Keywords Complications, hysterectomy, laparoscopy.

Transcription

1 DOI: /j x General gynaecology A series of 3190 laparoscopic hysterectomies for benign disease from 1990 to 2006: evaluation of complications compared with vaginal and abdominal procedures O Donnez, P Jadoul, J Squifflet, J Donnez Department of Gynecology, Université Catholique de Louvain, Brussels, Belgium Correspondence: Prof J Donnez, Department of Gynecology, Université Catholique de Louvain, Cliniques Universitaires St Luc, Avenue Hippocrate 10, 1200 Brussels, Belgium. jacques.donnez@uclouvain.be Accepted 31 August Published Online 13 November Objective The aim of this study was to evaluate the complication rate after laparoscopic total hysterectomy and laparoscopic subtotal hysterectomy (LASH) in case of benign disease. Design All complications were prospectively recorded at the time of surgery and analysed retrospectively. Setting University hospital. Population Among 4505 hysterectomies performed by the same team using the same techniques between 1990 and 2006, 3190 were performed by laparoscopy, 906 by the vaginal route and 409 by laparotomy. Methods Laparoscopic hysterectomies, defined as laparoscopic subtotal hysterectomy (LASH) and total laparoscopic hysterectomy [laparoscopy-assisted vaginal hysterectomy (LAVH) switched to total laparoscopic hysterectomy (TLH) in 2000], were compared with vaginal and abdominal hysterectomies. Main outcome measures and results Since the early 1990s, the number of laparoscopic procedures has continued to grow, while the number of abdominal and vaginal procedures has decreased. Both minor complications (fever >38.5 C after 2 days, bladder incision of <2 cm and iatrogenic adenomyosis) and major complications (haemorrhage, vesicoperitoneal fistula, ureteral injury, rectal perforation or fistula) have been observed during the surgical procedure itself and postoperatively. In the LASH group (n = 1613), the minor complication rate was 0.99% (n = 16) and the major complication rate 0.37% (n = 6). In the total laparoscopic hysterectomy (LAVH/TLH) group (n = 1577), the minor complication rate was 1.14% (n = 18) and the major complication rate 0.51% (n = 8). In the vaginal hysterectomy group (n = 906), minor and major complication rates were 0.77% (n = 7) and 0.33% (n = 3), respectively. In the abdominal hysterectomy group (n = 409), minor and major complication rates were 0.73% (n = 3) and 0.49% (n =2), respectively. Conclusion The results from our series of 4505 women clearly show that, in experienced hands, laparoscopic hysterectomy is not associated with any increase in major complication rates. Keywords Complications, hysterectomy, laparoscopy. Please cite this paper as: Donnez O, Jadoul P, Squifflet J, Donnez J. A series of 3190 laparoscopic hysterectomies for benign disease from 1990 to 2006: evaluation of complications compared with vaginal and abdominal procedures. BJOG 2009;116: Introduction Laparoscopy is now widely recognised as an indispensable tool in gynaecologic surgery. The first hysterectomy performed by laparoscopy was described by Reich et al. 1 In the following years, Mage et al., 2 Donnez and Nisolle 3 and Lyons, 4 respectively, described their first series of laparoscopy-assisted vaginal hysterectomy (LAVH) and laparoscopic subtotal hysterectomy (LASH). Surgical advantages to laparoscopy are related to the magnification provided in the pelvis, facilitating access to the uterine vessels, ureter, rectum and vagina. Patient advantages are also multiple and are related to the avoidance of a painful abdominal incision. They include shorter hospitalisation and recuperation time and an extremely low rate of infection and ileus. Indications for laparoscopic hysterectomy include numerous benign pathologies, but it may also be considered for endometrial cancer. 5 7 Laparoscopic hysterectomy has been associated with high rates of complications, particularly in terms of urinary tract injuries, compared with abdominal hysterectomy (OR 2.61, 492 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

2 Complications of laparoscopic hysterectomy 95% CI ) Having encountered major complication rates as high as 11.1 and 9.8% with laparoscopic hysterectomy, Johnson et al. 8,9 and Garry et al. 10 concluded that vaginal hysterectomy should be considered the technique of choice; only if vaginal hysterectomy is not possible, should laparoscopic hysterectomy be proposed to avoid abdominal hysterectomy. However, Karaman et al. 11 and Bojahr et al. 12 concluded that laparoscopic hysterectomy is a safe, reproducible technique associated with a low complication rate (between 0 and 1.4%), especially with respect to urinary tract injuries. We report a series of 4505 hysterectomies [LASH, total laparoscopic hysterectomy (TLH) and vaginal and abdominal hysterectomy] for benign disease performed in our department between 1990 and The aim of this study was to evaluate