Surgical routes and complications of hysterectomy for benign disorders: a prospective observational study in French university hospitals

Transcription

1 Human Reproduction Vol.22, No.1 pp , 2007 Advance Access publication September 1, doi: /humrep/del336 Surgical routes and complications of hysterectomy for benign disorders: a prospective observational study in French university hospitals E.David-Montefiore 1, R.Rouzier 1, C.Chapron 2, E.Daraï 1,3 and the Collegiale d Obstétrique et Gynécologie de Paris-Ile de France 1 Service de Gynécologie-Obstétrique, Hôpital Tenon, Université Pierre et Marie Curie-Paris VI and 2 Service de Chirurgie Gynécologique, Hôpital Baudelocque-Port-Royal, AP-HP, Université Paris V, Paris, France 3 To whom correspondence should be addressed at: Service de Gynécologie-Obstétrique, Hôpital Tenon, 4 Rue de la Chine, Paris, France. emile.daraï@tnn.ap-hop-paris.fr Services de Gynécologie Obstétrique: Prof. Fernandez and Prof. Friedman, Hôpital Antoine Béclère; Prof. Deval and Prof. Levardon, Hôpital Beaujon; Dr Dhainaut and Prof. Madelénat, Hôpital Bichat; Prof. Paniel, Hôpital Intercommunal de Créteil; Prof. Sibony and Prof. Oury, Hôpital Robert Debré; Prof. Lecuru, Hôpital Européen Georges Pompidou; Dr Ansquer and Prof. Mandelbrot, Hôpital Louis Mourrier; Prof. Fauconnier and Prof. Ville, Hôpital de Poissy; Prof. Bénifla, Hôpital Rothschild; Dr Poncelet and Prof. Cohen-Uzan, Hôpital Jean Verdier BACKGROUND: Despite the advantages of the vaginal and laparoscopic approaches, most hysterectomies carried out involve laparotomy. The objective of this prospective observational multicentre study was to examine the routes and complications of hysterectomy for benign disorders. METHODS: Of the 15 university hospitals belonging to Collégiale de Gynécologie-Obstétrique de Paris-Ile de France, 12 participated in this study that took place between June and December We analysed the characteristics of the patients, the indications for hysterectomy and intra- and postoperative complications (and their determinants) according to the surgical approach. RESULTS: In total, 634 women underwent hysterectomy for benign disorders during the study period. The patients mean age (±SD), BMI, parity and previous Caesarean sections were 51.4 ± 10.3 years, 25 ± 5.7 kg/m 2, 2 ± 1.6 children and 0.2 ± 0.6, respectively. Hysterectomy was performed by the laparoscopic, laparoscopically assisted vaginal hysterectomy (LAVH), laparotomic and vaginal routes in 19.1, 8.2, 24.4 and 48.3% of cases, respectively. The operating time was shorter with the vaginal route than with laparoscopy, laparotomy and LAVH (P < ). Intra- and post-operative complications were significantly more frequent in the laparotomic group (18%) compared with the vaginal group (8.2%), the laparoscopic group (5.8%) and the LAVH group (8.2%) (P < ). In a multivariate logistic regression model, obesity [odds ratio (OR): 2.84, 95% confidence interval (CI): , P = 0.001], history of pelvic surgery (OR: 2.47, 95% CI: , P = 0.002) and history of Caesarean section (OR: 2.04, 95% CI: , P = 0.046) were significantly associated with intra- and post-operative complications. Laparoconversion was necessary in 36 cases (7.5%) overall and was more frequent with laparoscopy and LAVH than with the vaginal route (P < ). CONCLUSIONS: This study confirms that the vaginal route is increasingly used for hysterectomy in France and that it is the route of choice for benign disorders. Key words: benign disorders/france/hysterectomy/laparoscopy/laparotomy Introduction About hysterectomies are performed each year in France, usually for benign disorders (uterine fibroids in onethird of cases) (Cosson et al., 1997). About hysterectomies are performed annually in the UK (University of York, 1991) and over in the USA (National Center for Disease Control and Prevention, 1997). Randomized studies comparing different routes suggest that vaginal or laparoscopic hysterectomy is