, hypertrophic cardiomyopathy, retinitis pigmentosa, or a fluctuating encephalopathy with epilepsy and stroke-like episodes. The 3243A>G mutation is found in at least one in 6000 of the general population, and appears to be much more prevalent in the hospital diabetic population, those with unexplained cardiomyopathy, and patients with unexplained ischaemic stroke presenting before 45 years of age (fig 3).¹⁶ Is it therefore reasonable to test for the 3243A>G mutation in these patient groups before doing a muscle biopsy? Simply on epidemiological grounds, testing is sensible. It is the most common pathogenic heteroplasmic mtDNA point mutation, but the result must be interpreted with great care. The percentage level of 3243A>G in blood is almost always less than the level in muscle and other clinically relevant tissues. Longitudinal studies have shown that the level in blood decreases over time sometimes falling below the detection threshold for most standard diagnostic tests. However, this paints a rather pessimistic view, and in most clinically affected individuals, 3243A>G can be detected in blood. There is also a much closer correlation between the percentage level of mutated mtDNA in urinary epithelium (obtained from a fresh urine sample) and the level in muscle, providing an alternative clinic based test.¹⁷ It is therefore reasonable to screen for 3243A>G in blood provided the investigations do not stop there. If the result is 0negative, other tissues should be studied, and urinary epithelium and cheek scrapings (buccal mucosa) provide sensible alternatives before proceeding to a muscle biopsy. Finally, it is important to remember that other point mutations of mtDNA can also cause the same phenotype, so a negative 3243A>G assay does not exclude an mtDNA defect as the cause.

• Other mutations are much less common. Although some can be reliably detected in blood (such as the 8993T>G/C mutation in NARP or MERRF), this is not always the case. In the Newcastle cohort of over 400 patients, a large proportion did not fit neatly into a defined clinical category, and required systematic investigation looking for evidence of mitochondrial dysfunction followed by molecular genetics tests.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Fundal appearance in Leber hereditary optic neuropathy (courtesy of Mr Philip Griffiths).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

T2-weighted MR brain scan showing ischaemic stroke-like lesions, typically in the posterior cerebral hemispheres, in MELAS (with thanks to Dr Andrew Scheafer)

WHEN SHOULD A MUSCLE BIOPSY BE PERFORMED?

If there is sufficient clinical suspicion of a mitochondrial disorder to warrant a molecular genetic blood test, and yet it is negative, then further investigation is mandatory. Urinary epithelium or buccal mucosa provide an alternative. In children, nuclear genetic defects are more common, and these can be detected on cultured skin fibroblasts or clinically affected tissue such as liver. In adults a muscle biopsy is most likely to reveal the diagnosis.

In Newcastle we rarely perform an open biopsy for a suspected metabolic disorder. In skilled hands a needle biopsy is quick, safe, and can be done as an outpatient in the neurological investigation unit. Muscle histochemistry is carried out to a high standard in most muscle diagnostic laboratories in the UK, and this is an invaluable screening test for adult mitochondrial disease.

The proliferation of mitochondria leads to the “ragged” appearance of some muscle fibres, but this is not specific for mitochondrial disease unless it is particularly prominent. Cytochrome c oxidase (COX) histochemistry is much more specific. Adults often show a mosaic (patchy) COX defect, but nuclear gene defects which specifically affect COX cause a global decrease in activity which can be difficult to appreciate. This becomes easier to see if the muscle section is stained with succinate dehyrogenase (SDH) after the COX reaction. SDH is entirely nuclear encoded and is not deficient in patients with mtDNA mutations, so “lighting up” the COX deficient fibres (fig 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Nuclear mitochondrial disorders. (A) Muscle histochemistry showing sequential cytochrome c oxidase (COX)/succinate dehydrogenase (SDH) reactions in skeletal muscle. The normal fibres are brown. The fibres with COX deficiency allow the sequential SDH reaction to stain them blue. (B) Long range PCR showing: lane c = control DNA with a single 16.5 Kb band, lanes 1 and 2 = two subjects with multiple mtDNA deletions, and a size marker. (C) Southern blot confirming the presence of multiple deletions in skeletal muscle from patient 2. (D) Fluorescent sequence chromatogram showing a heterozygous mutation in the C10ORF2 gene (arrow). With thanks to Dr Marcus Deschauer, Halle, and Dr Gavin Hudson, Newcastle (adapted from Hudson et al ²⁶)

IS THE MUSCLE BIOPSY INFALLIBLE?