the complication rates between laparoscopic, vaginal and abdominal hysterectomy. Methods From 1990 to 2006, 4505 hysterectomies were performed in our department for benign disease (Figure 1). Hysterectomies were performed in case of uterine fibroids, benign ovarian cysts associated with uterine pathology, adenomyosis, endometriosis, failure of endometrial ablation and/or myomectomy and uterine prolapse (as a step in laparoscopic sacrofixation). Our uterine volume limit for laparoscopic hysterectomy was equivalent to weeks of gestation, unless gonadotrophin-releasing hormone (GnRH) agonist was used preoperatively to decrease the size of the uterus. Hysterectomies for malignant disease, such as endometrial cancer, cervical cancer and ovarian cancer, were not included in the study. Duration of surgery was not statistically analysed because of great variation resulting from the training of fellows and residents; it would not have been truly representative. Of the 4505 operations, 3190 procedures (70.8%) were laparoscopic. Other procedures were vaginal hysterectomy (20.1%, n = 906) and abdominal hysterectomy (9.1%, n = 409). All the operations were performed by the same team using the same reproducible techniques. A total of six surgeons, assisted by residents in training, performed all of these procedures. During surgery, all complications were recorded. Any remarkable events occurring in the course of the operation, as well as immediate postoperative evolution, were noted in the hospital database. All women were reviewed after 4 6 weeks. Data on all surgical procedures were subsequently collected from the hospital database. Surgical procedures listed as abdominal hysterectomy, vaginal hysterectomy, LASH and laparoscopic hysterectomy were included in the study. They were reviewed in detail, and perioperative complications, correction and follow up were noted. Postoperative outcomes were also summarised from the medical file. Immediate postoperative adverse events were recorded, as was follow up until complete resolution of the complication. All emergency readmissions were also noted. Women were examined after 1 year Incidence (%) Laparoscopic procedures Vaginal procedures Abdominal procedures Figure 1. Incidence (%) of laparoscopic, vaginal and abdominal procedures in a series of 4505 hysterectomies for benign disease. ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 493

3 Donnez et al. They were then followed either by a doctor from this department or by a private gynaecologist every 2 years. We classified complications into minor and major complications. Minor complications included fever >38.5 C more than 2 days after surgery requiring intravenous antibiotherapy, bladder incision sutured during surgery with a favourable outcome after placement of a Foley catheter for 7 days and iatrogenic adenomyosis. 13,14 Major complications included conversion to laparotomy (due to serious complications such as major vessel injury), haemorrhage requiring blood transfusion or a second surgical procedure to perform adequate haemostasis and repair of urinary tract injuries or bowel perforation. Surgical procedure Different types of laparoscopic hysterectomies have been described in the literature and at least three types must be considered: 5,13, LAVH: this involves an initial laparoscopic surgical procedure, after which vaginal hysterectomy is carried out. 2 TLH: laparoscopic hysterectomy with laparoscopic dissection to free the uterus of all its attachments in the peritoneal cavity and suture of the vagina either laparoscopically or vaginally. 3 LASH: this entails extraction through a posterior colpotomy (from 1990 to 1993) or uterine morcellation (from 1993 to date). All laparoscopic procedures start in the same way. A Foley catheter is inserted during surgery to empty the bladder. Four laparoscopic puncture sites, including the umbilicus, are used. A cannula is placed in the cervix for appropriate uterine mobilisation. Abdominal and adnexal adhesions, if present, are lysed to mobilise the uterus, and the ureters are identified but not dissected. The round ligaments are then coagulated and removed with scissors. When adnexectomy is required, the infundibulopelvic ligament is also coagulated and cut. Bipolar coagulation is used to coagulate the pedicle, but staples or sutures may be applied. When adnexectomy is not required, the surgical procedure starts by grasping, coagulation and section of the round ligament. The vesicouterine peritoneum is then opened with scissors to expose the posterior leaf of the cardinal ligament. In case of LAVH, the procedure continues with a normal vaginal hysterectomy at this stage. When TLH or LASH is performed, the procedure is as follows. The uterine vessels are clearly identified and, after confirming the position of the ureter, they are desiccated with bipolar coagulation and cut. The same procedure is performed on the opposite side. In our department, staples or specialised disposable equipment are never used because of their very high cost and because they failed to demonstrate any advantage over bipolar forceps. When LASH is performed, a unipolar knife or unipolar scissors (Karl Storz, Tuttlingen, Germany) are used to cut the cervix below the level of the internal os and separate it from the corpus. In our department, until November 1993, longitudinal (vertical or horizontal) posterior colpotomy was performed either by laparoscopy or through the vagina to remove the uterus. Since then, however, the uterus has been removed through a 15-mm trocar after morcellation with Steiner s morcellator or using the new Rotocut (Karl Storz). 