associated with a shorter hospital stay and faster recovery than laparotomic hysterectomy (Johnson et al., 2005a). However, observational studies (Vessey et al., 1992; Harris and Olive, 1994; Dorsey et al., 1995; Johns et al., 1995; Davies et al., 1998; Härkki-Siren et al., 1998; Chapron et al., 1999) show that hysterectomy is still generally performed by laparotomy. A recent meta-analysis of 27 randomized trials involving a total of 3643 women (Johnson et al., 2005b) showed that women returned to their normal activities more quickly after 260 The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 Different gonadotrophin preparations in intrauterine insemination cycles vaginal or laparoscopic hysterectomy than after abdominal hysterectomy, whereas no difference was found between laparoscopic and vaginal hysterectomies. Also, more urinary tract injuries occurred during laparoscopic hysterectomy than during abdominal hysterectomy. The authors concluded that vaginal hysterectomy should be preferred to abdominal hysterectomy whenever possible. The aims of this study were to determine the frequency of use of the different routes in women undergoing hysterectomy for benign disorders, according to the patients epidemiological characteristics. We also compared the rates of complications and their determinants. Materials and methods This observational prospective multicentre study took place between June and December 2004 in the gynaecology and obstetrics departments of university hospitals belonging to Collégiale de Gynécologie- Obstétrique de Paris-Ile de France. Twelve (80%) of the 15 member hospitals agreed to participate in the study. All women with benign gynaecological disorders qualifying for hysterectomy were included in the study. The following epidemiological characteristics were recorded: age, BMI, parity, menopausal status and prior Caesarean section and/or pelvic surgery. The route chosen for hysterectomy depended on the surgeon s experience and the practices of each participating centre. Women with malignancies were excluded from the study, even when diagnosed during or after the study procedure. Uterine weight >280 g, a history of pelvic inflammatory disease, moderate or severe endometriosis and indications for adnexectomy were not considered to contraindicate vaginal or laparoscopic hysterectomy. All the women received prophylactic antibiotics at the beginning of the operation and prophylactic anticoagulation (low-molecular-weight heparin) the evening before the operation. The women were divided into four groups depending on the surgical approach: the laparotomic group consisted of women who underwent hysterectomy via a suprapubic or median incision; the laparoscopically assisted vaginal hysterectomy (LAVH) group consisted of women who underwent vaginal hysterectomy assisted by laparoscopic procedures (excluding uterine artery ligation); laparoscopic hysterectomy consisted of laparoscopic procedures including uterine artery ligation; and vaginal hysterectomy consisted of procedures performed with or without wedge morcellation, coring or bivalving. Laparoconversion was defined as any laparotomy procedure performed for any reason in the vaginal, laparoscopic and LAVH groups. The indications for laparoconversion were recorded. The choice for the route of hysterectomy was made in each participating centre depending on department options and experience of surgeons. Post-operative fever was defined by a body temperature of at least 38 C on two consecutive occasions at least 6 h apart, excluding the first 24 h. The operating time, the incidence of intra- and post-operative complications, transfusion requirements, the post-operative hospital stay and uterine weight were recorded in every case. The women were re-examined 6 8 weeks after surgery. Data on immediate and short-term post-operative outcomes were collected from the hospital and outpatient medical records. Owing to variations in anaesthetic and