Abstract

Mitochondrial respiratory chain disorders are the most prevalent group of inherited neurometabolic diseases. They present with central and peripheral neurological features usually in association with other organ involvement including the eye, the heart, the liver, and kidneys, diabetes mellitus and sensorineural deafness. Current treatment is largely supportive and the disorders progress relentlessly causing significant morbidity and premature death. Vitamin supplements, pharmacological agents and exercise therapy have been used in isolated cases and small clinical trials, but the efficacy of these interventions is unclear. The first review was carried out in 2003, and identified six clinical trials. This major update was carried out to identify new studies and grade the original studies for potential bias in accordance with revised Cochrane Collaboration guidelines.

To determine whether there is objective evidence to support the use of current treatments for mitochondrial disease.

We searched the Cochrane Neuromuscular Disease Group Specialized Register (4 July 2011), CENTRAL (2011, Issue 2, MEDLINE (1966 to July 2011), and EMBASE (January 1980 to July 2011), and contacted experts in the field.

We included randomised controlled trials (including cross-over studies). Two of the authors independently selected abstracts for further detailed review. Further review was performed independently by all five authors to decide which trials fit the inclusion criteria and graded risk of bias. Participants included males and females of any age with a confirmed diagnosis of mitochondrial disease based upon muscle histochemistry, respiratory chain complex analysis of tissues or cell lines or DNA studies. Interventions included any pharmacological agent, dietary modification, nutritional supplement, exercise therapy or other treatment. The review authors excluded studies at high risk of bias in any category. The primary outcome measures included an change in muscle strength and/or endurance, or neurological clinical features. Secondary outcome measures included quality of life assessments, biochemical markers of disease and negative outcomes.

Two of the authors (GP and PFC) independently identified studies for further evaluation from all abstracts within the search period. For those studies identified for further review, all five authors then independently assessed which studies met the entry criteria. For the included studies, we extracted details of the number of randomised participants, treatment, study design, study category, allocation concealment and other risk of bias criteria, and participant characteristics. Analysis was based on intention-to-treat data. We planned to use meta-analysis, but this did not prove necessary.

The authors reviewed 1335 abstracts, and from these identified 21 potentially eligible abstracts. Upon detailed review, 12 studies fulfilled the entry criteria. Of these, eight were new studies that had been published since the previous version of this review. Two studies which were included in the previous version of this review were excluded because of potential for bias. The comparability of the included studies is extremely low because of differences in the specific diseases studied, differences in the therapeutic agents used, dosage, study design, and outcomes. The methodological quality of included studies was generally high, although risk of bias was unclear in random sequence generation and allocation concealment for most studies. Otherwise, the risk of bias was low for most studies in the other categories. Serious adverse events were uncommon, except for peripheral nerve toxicity in a long-term trial of dichloroacetate (DCA) in adults.

One trial studied high-dose coenzyme Q10 without clinically meaningful improvement (although there were multiple biochemical, physiologic, and neuroimaging outcomes, in 30 participants). Three trials used creatine monohydrate alone, with one reporting evidence of improved measures of muscle strength and post-exercise lactate, but the other two reported no benefit (total of 38 participants). One trial studied the effects of a combination of coenzyme Q10, creatine monohydrate, and lipoic acid and reported a statistically significant improvement in biochemical markers and peak ankle dorsiflexion strength, but overall no clinical improvement in 16 participants. Five trials studied the effects of DCA: three trials in children showed a statistically significant improvement in secondary outcome measures of mitochondrial metabolism (venous lactate in three trials, and magnetic resonance spectroscopy (MRS) in one trial; total of 63 participants). One trial of short-term DCA in adults demonstrated no clinically relevant improvement (improved venous lactate but no change in physiologic, imaging, or questionnaire findings, in eight participants). One longer-term DCA trial in adults was terminated prematurely due to peripheral nerve toxicity without clinical benefit (assessments included the GATE score, venous lactate and MRS, in 30 participants). One trial using dimethylglycine showed no significant effect (measurements of venous lactate and oxygen consumption (VO₂) in five participants). One trial using a whey-based supplement showed statistically significant improvement in markers of free radical reducing capacity but no clinical benefit (assessments included the Short Form 36 Health Survey (SF-36) questionnaire and UK Medical Research Council (MRC) muscle strength, in 13 participants).

Despite identifying eight new trials there is currently no clear evidence supporting the use of any intervention in mitochondrial disorders. Further research is needed to establish the role of a wide range of therapeutic approaches. We suggest further research should identify novel agents to be tested in homogeneous study populations with clinically relevant primary endpoints.