18 When TLH is performed, a vaginal incision is made around a uterine mobiliser equipped with a vaginal cupula. The completely freed uterus is then pulled into the vagina and morcellated with a knife, if necessary. The vagina is sutured vaginally or laparoscopically using one or two running sutures between the uterosacral ligaments, making sure that both uterosacral ligaments are sutured together with the final stitch. The pelvis is then rinsed with saline solution. Blood clots are removed. A pelvic drain may be left in the pouch of Douglas for 24 hours. Intravenous antibiotics (cefuroxime 1.5 g and metronidazole 500 mg) are given perioperatively (one shot at the time of anaesthesia induction). Results From 1990 to 2006, 4505 hysterectomies were performed in our department. Of these, 3190 (70.8%) were performed by laparoscopy, 906 (20.1%) by the vaginal route and 409 (9.1%) by laparotomy. In this period, the rate of laparoscopic hysterectomy more than doubled, from 42.2 to 88.2%, while the rate of abdominal hysterectomy fell significantly from 23 to just 4% (P < 0.01) and the rate of vaginal hysterectomy from 34 to 8% (P < 0.001) (Figure 1). Laparoscopic hysterectomy (LAVH and TLH) The laparoscopic hysterectomy group (n = 1577) includes both LAVH and TLH, bearing in mind that we switched from LAVH to TLH in All complications encountered during these procedures are reported in Table 1. We found a 1.59% (n = 25) rate of total complications (minor and major) among the 1577 laparoscopic hysterectomies performed between 1990 and Minor complications The rate of minor complications was 1.14% (n = 18) after laparoscopic hysterectomy. Fever. Twelve women (0.76%) presented with fever >38.5 C, hypogastric tenderness and an inflammatory reaction [elevated C-reactive protein (CRP) and hyperleucocytosis] associated with slight induration of the vaginal vault. Symptoms appeared between 3 days and 3 weeks after surgery. All women recovered normal biological and clinical parameters 494 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

4 Complications of laparoscopic hysterectomy Table 1. Complications in a series of 3190 laparoscopic hysterectomies ( ) LAVH/TLH (n ), n (%) LASH (n ), n (%) Total (n ), n (%) Minor complications 18 (1.14) 16 (0.99) 34 (1.07) Fever.38.5 C (after second day) requiring 5 7 days of antibiotherapy 12 (0.76) 3 (0.19) 15 (0.47) Bladder incision (,2 cm, sutured by laparoscopy) 6 (0.38) 4 (0.25) 10 (0.31) Iatrogenic adenomyoma 0 (0.00) 9 (0.56) 9 (0.28) Major complications 8 (0.51) 6 (0.37) 14 (0.44) Haemorrhage 1 (0.06) 1 (0.06) 2 (0.06) Vesicoperitoneal fistula treated by Foley catheter for 14 days 1 (0.06) 0 (0.00) 1 (0.03) Ureteral lesions 5 (0.32) 3 (0.19) 8 (0.25) Rectal perforation 1 (0.06) 2 (0.12) 3* (0.09) Total minor and major complications 25 (1.59) 22 (1.36) 47 (1.47) *Two of the three women underwent colostomy (one in each subgroup). after 5 7 days of intravenous antibiotherapy (cefuroxime g/day and metronidazole mg/day). None of them complained of deep dyspareunia or recurrent pelvic pain. No vaginal vault abscesses were observed in any of the 1577 laparoscopic hysterectomy cases. Bladder incision. Six women (0.38%) suffered bladder incisions of less than 2 cm, which were all diagnosed during the hysterectomy procedure. Laparoscopic suture was performed, and a Foley catheter was left in place for 6 8 days depending on the size of the incision. All women recovered normal bladder function, and no vesicovaginal fistula was observed. Two women with bladder incisions suffered serious associated ureteral damage due to thermocoagulation (see Urinary tract lesions (ureteral lesions and vesicoperitoneal fistulas)). These two complications are also taken into account in the major complications. Severe adhesions were present in both cases. Major complications The rate of major complications was 0.51% (n = 8) after laparoscopic hysterectomy. All but three of these complications occurred before The major complication rate since 1995 is thus 0.1% (n = 3 of 2745). Urinary tract lesions (ureteral lesions and vesicoperitoneal fistulas). Five women (0.32%) suffered serious ureteral lesions postoperatively. Indeed, between 5 and 10 days after