Study	Reason for exclusion
Bresolin 1990	High risk of bias because entry criterion for the double‐blind portion of the study was decided retrospectively based upon the results of the open‐label phase. Furthermore, the double‐blind portion of the study was not randomised.
Cejudo 2005	High risk of bias due to lack of blinding.
Chen 1997	High risk of bias. Participants were not randomised. The study design (3 months of coenzyme Q10 or 1 month placebo, with one month washout if the participant started on coenzyme Q10) would allow a blinded assessor to know whether participant was on treatment or placebo depending on the duration of time in the study. There is a possibility of selective outcome reporting since participants were assessed monthly although results were only reported for end of three‐month treatment period
Gimenes 2007	High risk of bias due to possibility of selective outcome reporting, and inability to fully assess risk of bias in other domains, since this publication consists of an abstract only, and the study has not been published in a peer‐reviewed format
Koga 2005	High risk of bias due to the non‐randomised, open‐label design
Muller 1990	High risk of bias. Participants were not randomised. Data and statistical analysis were not reported. There is a possibility of selective outcome reporting given that study endpoints were not clearly stated, and only 7 participants were reported. This evidence is in abstract format and therefore it is not possible to fully assess the risk of bias
Stacpoole 2008	High risk of bias due to the non‐randomised, open‐label, non‐controlled study design
Suzuki 1998	High risk of bias due to the non‐randomised, open‐label, non‐placebo controlled study design
Taivassalo 2006	High risk of bias due to the non‐randomised, non‐controlled, non‐blinded study design