A systematic review and meta-analysis of Harmonic Focus in thyroidectomy compared to conventional techniques

Introduction Several meta-analyses have been performed comparing the use of a variety of ultrasonic devices in thyroidectomy to conventional procedures. These studies have shown the superiority of ultrasonic devices for most outcomes studied including faster operative time and less blood loss, and equivalent or better safety for recurrent laryngeal nerve paresis and hypocalcemia. The current work is the first to examine a single ultrasonic device specifically designed for thyroid surgery, the Harmonic Focus, in order to confirm its efficacy and safety in thyroidectomy. Methods A comprehensive literature search without language restrictions was performed for randomized clinical trials comparing Harmonic Focus and conventional clamp, cut and tie in thyroidectomy. Outcome measures included operating time, blood loss, post-operative pain, length of hospital stay, hypocalcemia and recurrent laryngeal nerve paresis. Risk of bias was analyzed for all studies. Meta-analysis was performed using random effects models with the inverse-variance method for mean differences of continuous variables and the Mantel-Haenszel method for risk ratios of dichotomous variables. Results A total of 14 studies met the inclusion criteria. Harmonic Focus reduced operative time by 29 min, a 31 % decrease (p < 0.001), intra-operative blood loss by 45 ml (p < 0.001), post-operative pain (p < 0.001), length of hospital stay by 0.68 days (p = 0.005), drainage volume by 29 ml (p = 0.01), and occurrence of transient hypocalcemia by 40 % (p = 0.001). There were no significant differences between Harmonic Focus and conventional procedures in rate of persistent hypocalcemia, or rates of transient and persistent recurrent laryngeal nerve paresis. Conclusion This is the first meta-analysis of Harmonic Focus in thyroid surgery. In agreement with meta-analyses previously performed on ultrasonic devices, use of the Harmonic Focus has been shown to be a more effective surgical procedure compared to conventional methods in thyroidectomy. The low occurrence of hypocalcemia and recurrent laryngeal nerve paresis confirms that Harmonic Focus can improve thyroidectomy efficiency without increasing the risk of complications.


Introduction
Ultrasonic cutting and coagulation of soft tissues was introduced in 1991 with a primary focus on laparoscopic surgery in general and cholecystectomy in particular [1,2]. There then followed a rapid evolution and adoption of the technology for vessel sealing that resulted in widespread use in a variety of intra-abdominal procedures including colectomy [3], fundoplication [4], and hysterectomy [5]. By the late 1990s and early 2000's, endocrine and head and neck surgeons had adopted such devices for thyroid surgery noting decreased operating times [6]. Since then numerous randomized control trials (RCTs), large case series and numerous meta-analyses have been published confirming the reduction in operating time [7]. Furthermore, they have shown less blood loss (albeit not clinically significant), smaller incisions, less postoperative pain, reduced hospitalization and less drainage. Perhaps most importantly, studies that have evaluated the economic impact of using ultrasonic energy have concluded that there is overall cost savings when using such devices [8,9].
A drawback of these studies is that they have used a variety of ultrasonic devices including Ultracision, Ace and Focus. However, since its introduction in 2007, the only specifically designed and approved device for use in thyroid and head and neck surgery is Harmonic Focus® (Ethicon Inc., Cincinnati OH). No meta-analysis or systematic review has been published to date evaluating specifically the impact of Harmonic Focus on the outcomes of thyroid surgery. Furthermore, some controversy exists in the literature concerning the impact of ultrasonic devices in general and Harmonic Focus in particular on parathyroid gland and recurrent laryngeal nerve function when these devices are used.
The purpose of this systematic review and metaanalysis is to answer two questions. First, what is the impact of Harmonic Focus on overall thyroid surgery outcomes when compared to the most commonly used clamp, cut and tie method that includes monopolar and or bipolar electrosurgery? Second, what impact does Harmonic Focus have on parathyroid and recurrent laryngeal nerve function following thyroid surgery?