surgery, they presented with relatively acute pelvic pain, abnormal tenderness and distension and an inflammatory reaction in the serum (elevated CRP and leucocytosis). A computed tomography (CT) scan revealed the presence of urine in the peritoneal cavity after intravenous contrast medium injection. Among these five women, four were treated by JJ stent placement. In three women, the JJ stent was removed after 3 months and the ureter recovered normal function. One of these women was followed in another institution, and the JJ stent was retrieved too soon (6 weeks after surgery), resulting in the development of a ureteroperitoneal fistula, which was treated by ureteral reimplantation. The last case of ureteral lesions was a ureterovaginal fistula. One month after surgery, the woman presented with urine loss through the vagina. Intravenous pyelography confirmed the presence of a fistula originating from the lower part of the right ureter. It was impossible to place a JJ stent, and a nephrostomy catheter was therefore used, before performing ureteral reimplantation 3 months later. One vesicoperitoneal fistula (0.06%) was diagnosed 6 days after surgery. The woman presented with acute pelvic pain and ultrasound revealed the presence of liquid in the peritoneal cavity. The diagnosis was made by CT scan. A Foley catheter was left in place for 14 days. Haemorrhage. In 1991, one woman (0.06%) underwent an emergency procedure for postoperative haemorrhage. The night following surgery, the woman presented with low blood pressure and acute pelvic tenderness. More than a litre of blood was observed in the peritoneal drain. Emergency laparoscopy was performed. One and a half litres of blood was found in the peritoneal cavity. Active bleeding from the uterine artery stump was diagnosed, and the stump was coagulated. A blood transfusion (1.5 l) was required. Rectal perforation. In 1992, a 3-cm rectal perforation (0.06%) occurred during surgery. There were extensive adhesions present due to previous pelvic inflammatory disease (PID). The perforation was immediately sutured. After calling for advice, the surgeon decided to perform a colostomy. Laparoscopic subtotal hysterectomy In the LASH group (n = 1613), a total (minor and major) complication rate of 1.36% (n = 22) was observed. Complications after LASH procedures are reported in Table 1. ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 495

5 Donnez et al. Minor complications The rate of minor complications was 0.99% (n = 16) after LASH. Fever. Following LASH procedures, three women (0.2%) presented with fever (>38.5 C after 2 days), but they all showed favourable evolution after intravenous antibiotherapy (cefuroxime g/day and metronidazole mg/day). Bladder incision. Bladder incisions were diagnosed in four women (0.25%). They were laparoscopically sutured, and the postoperative course was unremarkable. All these women had a history of more than two caesarean sections. Severe adhesions were found between the bladder and the isthmic portion of the uterus. Iatrogenic adenomyosis. Iatrogenic adenomyosis was noted in 0.56% of women (n = 9). These lesions were thought to have arisen from residual myometrium and endometrium left in place after morcellation. 13,18 We described eight women with tumours located in the pouch of Douglas, causing deep dyspareunia. Symptoms were due to iatrogenic adenomyomas, specimens of endometrium and myometrium from the morcellated uterus, which had not been removed during LASH. All these women underwent second-look laparoscopy to excise the adenomyotic lesions and to perform trachelectomy. Major complications The rate of major complications was 0.37% (n = 6) after LASH. Urinary tract complications (ureteral lesions and fistulas). Three cases of ureteral lesions (0.19%) were observed during and after LASH. One of these was diagnosed during surgery and was treated by immediate JJ stent placement. The other two were ureteral fistulas diagnosed by CT scan and intravenous pyelography when the women presented with acute pelvic pain and abdominal tenderness (7 and 9 days postoperatively, respectively). They were also treated by JJ stent insertion. These lesions were all due to thermal damage during coagulation of the uterine artery. After 3 months, the JJ stent was removed, and the ureter recovered its normal calibre and function. All these cases occurred before Haemorrhage. One case (0.06%) of serious haemorrhage was observed involving injury (5 mm in length) to the external iliac artery during section of the cervix with unipolar scissors. Emergency laparotomy was required to perform haemostasis and to suture the artery. The recovery was uneventful. Rectal perforation. Two rectal perforations (0.12%) occurred in the LASH group, one of which (2 cm) was detected during surgery. The woman had a history of two laparoscopic procedures for rectovaginal endometriosis, and severe adhesions were found