analgesic protocols from one centre to another, no attempt was made to evaluate the consumption of oral non-steroidal anti-inflammatory drugs (NSAIDs) or intramuscular opioids. The time to recovery was not evaluated. Data were collected by one clinical research assistant per centre and were centralized at Tenon hospital. Obesity was defined as BMI >30 kg/m 2. Categorical variables were compared between groups using the chi-square or Fisher s exact test, as appropriate. For continuous variables, statistical comparisons were performed using analysis of variance (ANOVA). Pairwise t-tests with Bonferroni s correction were used for post hoc tests. Age, obesity, parity, menopausal status, prior Caesarean section and/or pelvic surgery, indication and concomitant adnexectomy were tested as risk factors for complication in a multivariate logistic regression model with forward selection of variable. All tests were two tailed, and P values <0.05 were considered to denote significant differences. All analyses were performed using the R package with the Design, Hmisc and Lexis libraries. ( lib.stat.cmu.edu/r/cran). Results Epidemiological characteristics and indications From June to December 2004, 634 patients underwent hysterectomy for benign disorders. Mean age was 51.4 ± 10.3 years, mean BMI 25 ± 5.7 kg/m 2, mean parity 2 ± 1.6 children and the mean number of Caesarean sections per patient 0.2 ± 0.6. Twenty-nine percent of the women were post-menopausal and 27% had a history of pelvic surgery. Hysterectomy was total and subtotal in 600 (94.6%) and 34 patients (5.4%), respectively. Salpingo-oophorectomy was performed in 221 patients (34.9%). The indications for hysterectomy are summarized in Table I. The main indication was dysfunctional uterine bleeding (DUB), followed by symptomatic myomas and genital prolapse. Hysterectomy was performed by laparoscopy, LAVH, laparotomy and the vaginal route 19.1, 8.2, 24.4 and 48.3% of cases, respectively. The distribution of the different routes in the 12 centres is summarized in Table II. One centre used laparoscopy almost exclusively, one centre preferentially used laparotomy and the other 10 centres mainly used the vaginal route. The indications of hysterectomy differed among groups. We defined three main indications: DUB, pain/discomfort and genital prolapse. Indications according to the surgical approach are summarized in Table III. Genital prolapse was mainly treated using vaginal route. Mean uterine weight in the laparotomic, laparoscopic, LAVH and vaginal groups was 723 ± 1320, 280 ± 229, 230 ± 185 and 226 ± 203 g, respectively. Mean uterine weight differed among the groups (P < ). It was significantly Table I. Indications for hysterectomy (634 patients) Indications Patients, Adenomyosis 29 (4.6) Chronic pelvic pain 55 (8.7) Endometriosis 6 (0.9) Symptomatic myomas 186 (29.4) Endometrial hyperplasia 16 (2.5) Cervical intra-epithelial neoplasia 6 (0.9) Non-suspect adnexal mass 17 (2.7) Genital prolapse 107 (16.9) Dysfunctional uterine bleeding 202 (31.9) Ovarian borderline lesion 8 (1.3) Cervical pregnancy 1 (0.2) Haemostasis (post-caesarean section) 1 (0.2) 261

3 E.David-Montefiore et al. Table II. Rates of laparotomic, laparoscopically assisted vaginal hysterectomy (LAVH), laparoscopic and vaginal hysterectomy in 12 French university hospitals Centre Hysterectomy, Laparoscopy, LAVH, Laparotomy, Vaginal, 1 70 (11) 0 4 (5.7) 11 (15.7) 55 (78.6) 2 64 (10) 2 (3.1) 9 (14) 9 (14) 64 (68.8) 3 61 (9.6) 0 2 (3.3) 27 (44.3) 32 (52.5) 4 55 (8.7) 4 (7.3) 11 (20) 14 (25.5) 26 (47.3) 5 46 (7.2) 1 (2.1) 3 (6.5) 13 (28.3) 29 (63) 6 19 (3) 3 (15.8) 2 (10.5) 4 (21.1) 10 (52.6) 7 63 (9.9) 9 (14.3) 10 (15.9) 12 (19) 32 (50.8) (17.9) 92 (80.7) 0 10 (8.8) 12 (10.5) 9 10 (1.6) 0 1 (10) 5 (50) 4 (40) (4.4) 3 (10.7) 4 (14.3) 16 (57.1) 5 (17.9) (12.6) 7 (8.8) 6 (7.5) 25 (31.3) 42 (52.5) (3.8) (37.5) 15 (62.5) Total Table III. Distribution of the main indications for hysterectomy according to