There are two difficulties when interpreting muscle histochemistry. First, routine histochemistry only tests part of the respiratory chain. Some patients have a biochemical defect that principally or exclusively involves complex I (such as some with 3243A>G), and these individuals can have normal muscle histochemistry. Likewise, the muscle phenotype can be extremely subtle or absent in patients with POLG1 mutations, so normal muscle histochemistry does not exclude this diagnosis.¹⁸

The second problem is that mitochondrial abnormalities are seen in healthy aged individuals. Somatic mutation of mtDNA (mutations acquired during life) can accumulate in a very small proportion of muscle fibres as part of normal ageing, leading to a few COX deficient fibres in a biopsy of 200–300 fibres. These are not present in subjects less than 50 years of age, so even a few abnormal fibres is significant in younger individuals, and should be pursued further. The significance of a few COX negative fibres in older individuals depends very much on the clinical context—if there is a strong suspicion, then further biochemical and genetic investigations are warranted.

If there is sufficient clinical suspicion of a mitochondrial disorder to warrant a molecular genetic blood test, and yet it is negative, then further investigation is mandatory

FURTHER BIOCHEMICAL TESTS

It is useful to measure the activity of respiratory chain complexes in adults if the muscle histochemistry is normal or subtly abnormal. The results can also be a helpful guide for genetic studies if the common mutations are not present. Respiratory chain biochemistry in any tissue is technically difficult to do and also it can be difficult to interpret. Only a few laboratories in the UK perform the appropriate assays to a high standard, and even then there can be differences in interpretation (as recently demonstrated in a French study where markedly different values were obtained from different laboratories using the same tissue and ostensibly the same protocol). The aim is to look for a clearcut deficiency in one or more complexes which will confirm the diagnosis and suggest which genome or genes to screen in detail.

As with the other investigations, the outcome is not absolute. Secondary respiratory chain defects are well recognised and can be due to other multisystem neurological disorders (such as fatty acid oxidation disorders, disorders of amino acid metabolism, and neuroferritinopathy), and tissue specific defects that only affect the brain may not be detected in skeletal muscle. In general, respiratory chain complex assays are most helpful in children. Unfortunately, in adults, it is precisely those patients who have a mild phenotype and subtle histochemical abnormalities who are most likely to have normal complex assays. The biochemical defect in muscle may only affect a small proportion of muscle fibres which are “diluted out” when a block of muscle tissue is homogenised for further biochemical studies. New approaches using complex-specific antibodies are under development, but these have yet to be validated.

MOLECULAR GENETIC TESTS ON MUSCLE

Further molecular testing on muscle is indicated if the common mutations are not present in blood, and there is biochemical evidence of a respiratory chain defect, or there is a very strong clinical suspicion. The standard approach in Newcastle is to look for mtDNA deletions using a technique called long range PCR (polymerase chain reaction), which preferentially detects small fragments of mtDNA. If there are abnormalities, the sample is studied further with real-time PCR or a Southern blot which allows quantification of the rearrangements, and the detection of mtDNA depletion (loss of mtDNA, usually presenting in children and due to an underlying nuclear gene mutation (see fig 5)). These tests usually make it possible to distinguish between age related somatic mutation and genuine pathogenic deletions. The presence of multiple mtDNA deletions suggests an underlying nuclear gene mutation causing a disorder of mtDNA maintenance (see later, and fig 4). If there are no deletions, the next approach is targeted sequencing of specific mtDNA genes, and ultimately, complete genome sequencing if necessary.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Laboratory investigation of suspected mitochondrial disease in adults. Items in red refer to investigations carried out at specialised centres.

Complete mtDNA sequencing is technically straightforward and costs about the same as an MR brain scan. It is important to sequence muscle DNA because low levels of a heteroplasmic mutation may not be detected by automated fluorescent sequencing. Many patients have novel point mutations and deciding whether a particular substitution is pathogenic can be difficult. MtDNA is highly polymorphic, and it is not uncommon to find unique polymorphisms. If a mutation is heteroplasmic, then it can be relatively straightforward to show that high percentage levels are found in the pathologically abnormal muscle cells by single muscle fibre PCR analysis. This strongly suggests the mutation is causative. However, the recent description of homoplasmic mtDNA mutations in patients with mosaic COX defects poses a particular challenge.¹⁹ Previously these mutations would not have been considered pathogenic, but the boundaries have shifted, and proving that these mutations are not polymorphisms can take many years of laboratory research.

Until 2005, the role of nuclear gene sequencing had been limited in adults, restricted to rare families with dominant chronic progressive external ophthalmoplegia which can be due to mutations in three nuclear genes: POLG1 which codes for the mtDNA polymerase, C10ORF2 which codes for the mitochondrial helicase Twinkle, and ANT1 which codes for adenine nucleotide translocase.³ These genes code for proteins which maintain mtDNA and, when mutated, they cause secondary mtDNA mutations to accumulate in non-dividing cells such as skeletal muscle and neurons, leading to the clinical phenotype.