between the rectum and the posterior part of the uterus. Laparoscopic suture of the rectum was performed in two planes, and Tissucol Ò (Johnson & Johnson, Hamburg, Germany) was used to cover the suture. The postoperative course was entirely unremarkable. The second case of rectal perforation occurred 9 days after surgery. The woman presented with acute pelvic pain and signs of faecal peritonitis (distension, ileus, elevated CRP and leucocytosis). A CT scan revealed the presence of air and liquid in the peritoneal cavity, and a laparotomy was carried out. The peritoneal cavity was thoroughly washed out, the rectal defect was closed and a colostomy was performed. Three months later, the colostomy was closed. In this case, rectal perforation resulted from thermal damage during large-scale resection of the uterosacral ligaments due to endometriotic lesions, for which extensive coagulation was required to stop the bleeding. This coagulation was probably responsible for the thermal damage to the rectal wall, with subsequent necrosis. Table 1 summarises all the complications encountered with the different laparoscopic procedures. Taking all laparoscopic hysterectomies into account, a minor complication rate of 1.07% (n = 34) and a major complication rate of 0.44% (n = 14) were observed. The total rate of complications was therefore 1.47% (n = 48). Vaginal hysterectomy In our series of 4505 hysterectomies performed for benign disease, 20.1% (n = 906) were carried out vaginally (Table 2). Minor complications In the vaginal hysterectomy group, fever (>38.5 after 2 days) was observed in three women (0.33%). Bladder incisions Table 2. Complications in a series of 906 vaginal hysterectomies (VH) and 409 abdominal hysterectomies (AH) VH (n 5 906), n (%) AH (n 5 409), n (%) Minor complications 7 (0.77) 3 (0.73) Fever.38.5 C (after second day) 3 (0.33) 0 (0.00) requiring 5 7 days of antibiotherapy Bladder incision 4 (0.44) 3 (0.73) Major complications 3 (0.33) 2 (0.49) Ureteral lesions 3 (0.33) 0 (0.00) Rectal perforation (treated by colostomy) 0 (0) 2 (0.49) Total minor and major complications 10 (1.10) 5 (1.22) 496 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

6 Complications of laparoscopic hysterectomy occurred in four women (0.44%) and were immediately sutured. The rate of minor complications after vaginal hysterectomy was thus 0.77% (n = 7). Major complications No rectal perforations were observed after vaginal hysterectomy. In three women (0.33%), ureteral lesions were encountered. One involved ureteral section and the other two, ureteral ligation. All the three women were treated by ureteral reimplantation. Abdominal hysterectomy In our series of 4505 hysterectomies performed for benign disease, 9.1% (n = 409) were carried out by laparotomy (Table 2). In all women, laparotomy was indicated because of the size of the uterus. In our department, when uterine size exceeds weeks of gestation, a hysterectomy is systematically carried out through a Pfannenstiel incision. Minor complications Bladder incisions were observed in 0.73% (n = 3) of women and sutured during surgery. Major complications No ureteral lesions were observed in the abdominal hysterectomy group. Rectal perforation was observed in 0.49% (n =2) of women. In one case, the rectum was sutured during surgery. In the other, the woman presented to the emergency department 7 days later with signs of faecal peritonitis, and a colostomy was carried out. Discussion Since the first laparoscopic hysterectomy was described by Reich et al. back in and the first subtotal hysterectomy was performed in the early 1990s, 3,4 the laparoscopic approach has become the procedure of choice in gynaecological surgery. Numerous advantages for the patient (less pain and discomfort, quicker recovery time and shorter hospital stay) and the surgeon (magnification of the pelvic anatomy) are detailed in the literature. Complication rates in the literature: the Cochrane Review, a source of bias Laparoscopic procedures are associated with varying rates of complications. Johnson et al. 8,9 conducted a review of all existing randomised trials comparing abdominal, vaginal and laparoscopic techniques in the context of the Cochrane Collaboration. They observed that there were more urinary tract injuries with laparoscopic than with abdominal hysterectomy (OR 2.61, 95% CI ). The increased risk of urinary tract injuries was essentially seen in the evaluate study published by Garry et al. 10 in 2004, which involved two parallel randomised trials including 1380 women to evaluate the effects of laparoscopic hysterectomy compared with abdominal and vaginal hysterectomy. We critically analysed data from the evaluate study in a letter to the editor 19 published in the BMJ and showed that considerable bias in their method unfortunately led Garry et al. to what are probably erroneous conclusions. Their primary end-point was the occurrence of major