On 2014 Oct 15, Gerd Heusch commented:

The skepticism of Dr. Berthelsen is well taken. However, contrary to his expectation, the results of our study with its primary endpoint troponin release have already been replicated, see Candilio L, 2015.

On 2014 Oct 13, Preben Berthelsen commented:

This study was registered with ClinicalTrials (NCT01406678). According to the registration, the authors planned to include 500 patients in a randomised study of remote ischaemic preconditioning (RIP) in isoflurane anaesthetised patients. No interim statistical analyses were planned. The study was, however, stopped early after the inclusion of 162 patients in the RIP-group and 167 patients in the group where no RIP was used. (Only 121 and 137 patients could be included in a per protocol analysis).

It is an indispensable rule not to break the randomisation code before all patients have completed the investigation. The authors have not followed this imperative. They have kept running track of the results as can be seen from the admission on page 598 of this paper “After use in some patients, however, we became aware of apparent interference of propofol with remote ischaemic preconditioning and discontinued its use in the remainder of the study”. So at least theoretically, it has been possible for the authors to analyse the data continuously and therefore stop the investigation – when a statistical difference was demonstrated – and before the stipulated number of patients had been reached.

A re-calculation of the power of the study using the 17% difference in 72h-troponin AUCs and a SD of 200 (standardised difference 0.27) shows that there is less than a 50% chance that another study would replicate the authors’ findings. It is my opinion that the result of this investigation must be viewed with some scepticism. Preben G. Berthelsen MD. Charlottenlund, Denmark.

On 2014 Oct 13, Preben Berthelsen commented:

This study was registered with ClinicalTrials (NCT01406678). According to the registration, the authors planned to include 500 patients in a randomised study of remote ischaemic preconditioning (RIP) in isoflurane anaesthetised patients. No interim statistical analyses were planned. The study was, however, stopped early after the inclusion of 162 patients in the RIP-group and 167 patients in the group where no RIP was used. (Only 121 and 137 patients could be included in a per protocol analysis).

It is an indispensable rule not to break the randomisation code before all patients have completed the investigation. The authors have not followed this imperative. They have kept running track of the results as can be seen from the admission on page 598 of this paper “After use in some patients, however, we became aware of apparent interference of propofol with remote ischaemic preconditioning and discontinued its use in the remainder of the study”. So at least theoretically, it has been possible for the authors to analyse the data continuously and therefore stop the investigation – when a statistical difference was demonstrated – and before the stipulated number of patients had been reached.

A re-calculation of the power of the study using the 17% difference in 72h-troponin AUCs and a SD of 200 (standardised difference 0.27) shows that there is less than a 50% chance that another study would replicate the authors’ findings. It is my opinion that the result of this investigation must be viewed with some scepticism. Preben G. Berthelsen MD. Charlottenlund, Denmark.

On 2014 Oct 15, Gerd Heusch commented:

The skepticism of Dr. Berthelsen is well taken. However, contrary to his expectation, the results of our study with its primary endpoint troponin release have already been replicated, see Candilio L, 2015.