Methods
Twenty-one databases were systematically searched, including MEDLINE via PubMed, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), and 18 other national databases (Table 1). Reference lists of retrieved articles were reviewed and additional comprehensive searches were conducted through Google Scholar and Research-Gate. Publications of all languages were considered in the systematic review.
The PICOS categories (i.e., population, intervention, comparator, outcomes, and study design) were used to define study inclusion criteria. All published RCTs comparing the use of Harmonic surgical devices to conventional methods, such as monopolar or bipolar electrosurgery and suture, clips, or knot-tying in human subjects, for all surgery types, were considered for inclusion (Table 2). Benign and malignant disease were included if no lymph node dissection was performed. Full-text papers were excluded if, they were not a RCT, the principal surgical procedure was not thyroidectomy, devices other than the Harmonic Focus were used, and if lymphadenectomy, or only partial thyroidectomy, was conducted. The eligibility of each publication was evaluated by two independent reviewers (IS, NCF) and a third reviewer (HC) was consulted in the case of disagreements regarding study inclusion. When necessary, study authors were contacted for additional methodological details to confirm whether the study was randomized.
Study details (i.e., baseline characteristics and outcomes) of included publications were extracted through a standardized data extraction form. Two reviewers extracted data independently and any inconsistencies were resolved by consensus or by consultation with a third reviewer. Non-English publications were translated and data extraction was completed. Data extraction by one reviewer was subsequently cross-checked by a second reviewer.
The following clinical outcome measures were included: (1) operating time, (2) intra-operative blood loss, (3) postoperative pain, (4) length of hospitalization, (5) volume of drainage fluid, (6) hemorrhage requiring surgery, (7)  was assumed that there were zero cases of the persistent type. The same assumption was applied to the transient and persistent hypocalcemia outcome. Both clinically-and chemically-determined hypocalcemia were included in the hypocalcemia parameter. The standard deviation (SD) variance measure was not reported in one study [10] for operating time. Study authors were not contacted to retrieve missing data, however, standard methods provided by Cochrane [11] were used to impute the missing variance measure in this study.
The Cochrane Collaboration tool [11] for assessing risk of bias was used to evaluate the quality of the included studies. Based on seven pre-specified domains (sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias), publications were scored as having low, unclear, or high risk of bias. Final quality assessments were based on the combination of these factors and individual study characteristics. Two authors independently assessed the study quality and inconsistencies were resolved through consensus or by discussion with a third author.
Review Manager (Version 5.3, The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark, 2014) was used to perform the meta-analysis. The inverse-variance method was used to calculate the mean differences (MD) for continuous outcomes (operating time, intra-operative blood loss, post-operative pain, length of hospitalization, volume of drainage fluid). Risk ratios (RR) were calculated for dichotomous outcomes (hemorrhage requiring surgery, wound seroma, transient hypocalcemia, persistent hypocalcemia, transient RLN paresis, persistent RLN paresis) using the Mantel-Haenszel method. The meta-analysis used a random effects model and forest plots for all included outcomes were generated using Review Manager. Study heterogeneity was evaluated through the χ 2 test and I 2 measure.
The Harmonic Focus was compared to conventional methods in the primary analysis. Sensitivity analyses were conducted for the imputation of missing variance measures, where studies requiring imputation were excluded [10]. Further sensitivity analyses were completed for study quality, where studies with unclear or high risk of bias across several measures were excluded [8,[12][13][14][15]. Sensitivity analyses excluding studies that did not include electrocautery as part of the conventional technique [8,[15][16][17][18] were also performed.

Results
The systematic search resulted in the identification of 4542 total records, of which 4153 were excluded following title and abstract screening (Fig. 1). Of the 389 full text articles retrieved and reviewed, 375 were further excluded if studies were non-RCTs, had an undefined manufacturer, did not use Harmonic Focus, the publication was unavailable and had nonhuman subjects, or the surgical procedure was not total thyroidectomy. Overall, 14 studies consisting of 2516 patients in total reporting on Harmonic device (Focus) use in total thyroidectomy were included in the meta-analysis [8][9][10][12][13][14][15][16][17][18][19][20][21][22].
The risk of bias varied across the included studies. The overall results of the risk of bias assessments are reported in Fig. 2 and individual study quality assessments are summarized in Table 3. Randomization method was known in nine studies [8-10, 12, 16, 17, 19-22]. Three studies described randomization through the use of envelopes [10,12,17], two used a random permuted block design [9,22], and two described the use of a drawing technique [8,21]. One study used a computer-generated schedule [19] and one described a fixed simple randomization method [20]. Six studies [9,10,12,17,19,22] described concealment of the randomization sequence. Blinding of patients to the surgical technique was reported in six studies [8-10, 14, 15, 20], one study [22] reported blinding of the surgeon to the surgical technique, and three studies [9,10,19] described blinding of outcome assessors. Risk of performance bias was deemed low in non-blinded studies, as outcomes were considered objective and unlikely to be affected by a lack of blinding. There were no patient withdrawals in seven studies [8,9,12,[16][17][18]20] and one study [22] reported exclusions, but was assumed to have no clinically relevant impact. Reporting of attritions or exclusions was insufficient in six studies [10, 13-15, 19, 21]. Selective reporting remained unclear in nine studies [9, 10, 14-18, 20, 21], while three studies were deemed to have a high risk of bias as certain outcomes noted in protocol were not included in the results [8,12,13].  the Harmonic Focus compared to conventional methods in total thyroidectomy (Fig. 3).