complications, which was as high as 11.1 and 9.8% in the laparoscopic hysterectomy groups of the two trials. We agree that such rates are totally unacceptable but, in our opinion, they are a direct result of the major bias in the evaluate study. Indeed, 43 gynaecologists from 30 centres took part, so that the mean number of laparoscopic hysterectomies (n = 920) per gynaecologist was just 21 over 4 years. According to the literature, 20 the learning curve should exceed 21 cases. In our department, 3190 laparoscopic hysterectomies were performed by the same team between 1990 and 2006, giving an average of 200 laparoscopic hysterectomies per year by the same surgical team (Table 1), which is far more than required for the learning curve. The experience of the 43 gynaecologists in the evaluate study most certainly differed from centre to centre, but the rate of complications was not analysed according to the gynaecologists experience. The high complication rates found were probably due more to the relative inexperience of the surgeons in laparoscopic hysterectomy than to the technique of laparoscopic hysterectomy itself. Indeed, the experience and skill of the surgeons involved obviously differed, with some of them still in the learning curve. It is likely that more experienced laparoscopists would never have been associated with such high complication rates. Complication rate after laparoscopic hysterectomy: comparison with vaginal and abdominal hysterectomy Initially, the Cochrane Review published by Johnson et al. in and concluded that vaginal hysterectomy should be performed in preference to abdominal hysterectomy where possible; if not possible, a laparoscopic approach may be used to avoid abdominal hysterectomy. The authors considered that laparoscopic hysterectomy was a longer operation, carrying a greater risk of damage to the bladder or ureter. In our series, however (Tables 1 and 2), ureteral lesions were observed in 0.33% of women (n = 3) (95% CI ) after vaginal hysterectomy and 0.25% of women (n = 8) (95% CI ) after laparoscopic procedures (both LAVH/TLH and LASH). These results are not significantly different. Bladder injuries were observed in 0.44% of women (n = 4) (95% CI ) after vaginal hysterectomy and 0.31% of women (n = 10) (95% CI ) after laparoscopic procedures. This clearly shows that the risk of ureteral and bladder damage after laparoscopic hysterectomy is similar to that found after vaginal hysterectomy. Most of our major ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 497

7 Donnez et al. complications occurred before 1995 and were due to thermal damage to the ureter or bladder. Careful identification of the ureter (and subsequent dissection if necessary) was thus required before coagulation and section of the uterine blood vessels along the ascending part of the uterus. It is important to note that serious complications relating to thermal damage can happen, even if the rate of complications is low. For this reason, the laparoscopic technique requires a high level of expertise and perfect knowledge of all the safety rules. In our series (Tables 1 and 2), a complication rate (major and minor) of 1.59% (95% CI ) was observed after TLH and 1.36% (95% CI ) after LASH. Complication rates after vaginal and abdominal hysterectomy were 0.66% (95% CI ) and 1.22% (95% CI ), respectively. None of these complication rates differ significantly when surgery is performed by the same team using reproducible surgical techniques. Our complication rates are comparable with those found in the literature. In a series of 1706 laparoscopic supracervical hysterectomies, Bojahr et al. 12 obtained a urinary tract injury rate of just 0.23% (n = 4). We agree with their conclusion that LASH is a minimally invasive surgical approach associated with low perioperative morbidity and rapid convalescence. Karaman et al. 11 observed an overall major complication rate of 1% with no urinary tract injuries at all in a series of 1120 laparoscopic hysterectomies. More recently, Brummer et al. 21 observed low complication rates in a multicentric retrospective study of laparoscopic hysterectomies in Finland. From 2000 to 2005, urinary tract injuries decreased from 1.4 to 0.7%, and ureteral injuries decreased from 0.9 to 0.3%. They also concluded that laparoscopic hysterectomy is a good and safe option for hysterectomy. Concerning the risk of conversion to laparotomy, David- Montefiore et al. 22 observed, in a prospective study, that conversion was more frequently required during laparoscopic hysterectomy (19%) than vaginal hysterectomy (4.2%) (P < ). However, among the 12 centres taking part in the study, only 1 centre regularly performed laparoscopic surgery, while 10 centres mainly used the vaginal route. The mean uterine weight in the laparoscopic surgery group (230 ± 185 g) was not especially high compared with the vaginal route group (226 ± 203 g). In the publications