Detecting POLG1 mutations in adults with ataxia without ophthalmoplegia poses a particular challenge because these patients do not always have abnormal muscle histochemistry.¹¹ The phenotype of POLG1 disease appears to be particularly wide—some of the patients with ataxia presented with headache and epilepsy preceding the ataxia by many years,¹² and we have seen patients with muscle pain and fatigue many years before they developed objective signs which led to further investigation. Should we be screening POLG1 in these patients either before carrying out a muscle biopsy, or if the biopsy is negative? POLG1 has 22 coding exons (regions of the gene that code for the protein), and mutations in adults appear to be scattered throughout the gene. Direct sequencing is therefore not a trivial exercise, and there are common polymorphisms in the gene that complicate interpretation.

Current evidence suggests that all patients with neurological disease due to POLG1 mutations have a secondary mtDNA abnormality (mtDNA deletions in muscle, or depletion in an affected tissue), so this is a useful screening test. Of course, normal healthy elderly individuals also have secondary mtDNA deletions in muscle, so the presence of multiple deletions is not specific for POLG1 disease. At present each case must be interpreted on an ad hoc basis, and this field is changing rapidly.

OLIGOSYMPTOMATIC PATIENTS AND PROVING IT IS NOT MITOCHONDRIAL DISEASE: HOW FAR SHOULD YOU GO?

This is a difficult question, and the answer depends very much on the clinical context. There is no doubt that an isolated posterior circulation ischaemic stroke can be the presenting feature of the 3243A>G mutation even late in life,²⁰ exercise intolerance may be the only feature in patients with mtDNA cytochrome b gene mutations,²¹ and subtle ptosis can be transmitted as a dominant trait in families with a heterozygous POLG1 mutation.²² What are the distinguishing features? Unfortunately there may not be any, and the diagnosis only becomes clear with time. Even with a high index of suspicion in a tertiary centre, we have missed a few patients passing through our hands. The chance of missing these patients is minimised by keeping an open mind, and taking a meticulous family history.

Proving a disorder is not mitochondrial disease is even more difficult, and arguably impossible. The extent of investigation really depends on the probability of the diagnosis from the outset. Every appropriate negative test reduces the likelihood of the disorder, but does not exclude it completely. If there is any doubt, sometimes the best approach is to monitor the situation in the outpatient clinic over a year or two.

SHOULD I REFER TO A TERTIARY CENTRE? A PERSONAL VIEW

Working in a tertiary referral centre for mitochondrial disorders, it is difficult to be wholly objective about this issue. For many adults with classical phenotypes, making a diagnosis is relatively straightforward, and can be done either through local Molecular Genetics Services, or by sending tissue to a specialised laboratory. In Newcastle we receive blood samples and muscle biopsies every day, and carry out specialised investigations directly liaising with the referring clinician. From a practical stance, given the disability experienced by many patients with suspected mitochondrial disease, it seems sensible to do as many investigations as close as possible to home, and only to refer to a tertiary centre when the routine histochemical and basic molecular genetic tests are negative or difficult to interpret. This approach makes sense, provided the referring clinician feels comfortable with the immediate management issues: prognostic and genetic counselling, screening for possible complications, and therapeutic options.

I would argue that tertiary referral is appropriate for a number of reasons. First, mitochondrial disorders are relatively uncommon and heterogeneous. Although management is largely supportive at present, treating some of the complications is highly specialised (such as ptosis surgery) and it makes sense for one or a few national centres to develop expertise in this area. Second, the genetic implications are often complex and our knowledge of the area is rapidly evolving. For example, we previously thought that mtDNA deletions were essentially sporadic diseases, but recent work has shown that the recurrence risk is in the order of one in 24.²³ Again, and particularly because there is no effective treatment, it is essential that families get the best advice. Finally, in Newcastle we are currently leading a pan-European longitudinal clinical study of mitochondrial disease in order to understand the natural history of these disorders and develop cohorts for clinical trials with new treatments. This will only be successful if we collect data on as many cases as possible. Even if patients are not referred to our centre, it would be helpful to know where they are and what the problem is.

CONCLUSIONS

The algorithm shown in figure 5 summarises the main points, and can be used as a field guide when considering a mitochondrial disorder. This approach will lead to a biochemical or molecular diagnosis in most patients. We would be happy to discuss difficult cases, and receive clinical referrals or specimens for further analysis. Looking back over the last 10 years, some disorders have become very easy to diagnose, but the expanding phenotypic spectrum makes the question “Could it be mitochondrial disease?” even more difficult to answer. Management still remains a challenge. A recent Cochrane review failed to identify any clearcut evidence in favour of any particular oral treatment agent,²⁴ but trials are ongoing, including the role of exercise therapy.²⁵ Hopefully systematic international studies will put the growing discipline of mitochondrial medicine onto a more solid clinical foundation