Post-operative pain
On the basis of three studies comparing the Harmonic Focus to conventional techniques in total thyroidectomy, a statistically significant reduction in mean VAS reported post-operative pain (Harmonic Focus: 2.54, conventional technique: 3.87) by 1.33 points (95 % CI: −1.99 to −0.67; P < 0.0001; 3 studies; I 2 = 85 %) was observed with the Harmonic Focus (Fig. 5).

Drainage volume
In contrast to conventional methods in total thyroidectomy, mean drainage volume (Harmonic Focus: 16.25 mL, conventional technique: 45.63 mL) was statistically significantly   (Fig. 7).

Hemorrhage requiring surgery
Five of the eight studies examining post-operative bleeding requiring re-operation reported on hemorrhage events. Results demonstrated no statistically significant difference in bleeding events requiring re-operation between the Harmonic Focus and conventional methods (RR = 0.68; 95 % CI: 0.19 to 2.46; P = 0.55; 8 studies; I 2 = 0 %) (Fig. 8).

Hypocalcemia
Compared to conventional techniques in total thyroidectomy, the Harmonic Focus resulted in a statistically significant reduction in transient hypocalcemia with a RR of 0.60 (95 % CI: 0.44 to 0.82; P = 0.001; 12 studies; I 2 = 32 %) (Fig. 9). Persistent hypocalcemia events were reported in only two of the twelve studies examining this outcome. Results demonstrated a lower risk of persistent hypocalcemia with the Harmonic Focus than with conventional methods, although not statistically significant (RR = 0.35; 95 % CI: 0.07 to 1.91; P = 0.23; 12 studies; I 2 = 0 %) (Fig. 10).

RLN paresis
Harmonic Focus use was associated with fewer transient RLN paresis events compared with conventional thyroidectomy techniques, but these results were not statistically significant (RR = 0.64; 95 % CI: 0.28 to 1.44; P = 0.28; 13 studies; I 2 = 0 %) (Fig. 11). Of the thirteen studies examining persistent RLN paresis, only one study reported an event. In this study, the Harmonic Focus reduced the risk of persistent RLN paresis, although results were not statistically significant (RR = 0.33; 95 % CI: 0.01 to 8.03; P = 0.50) (Figure not shown).

Sensitivity analyses
Sensitivity analyses demonstrated results similar to the primary analysis and were relatively robust to variables tested. Results for operating time, intraoperative blood loss, post-operative pain, length of hospitalization and transient hypocalcemia remained statistically significantly lower with the Harmonic Focus when studies with a higher risk of bias were excluded (Table 4). Additionally, primary analysis results were relatively robust to the exclusion of studies with conventional techniques that did not utilize monopolar or bipolar electrocautery or when imputed results by Miccoli et al. [10] were excluded for operating time.