the routes Laparoscopy LAVH Laparotomy DUB, dysfunctional uterine bleeding; LAVH, laparoscopically assisted vaginal hysterectomy. Vaginal Pain/discomfort Prolapse DUB higher in the laparotomic group than in the laparoscopic group (P < ), the vaginal group (P < ) and the LAVH group (P = ). It was also significantly higher in the laparoscopic group than in the vaginal group (P = 0.01). No difference in uterine weight was found between the LAVH group, the laparoscopy and the vaginal group. Epidemiological characteristics according to the surgical approach Mean age differed according to the route (P = ). The women in the vaginal group were significantly older than those in the laparoscopic group (P = ), the LAVH group (P = 0.01) and the laparotomic group (P = 0.001). Mean age did not differ between the laparoscopic and LAVH groups, the laparoscopic and laparotomic groups or the LAVH and laparotomic groups (Table IV). Mean BMI differed according to the surgical approach (P = 0.03). BMI was significantly lower in the laparoscopic group than in the laparotomic group (P = 0.05). The other post hoc t-tests adjusted with Bonferroni correction were not statistically different. Mean parity differed according to the surgical approach (P < ). Parity was significantly higher in the vaginal group than in the laparoscopic group (P < ), the LAVH group (P = 0.003) and the laparotomic group (P < ). Nulliparous patients were statistically less frequent in the vaginal group (P < ). The proportion of post-menopausal women differed among the groups (P = 0.006). The frequencies of prior Caesarean section and prior pelvic surgery differed among the groups (P = and P < , respectively). Patients operated on by vaginal route were less likely to have had prior pelvic surgery. Operating times and intra- and post-operative complications The mean operating time differed among the groups (P < ). It was significantly shorter in the vaginal group than in the laparoscopic group (P < ), the laparotomic group (P < ) and the LAVH group (P = ). No difference was found among the laparoscopic, laparotomic and LAVH groups. Intra- and post-operative complications were significantly more frequent in the laparotomic group (18%) compared with the vaginal group (8.2%), the laparoscopic group (5.8%) and the LAVH group (5.8%) (P < ) (Table V). In a multivariate logistic regression model, obesity [(odds ratio (OR): 2.84, 95% confidence interval (CI): , P = 0.001)], history of pelvic surgery (OR: 2.47, 95% CI: , P = 0.002) and history of Caesarean section (OR: 2.04, 95% CI: , P = 0.046) were significantly associated with intra- and postoperative complications. We then tested the interaction between surgical approach and risk factors for complications in Table IV. Epidemiological characteristics of the patients undergoing laparotomic, LAVH, laparoscopic or vaginal hysterectomy Laparoscopic group (n = 121) LAVH group (n = 52) Laparotomic group (n = 155) Vaginal group (n = 306) P Age ± SD (years) 49.3 ± ± ± ± BMI ± SD 24.1 ± ± ± ± Parity ± SD 1.6 ± ± ± ± 1.6 < Nulliparous < Menopausal status Prior Caesarean sections per patient ± SD 0.2 ± ± ± ± Prior Caesarean sections Prior pelvic surgery < Operating time ± SD (min) ± ± ± ± 43.1 < Uterus weight ± SD (g) 280 ± ± ± ± 203 < Hospital stay ± SD (days) 4.6 ± 2 5 ± ± ± 1.6 < LAVH, laparoscopically assisted vaginal hysterectomy. 262