of Garry et al. and Johnson et al., a conversion rate from laparoscopy to laparotomy of 3.5% (32/920) seems very high to us, given a mean uterine size equivalent to only 6 (0 12) weeks of gestation. Only one conversion to laparotomy was needed in our series of 3190 laparoscopic hysterectomies for a serious haemorrhage from the external iliac artery. Conversion was never required due to uterine size because careful vaginal examination of uterine volume and uterine mobility was performed prior to surgery. Our uterine volume limit for laparoscopic hysterectomy is equivalent to weeks of gestation, unless GnRH agonist is used preoperatively to decrease the size of the uterus. Complication rates and comparison between LASH and TLH In a comparison of the two different laparoscopic techniques (LAVH/TLH and LASH) used in our series (Table 1), the complication rates were also found to be similar. In the LAVH/ TLH group, a major complication rate of 0.51% (n = 8) (95% CI ) was observed compared with 0.37% (n = 6) (95% CI ) in the LASH group. These results are not significantly different. Total complication rates, 1.59% (95% CI ) and 1.36% (95% CI ) in the LAVH/ TLH and LASH groups, respectively, were also similar. According to Kilkku et al. 23 and Virtanen et al. 24, libido and orgasmic frequency are not affected by subtotal hysterectomy but are significantly reduced by total hysterectomy. However, Thakar et al. 25 demonstrated a clear improvement in quality of life after both total and subtotal hysterectomy and no difference in pelvic organ function. Lethaby et al. 26 did not confirm the perception that subtotal abdominal hysterectomy offers improved outcomes for sexual, urinary or bowel function compared with total abdominal hysterectomy. El-Toukhy et al. 27 were the only authors in the literature to compare urinary and sexual function after abdominal, vaginal, laparoscopic total hysterectomy and LASH. No adverse effect on sexual function was observed with any of these techniques 6 months after surgery. In a randomised controlled trial, Kluivers et al. 28 observed that laparoscopic hysterectomy is superior to abdominal hysterectomy with respect to postoperative symptoms. The authors also reported that laparoscopic hysterectomy results in greater postoperative vitality compared with abdominal hysterectomy. 29 Conclusions We firmly believe that laparoscopic hysterectomy offers multiple advantages over abdominal and vaginal hysterectomy and that laparoscopy is clearly the most appropriate technique. The findings from our series of 3190 women plainly show that LASH and TLH are safe procedures with complication rates of just 1.36% (95% CI ) and 1.59% (95% CI ), respectively, compared with 1.10% (95% CI ) and 1.22% (95% CI ), respectively, after vaginal and abdominal hysterectomy. Therefore, there is no statistically significant difference in complication rates when surgery is performed by the same team using reproducible surgical techniques. Do we really need randomised trials to endorse the value of laparoscopic hysterectomy vis-à-vis other techniques, after encountering only 0.44% of major complications among 3190 laparoscopic procedures? 498 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

8 Complications of laparoscopic hysterectomy In the USA and in Europe, however, 75% of hysterectomies are still performed by an abdominal approach. If laparoscopic hysterectomy were more fully incorporated into our surgical armamentarium, almost all hysterectomies (95%) could be carried out without an abdominal incision. In our department, the rate of abdominal hysterectomy decreased from 23% in the early 1990s to less than 3.8% in 2006 (Figure 1). In experienced hands, laparoscopic hysterectomy is a safe, cost-effective and beneficial technique for both the patient and the surgeon. Expertise in laparoscopic procedures and adherence to the safety rules are nevertheless paramount to avoid any serious complications that may occur. Abdominal, vaginal and laparoscopic procedures all have their own indications and need to be performed with the same level of application and skill. Disclosure of interests The authors have nothing to disclose. Contribution to authorship O.D., P.J., J.S. and J.D. are surgeons and have collaborated in the surgical procedures and clinical follow up. They have participated in the preparation of the manuscript and its discussion. O.D. collected data, conducted statistical analysis and wrote the manuscript. Details of ethics approval No ethics approval was needed. Funding No funding was received for this study. Acknowledgements The authors thank Mira Hryniuk, BA, and Anne Lepage for reviewing and correcting the manuscript. j References 1 Reich H, De Caprio J, Mac Glynn F. Laparoscopic hysterectomy. J Gynecol Coll Surg 1989;5: Mage G, Wattiez A, Chapron C, Canis M, Pouly JL, Pingeon JM, et al. Hystérectomie per-coelioscopique: résultats d une série de 