Comment on Garas et al., “Which Hemostatic Device in Thyroid Surgery? A Network Meta-Analysis of Surgical Technologies”

Abstract

Dear Editor: We disagree with the conclusions of the recent article by Garas et al. (1) on permanent recurrent laryngeal nerve injury (p-RLN) following thyroidectomy with different energy-based technologies using network meta-analysis. Since network meta-analysis extends the number and type of studies being combined, there is potential for combining studies that are not adequately similar, possibly leading to erroneous conclusions (2). One of the fundamental requirements of network meta-analysis to improve accuracy is that they include only very similar studies, under similar conditions, with similar patient populations, and clearly defined endpoints. Evaluation of the articles included by Garas et al. shows they have included a study using Starion (3), an electrocautery device, in the Harmonic arm and a nonrandomized trial in the Harmonic arm (4). Furthermore, not all devices are similar in a given treatment arm nor are the procedures performed, patients studied, or diseases treated similar. For example, within the Ultrasound group, devices used include CS-14C, Wave, and Focus, and within the LigaSure arm, Precise, Small Jaw, and unstated devices. Many of the studies in the Harmonic and LigaSure treatment arms combine the use of monopolar and/or bipolar electrosurgery. The disease state and procedures performed vary significantly from total thyroidectomy with central neck dissection to partial thyroidectomy to videoscopic thyroidectomy and from cancer to Graves' disease to multinodular goiter. Thus, to conclude that the populations are suitably similar and satisfactory for comparison through network meta-analysis is erroneous and raises serious questions regarding the conclusions reached. The conclusions of network meta-analysis performed by Garas et al. are that ultrasonic devices and advanced bipolar devices provide less blood loss, less drain output, shorter operative time, and shorter hospital stay compared to the conventional methods of clamp-cut-and-tie (CCT) thyroidectomy, but that the rate of p-RLN paralysis is significantly higher for ultrasonic devices and LigaSure devices compared to CCT. How they came to that conclusion is not evident from the studies they reviewed. When the total actual number of individual cases of p-RLN paralysis reported in the randomized control trials used by Garas et al. (excluding the two studies noted above) are totaled, there were 2/992 cases (0.2%) of p-RLN paralysis for CCT versus 1/1006 (0.1%) for Harmonic in RCTs where both are compared (and 1/266 cases [0.4%] of p-RLN paralysis for CCT vs. 1/266 [0.4%] for LigaSure in studies where CCT and LigaSure were compared directly). Consistent with the absence of a difference in p-RLN paralysis is the I2 value of 0.00% for p-RLN paralysis reported by the authors in Table 2. This statistic is a measure of heterogeneity, and values below 25% represent low heterogeneity. A value of 0.00% suggests no heterogeneity or variability across the studies and is consistent with the finding that the vast majority of studies reported no p-RLN paralysis. What is particularly troubling is that the authors state that the results of this network meta-analysis suggest that the superior safety profile of the CCT technique can contribute to future international thyroidectomy guidelines because they believe there is relative risk of permanent RLN. Based on direct comparison methods using the randomized clinical trials identified in the Garas et al. study, one would logically conclude that there is no increase in the risk of recurrent laryngeal nerve injury when Harmonic or LigaSure devices are used and that the use of these energies will provide less blood loss, less drain output, shorter operative time, and shorter hospital stay compared to the conventional methods of CCT thyroidectomy, all these being of significant benefit to patients, hospitals, and societies.