4 Different gonadotrophin preparations in intrauterine insemination cycles Table V. Intra- and post-operative complications in the laparotomic, LAVH, laparoscopic and vaginal hysterectomy groups Complications Laparoscopy (n = 121) LAVH (n = 52) Laparotomy (n = 155) Vaginal (n = 306) Intra-operative complication Intestinal injury 1 (0.8) (0.7) Urinary tract injury 1 (0.8) 1 (1.9) 2 (1.3) 6 (2) Haemorrhage 1 (0.8) 0 1 (0.6) 3 (1) Post-operative complication Parietal abscess/haematoma (3.9) 0 Pelvic abscess (1.3) 1 (0.3) Pelvic haematoma (1.3) 1 (0.3) Haemorrhage 0 1 (1.9) 0 4 (1.3) Hyperthermia 2 (1.7) 0 3 (1.9) 1 (0.3) Urinary retention 0 1 (1.9) 1 (0.6) 0 Anal incontinence (0.6) 0 Phlebitis (0.3) Intestinal occlusion 1 (0.8) 0 2 (1.3) 0 Rupture of suction drain (0.6) 0 Urinary tract infection (1.9) 5 (1.6) Blood transfusion 1 (0.8) 0 5 (3.2) 2 (0.7) Total LAVH, laparoscopically assisted vaginal hysterectomy; P < the multivariate model and found that laparotomy route was associated with greater morbidity in obese women (P = 0.01), whereas there was no significant interaction between surgical route and history of pelvic surgery or Caesarean section. Repeat surgery was required in 14 cases (2.2%) for haemorrhage in five cases, haematoma in four cases, intestinal occlusion in two cases, pelvic abscess in one case and recovery of a suction drain and of a surgical sponge in one case each. The frequency of repeat surgery tended to differ among the groups (P = 0.08). Eight women required blood transfusion, although the frequency did not differ among the groups (P = 0.13). Conversion to laparotomy and its determinants according to the surgical approach Conversion to laparotomy was necessary in 36 cases (7.5%). Thirteen laparoconversions (4.2%) were required in the vaginal group and 23 (19%) in the laparoscopic and LAVH groups combined (P < ). The indications for conversion are summarized in Table VI. In a multivariate logistic regression model, obesity (OR: 3.11, 95% CI: , P = 0.01), history of pelvic surgery (OR: 2.82, 95% CI: , P = 0.02) and uterus weight >500 g (OR: 3.86, 95% CI: , P = 0.01) were significantly associated with laparoconversion. We did not find any interaction between surgical route and risk factors for laparoconversion. The rates of conversion to laparotomy are summarized in Table VII according to the route and centre. The vaginal hysterectomy conversion rate was similar across the centres. In contrast, the conversion rate was higher in centres with little experience of laparoscopy. Hospital stay The mean hospital stay in the laparotomic, laparoscopic, LAVH and vaginal groups was 7.6 ± 5.1, 4.6 ± 2, 5 ± 1.5 and 4.8 ± 1.6 days, respectively. The hospital stay differed among the groups (P < ). The pairwise comparisons using t tests with Bonferroni correction revealed that intergroup significant differences were between the laparotomic group and other groups (laparoscopic group, P < ; LAVH, P < and vaginal group, P < ). Table VI. Indications for the 36 conversions to laparotomy Indication Patients, Adhesions 4 (10.3) Insufflation failure 2 (5.1) Gas embolism 1 (2.6) Endometriosis 2 (5.1) Poor accessibility 5 (12.8) Haemorrhage 9 (23.1) Suspect adnexal mass 3 (7.7) Unspecified technical problem 1 (2.6) Uterus volume 9 (23.1) Table VII. Rates of conversion to laparotomy in the 12 university hospitals Centre Laparoscopy group, LAVH group, Vaginal group, 1 NA 0 3 (5.5) 2 2 (100) 0 5 (7.8) 3 NA (75) 0 1 (3.8) (3.4) 6 2 (67) (22) (4.3) NA 0 9 NA (100) (71) 0 2 (4.8) 12 NA NA 0 LAVH, laparoscopically assisted vaginal hysterectomy; NA, not applicable. 263

5 E.David-Montefiore et al. Discussion This prospective observational study of university hospitals in the Paris and Ile-de-France region shows that hysterectomy for benign disorders is mainly performed by the vaginal route. Vaginal hysterectomy is quicker than laparoscopic or laparotomic hysterectomy. However, the decision of surgical route depends on patient characteristics and uterus weight. Complications are rare and independent of surgical route. Hysterectomy is the second most common surgical procedure employed worldwide (Wu et al., 2005). Despite >30 prospective comparative trials, the choice of route still depends on the surgeon s experience and on local preferences (Garry et al., 2004b). The most striking findings in our study are the high frequency of vaginal hysterectomy (48.3% of cases) and the low frequency of laparotomic hysterectomy (24.4%). When we excluded a centre in which hysterectomy was almost always performed by the laparoscopic route, the rate of vaginal hysterectomy rose to 56.5%. In previous studies, the rates of vaginal and laparotomic hysterectomies ranged from 10.9 to 52.4% and from 42.4 to 83.8%, respectively (Vessey et al., 1992; Harris and Olive, 1994; Dorsey et al., 1995; Johns et al., 1995; Davies et al., 1998; Härkki-Siren et al., 1998). In a previous French multicentre study (Chapron et al., 1999), the rates of laparoscopic, vaginal and laparotomic hysterectomies were 9.6, 47 and 43.4%, respectively. Compared with the previous study (Chapron et al., 1999) that had the same design, we show that the decrease of laparotomic hysterectomy was related to the increase of laparoscopic approach. These findings suggest that it is possible to reduce the use of laparotomy for hysterectomy with adequate training in both vaginal and laparoscopy surgeries. All the participating centres in our study were university hospitals; this may explain that the rates of laparoscopic hysterectomy and LAVH were higher in our study than those in the previous reports (Vessey et al., 1992; Harris and Olive, 1994; Dorsey et al., 1995; Johns et al., 1995; Davies et al., 1998; Härkki-Siren et al., 1998), in which laparoscopy was used in <10% of cases. Corroborating our results, Mabille de Poncheville (1998) showed that laparoscopic hysterectomy was significantly more frequent in university hospitals than in general hospitals and private clinics. We report in this study that the choice of surgical route depends on four factors: personal history of patients (parity and history of pelvic surgery), indication, uterus weight and treatment centre. Vaginal route was used mainly in post-menopausal women, women with high parity and women with no history of Caesarean section or pelvic surgery. Previous studies have shown the feasibility of vaginal hysterectomy in nulliparous women (Chauveaud et al., 2002). However, the use of vaginal route is more difficult in nulliparous patients because of the absence of prolapse. We also demonstrated in this study that hysterectomy was performed by vaginal route for the cure of a prolapse in >25% of patients. We found that uterine weight was significantly higher in the laparotomic group. Although hysterectomy is feasible by the laparoscopic or vaginal route in women whose uterus weighs >280 g, with acceptable complication rates (Daraï et al., 2001; Soriano et al., 2001), we confirmed that uterine weight is the main limitation on vaginal hysterectomy. 264 We report in our study that vaginal hysterectomy is quicker than using other approaches (laparotomy or laparoscopy). These data are in keeping with a recent meta-analysis (Johnson et al., 2005b) showing that the operating time is always shorter with vaginal hysterectomy than with laparotomy (Hwang et al., 2002; Ribeiro et al., 2003), LAVH (Moller et al., 1999) or laparoscopy (Daraï et al., 2001; Garry et al., 2004b). We found no difference in intra-operative complications across the different routes. In contrast, a Cochrane review (Johnson et al., 2005a) showed that, at least in some of the studies analysed, laparoscopic hysterectomy was associated with a higher incidence of urinary tract injury (bladder or ureter) than was abdominal hysterectomy, whereas other authors and we have reported that laparoscopic hysterectomy presents a higher risk of urinary tract injury than other routes (Daraï et al., 2001; Garry et al., 2004a). The rate of post-operative complications in our laparotomic group was 13.5%, but only a trend towards a difference among the groups was observed. This apparent discrepancy could be explained by the relatively low rate of complications and the sample size of our study population. Nevertheless, our data support the view that, when possible, the vaginal route has significant advantages over LAVH and laparoscopic hysterectomy and should be the route of first choice. There are few published data on determinants of complications in