44 cas. J Gynecol Obstet Biol Reprod 1992;21: Donnez J, Nisolle M. Laparoscopic supracervical (subtotal) hysterectomy (LASH). J Gynecol Surg 1993;9: Lyons TL. Laparoscopic supracervical hysterectomy: a comparison of morbidity and mortality results with laparoscopic assisted vaginal hysterectomy. J Reprod Med 1993;38: Reich H. New techniques in advanced laparoscopic surgery. Clin Obstet Gynecol 1989;3: Querleu D, Leblanc E, Castelain G. Laparoscopic pelvic lymphadenectomy in the staging of early carcinoma of the cervix. Am J Obstet Gynecol 1991;164: Reich H, McGlynn F, Wickie W. Laparoscopic management of stage 1 ovarian cancer: a case report. J Reprod Med 1990;35: Johnson N, Barlow D, Lethaby A, Tavender E, Curr L, Garry R. Methods of hysterectomy: systematic review and meta-analysis of randomized controlled trials. BMJ 2005;330: Johnson N, Barlow D, Lethaby A, Tavender E, Curr L, Garry R. Surgical approach to hysterectomy for benign gynaecological disease (Review). Cochrane Database Syst Rev 2006;CD Garry R, Fountain J, Mason S, Hawe J, Napp V, Abbott J, et al. The evaluate study: two parallel randomized trials, one comparing laparoscopic with abdominal hysterectomy, the other comparing laparoscopic with vaginal hysterectomy. BMJ 2004;328: Karaman Y, Bingol M, Günenc Z. Prevention of complications in laparoscopic hysterectomy: experience with 1120 cases performed by a single surgeon. J Minim Invasive Gynecol 2007;14: Bojahr B, Raatz D, Schonleber G, Abri C, Ohlinger R. Perioperative complication rate in 1706 patients after a standardized laparoscopic supracervical hysterectomy technique. J Minim Invasive Gynecol 2006;13: Donnez O, Jadoul P, Squifflet J, Donnez J. Iatrogenic peritoneal adenomyoma after laparoscopic subtotal hysterectomy and uterine morcellation. Fertil Steril 2006;86: Donnez J, Squifflet J, Donnez O, Smets M, Jadoul P. LASH: laparoscopic subtotal hysterectomy a series of 1400 cases. In Donnez J, editor. An Atlas of Operative Laparoscopy and Hysteroscopy. Oxon, UK: Informa Healthcare; pp Donnez J, Nisolle M, Anaf V. Place de l endoscopie dans le cancer de l endomètre. In: Dubuisson JB, Chapron CH, Bouquet de Jolinière J, editors. Coelioscopie et Cancérologieen Gynécologie. Paris: Arnette; pp Donnez J, Smets M, Squifflet J, Jadoul P. Laparoscopic hysterectomy in benign disease: a series of 1233 cases. In: Donnez J, editor. An Atlas of Operative Laparoscopy and Hysteroscopy. Oxon, UK: Informa Healthcare; pp Johns A. Laparoscopic assisted vaginal hysterectomy (LAVH). In: Sutton C, Diamond D, editors. Gynecologic Endoscopy for Gynecologists. London, UK: Saunders; pp Donnez O, Jadoul P, Squifflet J, Donnez J. Post-hysterectomy pelvic adenomyotic masses observed in 8 cases out of a series of 1405 laparoscopic subtotal hysterectomies. J Minim Invasive Gynecol 2007;14: Donnez J, Squifflet J, Jadoul P, Smets M. High rate of complications needs explanation (Letter). BMJ 2004;328: Wattiez A, Soriano D, Cohen SB, Nervo P, Canis M, Botchorishvili R, et al. The learning curve of total laparoscopic hysterectomy: comparative analysis of 1647 cases. J Am Assoc Gynecol Laparosc 2002;9: Brummer T, Seppälä T, Härkki PS. National learning curve for laparoscopic hysterectomy and trends in hysterectomy in Finland Hum Reprod 2008;23: David-Montefiore E, Rouzier R, Chapron C, Daraï D; the Collégiale d Obstétrique et Gynécologie de Paris-Ile de France. Surgical routes and complications of hysterectomy for benign disorders: a prospective observational study in French university hospitals. Hum Reprod 2007; 22: Kilkku P, Gronroos M, Rauramon L. Supravaginal amputation with preoperative electrocoagulation of endocervical mucosa. Description of the method. Acta Obstet Gynecol Scand 1985;64: Virtanen H, Makinen J, Tenho T, Kiilholma P, Pitkänen Y, Hirvonen T. Effects of abdominal hysterectomy on urinary and sexual symptoms. Br J Urol 1993;72: ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 499

9 Donnez et al. 25 Thakar R, Ayers S, Clarkson P, Stanton S, Manyonda I. Outcomes after total versus subtotal abdominal hysterectomy. N Engl J Med 2002;347: Lethaby A, Ivanova V, Johnson NP. Total versus subtotal hysterectomy for benign gynaecological conditions. Cochrane Database Syst Rev 2006;CD El-Toukhy TA, Hefni M, Davies A, Mahadevan S. The effect of different types of hysterectomy on urinary and sexual functions: a prospective study. J Obstet Gynaecol 2004;24: Kluivers K, Mol B, Bremer G, Brölmann H, Vierhout M, Bongers M. Pelvic organ function in randomized patients undergoing laparoscopic or abdominal hysterectomy. J Minim Invasive Gynecol 2007; 14: Kluivers K, Hendriks J, Mol B, Bongers M, Bremer G, de Vet H, et al. Quality of life and surgical outcome after total laparoscopic hysterectomy versus total abdominal hysterectomy for benign disease: a randomized, controlled trial. J Minim Invasive Gynecol 2007;14: ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology