Despite heterogeneity in the methods of the included trials, it was possible to combine sets of data for 6 outcomes: operating time for elective laparoscopic cholecystectomy, gallbladder perforation with bile leak only, mean length of hospital stay, number of patients who needed overnight hospital stay, mean duration of sick leave, and abdominal pain scores at 24 hours postoperatively.
As shown in , the point estimate of the 5 included trials shows no statistically significant difference in mean patient age in the comparison groups (P = 0.57). Heterogeneity between included trials was statistically significant.
Operating time is significantly shorter in elective laparoscopic cholecystectomy when ultrasonic dissection is used.
shows a point estimate that favors the ultrasonic dissection with effect estimate (WMD) of −8.19 (95% CI −10.36 to −6.02). Heterogeneity between studies was statistically insignificant [ch2
=0.65, dt=2 (P=0.72)]. Furthermore, in the Janssen trial,8
the mean operating time was significantly shorter in ultrasonic versus electrocautery groups:
- When surgery was performed by the least experienced surgeons (who performed <10 laparoscopic cholecystectomies): 66.7 (range, 45 to 95) minutes versus 85.4 (range, 60 to 180) minutes (P=0.043).8
- When operating on patients with complicated gallbladders (eg, distended gallbladders, adhesions): 60 (range, 28 to 120) minutes versus 80 (range, 32 to 180) minutes (P=0.049).8
Operating time in minutes (p< 0.00001).
Most trials excluded patients with acute cholecystitis.2–6,8
Only Cengiz et al7
discussed differences in operating time in acute cholecystitis separately. The Mann-Whitney U test was used to calculate P value (0.004). This trial concluded that, in acute cholecystitis, the operating time is significantly shorter using the laparoscopic fundus-first technique with ultrasonic dissection (WMD −17, 95% CI −28.68 to −5.32, P=0.004).
Complicated cases were included in the Janssen trial.8
These are patients with hydrops of the gallbladder, gallbladder shrinkage, stones trapped in the cystic duct, and adhesions around the gallbladder.8
Interestingly, ultrasonic dissection was associated with a shorter mean operating time in this subgroup compared with electrocautery, with a statistically significant difference (WMD −15.00, 95% CI −28.15 to −1.85, P=0.03).
Detailed data on operating time with and without gallbladder perforation were only described by Bessa.6
The majority of patients did not have gallbladder perforation during dissection, and the operating time was found to be statistically shorter in the ultrasonic group knowing that clipless technique was used in this group (WMD −4, 95% CI −6.48 to −1.52, P=0.002). Six of 60 patients had gallbladder perforations in the ultrasonic group compared with 20 of 60 patients in the electrocautery group. To illustrate the effect of gallbladder perforation on the operating time in both comparison groups, the mean operating time (±SD) was 59.2(14) minutes and 61.9(12.2) minutes, respectively, and the difference was not statistically significant (P=0.26).6
Gallbladder perforation with bile leak, stone loss, or both, was discussed in 3 trials.3,6,8
Thirty out of 256 patients had gallbladder perforations with bile loss in the ultrasonic group compared with 86 out of 263 patients in the electrocautery group. This is statistically significant (OR 0.27, 95% CI 0.17 to 0.42, P<0.00001) (
). There is no statistically significant heterogeneity between the trials [ch2
=3.89, dt=2 (P=0.14)]. In the Janssen trial,8
although there was a significant difference in the gallbladder perforation rate with subsequent bile leak between junior trainees (who performed ≤20 procedures) and more senior surgeons, in favor of the seniors when monopolar electrocautery was used for dissection (P=0.026), there was no statistical difference in that outcome between juniors and seniors when ultrasonic dissection was used.
Gallbladder perforation with bile leak or stone loss (p< 0.00001).
The number of gallbladder perforations with stone loss was only described by Janssen et al.8
Three out of 96 patients had gallbladder perforation along with stone loss in the ultrasonic dissection group compared with 20 out of 103 patients in the electrocautery group. This is a statistically significant difference (OR 0.13, 95% CI 0.04 to 0.47, P=0.002).
The number of gallbladder perforations in complicated cases was also described by Janssen et al.8
Examples of complicating factors include adhesions, cystic duct stone(s), and distended gallbladders. Forty-five patients in the Ultracision group (n = 96) had complicating factors compared with 51 in the electrocautery group (n=103). Ultrasonic dissection resulted in a significantly lower number of perforations compared with electrocautery (OR 0.24, 95% CI 0.09 to 0.61, P=0.003).
Postoperative bile leakage was only discussed in 2 trials.3,6
No patients had subhepatic drains or postoperative bile leakage in the ultrasonic group (n=160), while in the electrocautery group (n=160), 3 patients had bile observed in their subhepatic drains reported by Tsimoyiannis et al.3
In 2 patients, bile leakage was observed during the first 24 postoperative hours, while in the third patient, bile leakage continued for 6 days. In all, endoscopic retrograde cholangiopancreatography (ERCP) confirmed bile leakage from the gallbladder's liver bed.3
There is no statistical difference between the 2 groups with regards to this outcome (P=0.19) despite the claim in other reports that Harmonic scalpel dissection of the liver bed can more effectively close the ducts of Luschka11
Number of patients with postoperative bile leakage (p= 0.19).
A total of 13 patients were excluded due to conversion; some from the analysis only2,7
or from the whole trial and randomly replaced by new patients.3
Conversion to open surgery was not considered in another study that looked at clinical outcomes,8
or looked for them but they simply did not occur.6
Trials, such as those of Sietses et al4
and Brokelman et al5
have legitimately ignored this outcome as both have looked at surrogate rather than clinical outcomes. Therefore, only 2 reports have determined the number of conversions in each arm of the study6,7
; hence, we were able to compare the conversion rate, which is not statistically different in a total of 200 patients (
Rate of conversion to open surgery (p= 0.55).
