Use of edaravone to decrease perioperative neurocognitive disorders in elderly patients with hip replacement

To the Editor: By a randomized controlled trial, Xi et at^[1] assessed the effects of edaravone on the development of postoperative delirium (POD) and perioperative neuro-cognitive disorders (PND) in elderly patients with hip replacement and showed that edaravone significantly decreased the incidence of POD within 7 days after surgery and the incidences of PND at 1 and 12 months after surgery. Given that both POD and PND are common complications after hip surgery in elderly patients and have been significantly associated with postoperative death, hospital-acquired complications, persistent cognitive impairments, poor postoperative functional recovery, prolonged duration of hospital stay, and increased healthcare costs,^[2,3] their findings have potential implications. However, we noted several issues in the methodology and results of this study on which we would like to invite authors’ comments.

First, the study objects were elderly surgical patients with a mean age >72 years. The authors only assessed preoperative cognitive function by the Montreal cognitive assessment score but did not determine whether patients suffered from preoperative neuropsychiatric comorbidities such as anxiety, delirium, depression, and sleep disorders. In fact, these comorbidities are common among elderly patients undergoing hip surgery and cannot be determined by the Montreal cognitive assessment score. It has been shown that these preoperative neuropsychiatric comorbidities are the most established predisposing factors of PND and POD after hip surgery in elderly patients.^[2-5] Furthermore, the history of elderly patients’ preoperative medications was not included in the baseline data. The available evidence indicates that anticholinergic drugs and benzodiazepines are widely utilized in managing elderly patients and have been significantly associated with the development and severity of POD.^[6] In addition, preoperative hemoglobin and albumin levels were also not provided in baseline data, though preoperative anemia and hypoalbuminemia have been significantly associated with an increased risk of POD in elderly surgical patients.^[7] We are concerned that any imbalance in the above preoperative risk factors would have biased their findings.

Second, in this study, the modified telephone interview for cognitive status was used to assess the cognitive function of all patients before and after surgery. Furthermore, the incidences of PND at 1 and 12 months after surgery were significantly lower in the edaravone group than in the control group. In the methods, however, the authors did not provide the diagnostic criteria of PND. In 2018, the International Nomenclature Consensus Working Group recommends that definitions of PND in a clinical study must meet the diagnostic criteria of neurocognitive disorders in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition.^[8] Evidently, the use of decreased scores of modified telephone interview alone for cognitive status to diagnose PND in this study cannot achieve the requirements of new recommendations for the definition of PND. Thus, we question the incidence of PND reported in this study.

Third, the duration of the hospital stay was shorter in the edaravone group than in the control group, but the authors did not provided the reasons for a prolonged hospital stay in the control group. Most importantly, it was also unclear whether the two groups were comparable with respect to early postoperative complications, such as infection, hemodynamic instability, arrhythmia, sleep disorders, accidental fall, anemia, and pneumonia, which are common after major surgery in elderly patients. It has been shown that these early postoperative complications can significantly increase the risk of POD and prolong the duration of hospital stay after hip surgery in elderly patients.^[9] To differentiate the real effect of one factor on the primary endpoint in a randomized controlled trial, we argue that all other possible influencing factors must be standardized for the avoidance of potential bias.

Finally, the authors described that primary endpoint were the incidence of POD within 7 days after surgery, the scores of the modified telephone interview for cognitive status, and the activities of daily life at 1 and 12 months after surgery. However, they used the incidence rate of decline in postoperative cognitive function scores to calculate the sample size. In fact, as a basic principle, in a randomized controlled trial, only an important observed parameter can be designed as the primary endpoint and sample size calculation must be performed on solely the primary endpoint. Furthermore, analysis of multiple secondary outcome parameters requires significance levels to be adjusted, for example, using a Bonferroni correction.^[10] We believe that clarification of these statistical issues will improve the transparency of this study design.

Conflicts of interest

None.