Discussion
Although thyroidectomy was not one of the initial applications of ultrasonic surgery, it was realized that use of the Harmonic scalpel provides many advantages over the conventional procedures. Because of the high vascularization of the thyroid gland, efficient and meticulous hemostasis is required to reduce the risk of complications and postoperative morbidity. As with any surgical procedure, a reduction in operative time will typically lessen the chance of surgical-site infection and may lead to faster patient recovery time [23]. Rapid coagulation with the Harmonic Focus can substantially increase the speed of the operation. In addition, the precision of the Harmonic Focus can ensure that the risk of the two principal adverse events, namely hypocalcemia and recurrent laryngeal nerve paresis, are not elevated, or may even be reduced. Five standard meta-analyses have compared ultrasonic devices to conventional techniques in thyroidectomy. These previous findings are in general similar to those in this study. In analyzing 9 studies, Melck [24] observed significantly faster operative time (−23.1 min, p < 0.001), and less transient hypocalcemia (RR 0.69, p = 0.01) for use of ultrasonic devices compared to conventional methods. In 7 studies, Cirocchi [25] found significantly faster operative time (−18.7 min, p < 0.001), less blood loss (−60.1 ml, p = 0.04), and less drainage volume (−35.3 ml, p < 0.001). In 11 studies, Zhang [26] found significantly faster operative time (−22.4 min, p < 0.001), and less intraoperative bleeding (−26.6 ml, p = 0.02). In 12 studies, Ecker [27] observed significantly faster operative time (−22.7 min, p < 0.001), less blood loss (−20.0 ml, p < 0.001), less post-operative pain (−0.86 units, p = 0.02), and reduced length of hospital stay (−0.12 days, p = 0.05). Finally, in 13 studies, Zhao [28] found significantly faster operative time (−21.1 min, p < 0.001), less intra-operative blood loss (−14.4 ml, p < 0.001), less drainage (−7.5 ml, p < 0.001), and lower hospitalization charges (−118 USD, p < 0.001). None of these meta-analyses found a higher rate of transient or persistent RLN paresis for ultrasonic devices compared to conventional techniques.
Recently two network meta-analyses have been performed. In these studies, ultrasonic devices were compared to both conventional techniques and advanced bipolar technology. In one of these network metaanalyses, Contin [7] evaluated 21 studies and found significantly faster operative time (−22.3 min, p < 0.001), less intraoperative blood loss (−28.5 ml, p < 0.001), shorter hospital stay (−0.28 days, p = 0.016) and less postoperative bleeding (−11.2 ml, p < 0.001) for ultrasonic devices. There was a trend toward lower transient hypocalcemia (p = 0.066) for ultrasonic, and importantly no difference in rates of transient RLN paresis (p = 0.847) or persistent RLN paresis (p = 0.711). Interestingly, the mean difference of operative time for investigator-initiated trials was greater than for industry-sponsored trials, belying the notion in this case that industry trials are necessarily biased toward their own commercial product.
In the other network meta-analysis, Garas [29] claimed to use 25 RCT's, although one of the studies misidentified an electrosurgical device as ultrasonic and another study was not randomized [7,30]. In comparing ultrasonic to conventional, they reported significantly faster operative time (p < 0.01), less blood loss (p < 0.01), less drain output (p = 0.03), and lower cost (p = 0.03), but a trend towards higher persistent RLN paresis (p = 0.08) for ultrasonic. A re-tabulation of the studies included in the Garas analysis [30]   used, there was no difference in the rate of persistent RLN paresis (p = 0.50), with zero cases among a total of 1119 subjects for Harmonic (0.0 %) and one case among 1085 subjects for conventional techniques (0.1 %). This lack of difference was also confirmed in the Contin network meta-analysis [7], which suggested that the results of Garas "should be read with caution." Based on the results of the standard and network meta-analyses, including re-analysis of the Garas study, and the current study, all of which show no difference between ultrasonic devices and conventional techniques, there should be high confidence that Harmonic Focus can be used thyroidectomy with a low risk of RLN paresis, equivalent to conventional clamp, cut and tie.
In contrast to all previous efforts, the current study was limited to only Harmonic Focus as the ultrasonic device of interest. The Harmonic Focus is designed for use in open procedures and is specifically cleared for applications in head and neck surgery. A small end effector enables precise placement and the ultrasonic energy provides rapid dissection and coagulation. As evidenced in all studies to date, the design of the Harmonic Focus gives substantial advantages over conventional techniques in thyroidectomy.