In 2 trials, subhepatic closed drains were left for the first 24 postoperative hours either in all patients,6
or in patients who were likely to ooze blood or to have bile leakage.3
This comparison describes the difference in the numbers of patients who actually needed
subhepatic drainage at the discretion of surgeons, therefore excluding patients from the former trial where drains were routinely inserted as part of the study protocol.6
A total of 26 patients in the ultrasonic dissection group (n=100) had subhepatic drains, compared with 37 patients in the electrocautery dissection group (n=100).8
The difference in the need for subhepatic drains in the immediate postoperative period was statistically insignificant.
The mean length of hospital stay (LHS), in days, is discussed in 5 trials.2–4,6,7
It was determined by the patients needs and speed of postoperative recovery except in 2 trials, where patients were kept in the hospital for 1 or 2 postoperative days as a part of their protocols.4,6
A total of 246 patients had ultrasonic dissection compared with 223 patients who had dissection by monopolar electrocautery. Based on Tsimoyannis et al results,3
there is a statistically significant shorter LHS with ultrasonic dissection (WMD −0.3, 95% CI −0.51 to −0.09, P=0.005) (
). Nevertheless, it is important to emphasise that, in unblinded studies, such as that of the Tsimoyiannis et al trial,3
LHS is prone to unconscious bias, manipulation, or both.
The mean duration of sick leave was discussed in 2 trials.2,7
A total of 77 patients who had ultrasonic dissection were compared with 54 patients who had electrocautery. There is a statistically significant shorter duration of sick leave with ultrasonic dissection compared electrocautery, (WMD −3.8, 95% CI −6.21 to −1.39, P=0.002) (
Mean duration of sick leave in days (p= 0.002).
Postoperative pain scores studied in the Cengiz et al trial7
at the first and fourth hours of recovery are statistically lower with ultrasonic dissection (WMD −1.10, 95% CI-2.16 to −0.04, P=0.04 and WMD −0.80, 95% CI −1.34 to −0.26, P=0.004).7
Pain scores at 24 hours of recovery from Cengiz and Tsimoyiannis trials were combined with a lower estimate in the ultrasonic dissection group (WMD −0.94, 95% CI −1.06 to −0.82, P<0.00001).3,7
Heterogeneity between trials is statistically significant [ch2
=41.18, dt =1, P<0.00001] (
Postoperative pain scores at 24 hours (p< 0.00001).
Postoperative nausea scores at 2, 4, and 24 hours were statistically lower with ultrasonic dissection (WMD −0.90, 95% CI −1.62 to −0.18, P=0.01, WMD −0.80, 95% CI −1.31 to −0.29, P=0.002, and WMD −1.20, 95% CI −2.02 to −0.38, P=0.004, respectively), while there was no statistical difference in the number of patients who experienced a clinically significant postoperative nausea,3
nor was there a statistical difference in the number of patients who suffered from vomiting in the early postoperative period (P=0.65).3
Postoperative suppression of immune function was discussed in detail only in the Sietses et al trial.4
With a small sample size (n=18), he compared the preoperative and postoperative levels of HLA-DR expression within each group and between the 2 dissection groups, with no statistical difference.
In the same trial conducted by Sietses and co-researcher,4
measurement of the postoperative inflammatory response was expressed by the preoperative and postoperative levels of C-reactive Protein (CRP) and the white cell count (WCC). CRP levels were expectedly significantly higher in both groups postoperatively, but postoperative mean CRP levels were not significantly different between the 2 dissection groups (P=0.95).
Preoperative mean WCC were recorded by 2 trials in both the ultrasonic and dissection groups.2,4
There is no statistically significant difference in WCC between ultrasonic and electrocautery dissection groups during the first postoperative day (P=0.58).
Perioperative levels of peritoneal growth transforming factor beta-1 (GTF- β1) expression was only discussed by Brokelman in his trial on GTF-β1 peritoneal expression during laparoscopic cholecystectomy.5
GTF-β1 levels were measured both at the start and at the end of the procedures in 2 randomized groups (n=10 each) with equal intraabdominal pressures and lighting. The finding that ultrasonic scalpel dissection is associated with lower peritoneal total and active GTF-β1 levels compared with electrocautery (P<0.005 and P<0.01, respectively) at the end of the surgery, suggests a reduced risk of peritoneal adhesion formation with the former dissection device.
Postoperative complications were described in 2 studies.3,6
In total, there were 10 patients who experienced postoperative complications out of 320 patients (3.1%). In the ultrasonic group, only 3 of 160 patients had complications (1.9%) compared with 7 in the electrocautery group (4.4%). The difference in the overall postoperative complication rate is not statistically significant (OR 0.45, 95% CI 0.12 to 1.65, P=0.23). There is no significant heterogeneity between the 2 trials [Ch2
=0.99, df=1 (P=0.32)] (
Postoperative complications (p= 0.23).
As described by Tsimoyiannis et al,3
there were no patients with postoperative bile leakage in the ultrasonic group (n=100), compared with 3 patients in the electrocautery group (n=100). In 2 patients, bile leakage was observed during the first 24 postoperative hours, while in the third patient, bile leakage continued for 6 days. In all, endoscopic retrograde cholangiopancreatography (ERCP) confirmed bile leakage from the gallbladder's liver bed.3
In the Bessa et al trial,6
no minor or major bile leaks were reported in the drains postoperatively, but the authors did report port-site and chest infections in 5 and 2 patients, respectively.6
Port-site infections occurred in 2 patients in the ultrasonic group (3.3%) and in 3 others in the electrocautery group (5%). The incidence of chest infection was equal in both groups (1 patient in each arm or 1.6%).6