In the current study, we observed operative time was significantly faster for Harmonic Focus compared to conventional clamp, cut, and tie. This speed increase is a result of both faster dissection and better hemostasis. Shorter operative time can lead to improved patient outcomes and faster recovery, and provide operative and overall hospital cost savings [8,9]. The difference of 29 min observed in this study represents a 31 % decrease in operative time compared to conventional technique. Other measures related to the hemostasis capabilities are intra-operative blood loss and drainage volume, where use of Focus led to decreases of 60 % and 64 %, respectively, compared to clamp, cut and tie.
Post-operative pain was 34 % lower for Harmonic Focus than clamp, cut, and tie. This advantage may be related to shorter operative time noted above and observations that ultrasonic devices generally cause less inflammation [31][32][33]. The decrease of 1.33 points represents a 34 % reduction in pain compared to conventional surgery. The decrease in length of hospital  For the dichotomous results evaluated, namely, hemorrhage requiring surgery, hypocalcemia, RLN paresis, and wound seroma, all were less frequent with Harmonic than conventional technique, but because of the low frequency of all these events, only for transient hypocalcemia was there a significant difference, with Harmonic having a 40 % lower rate of occurrence than conventional clamp, cut, and tie. This significantly lower rate of transient hypocalcemia may be a result of the shorter operative time and reduced overall systemic impact of surgery with Harmonic Focus compared to conventional surgery. A drop in calcium blood levels may result not just from unintended damage to the parathyroid gland by clamp, cut and tie, but also be a typical systemic response to longer, more involved surgical procedures. While Harmonic Focus provides a lower rate of transient hypocalcemia, there does not appear to be a long-term difference, as the rate of persistent hypocalcemia is not significantly different from conventional procedures.
As demonstrated in this and all other reliable metaanalyses, the rate of both transient and persistent RLN paresis with ultrasonic devices is not greater than with clamp, cut & tie procedures. In fact, the trend with Harmonic Focus suggests that with more data it may be possible to show a reduction in transient RLN paresis. The cause of transient RLN paresis is generally assigned to excess traction. Use of continuous intra-operative nerve monitoring with stimulation to the vagus nerve can avoid paresis if the traction is reduced when the electromyographic reading has decreased by 50 % or more [34]. The improved dissecting ability of ultrasonic devices may decrease the amount of traction compared to clamp, cut and tie. The etiology of persistent paresis is typically prolonged thermal, electrical or direct contact with transection of the RLN. To avoid this complication, it is critical to visually identify the RLN during     [15] surgery. Use of an ultrasonic device eliminates electrical current and gives a wider margin of error for thermal effects compared to monopolar electrosurgery. Numerous preclinical studies have evaluated the impact of Harmonic technology when used close to a nerve [32]. All studies to date have shown an effect on nerve function similar to that of cold steel scalpel when used up to 2 mm from the nerve. Important caveats to remember are that activation time should be less than 15 s and the nerve itself should never be touched directly either during activation or immediately afterwards. As all energy technologies generate heat, they must be carefully managed. In contrast to ultrasonic energy, monopolar and bipolar technologies produce at minimum transient injury when used up to 5 mm from the nerve. Recently, the potential adverse effects of advanced bipolar technologies were demonstrated with a doubling of βamyloid precursor protein, a marker of impaired axonal transport, and a 133 % increase in neural inflammation when compared to ultrasonic energy [35].
Despite the comfort that can be developed from studies evaluating Focus, it must be stressed that no technology will replace meticulous surgical techniques and no technology is fool-proof. In that regard, the important elements must always be kept foremost in the surgeon's mind when using Harmonic Focus. First, activations should occur as far from nerves or glands as possible, and no closer than 2 mm. Second, continuous activation should be less than 15 s as one approaches the nerve. Finally, the heat of the instrument must be actively managed by cooling, and an activated instrument should never touch the nerve during dissection.

Conclusions
The Harmonic Focus is a more effective surgical device compared to conventional techniques in thyroidectomy. Its use offers several clinical advantages, including reduced operating time, intra-operative blood loss, drainage volume, post-operative pain, length of hospital stay, and transient hypocalcemia which can ultimately benefit the surgeon, patient and hospital, without the addition of safety concerns.
All studies performed to-date have used Harmonic Focus, prior to the introduction of an improved version, Harmonic Focus+, which has a substantially smaller end effector and produces a significantly smaller thermal effect [35]. All of the benefits observed in this study are expected to be maintained or even improved upon with the increased precision of the new device.