women undergoing hysterectomy. Whatever the route, we found that obesity, prior Caesarean section and prior pelvic surgery were the main determinants of intra- and post-operative complications. In contrast, age and menopausal status did not affect the rate of complications. However, we did not find any interaction between surgical route and risk factors for complication. Therefore, we cannot make any recommendation for the choice of surgical route according to obesity, history of pelvic surgery or history of Caesarean section. Moreover, we have to acknowledge that determining complications with the different approaches might be biased because our study is an observational study and not a randomized trial. As previously reported (Johnson et al., 2005a), we found that the hospital stay was longer after laparotomy than with the other routes. In contrast to Johnson et al. s results, the hospital stay was shorter after laparoscopic hysterectomy than after vaginal hysterectomy. The higher mean age of our patients, and the high frequency of associated procedures related to genital prolapse in the vaginal hysterectomy group, could explain this discrepancy. Women must be warned of the risk of conversion to laparotomy, as this prolongs the hospital stay and the recovery time. The rate of laparoconversion in our study was 7.5% overall and was significantly higher with LAVH and laparoscopic hysterectomy (19%) than with vaginal hysterectomy (4.2%). Moreover, laparoscopic hysterectomy was associated with a higher rate of laparoconversion than was LAVH. Regarding vaginal hysterectomy, our rate of conversion is in keeping with that found in previous prospective studies ( %) (National Center for Disease Control and Prevention, 1997; Daraï et al., 2001; Soriano et al., 2001). In contrast, the conversion rates during LAVH and laparoscopic hysterectomy were higher than

6 Different gonadotrophin preparations in intrauterine insemination cycles those in previous prospective studies (Vessey et al., 1992; Harris and Olive, 1994; Dorsey et al., 1995; Johns et al., 1995; Davies et al., 1998; Härkki-Siren et al., 1998). This discrepancy may be related to differences in the surgeons experience with laparoscopy surgery. Indeed, <10 LAVH or laparoscopic hysterectomy procedures were performed in 10 of the 12 participating centres during the study period. In keeping with previous studies (Vessey et al., 1992; Harris and Olive, 1994; Dorsey et al., 1995; Johns et al., 1995; Davies et al., 1998; Härkki-Siren et al., 1998), we found that most conversions were related to uterine volume and haemorrhage. Although both laparoscopic and vaginal hysterectomies are feasible and safe in women with an enlarged uterus (Chapron et al., 1999; Daraï et al., 2001), the risk of conversion was increased by a uterine weight of >500 g in our patients. Prior Caesarean section and prior pelvic surgery were risk factors for both laparoconversion and the incidence of complications. Recently, using multivariate analysis, Leonard et al. (2005) demonstrated that risk factors for conversion during laparoscopic hysterectomy were uterus width, BMI and a history of pelvic surgery. Despite advances in laparoscopy and vaginal surgery, these data confirm that laparotomy is still indicated in specific circumstances. In conclusion, this prospective observational study confirms that the vaginal route is increasingly used for hysterectomy in France and that it is the first-choice procedure in this setting. Laparoscopy may help reduce the incidence of laparotomic hysterectomy, especially in women with a history of laparotomy. Further studies taking into account the characteristics of both the patients and the uterus (e.g. volume) are required to clarify the respective places of the laparotomic, laparoscopic and vaginal routes for hysterectomy. Moreover, what is striking when analysing the data is that all routes are used in all university centres participating in this study, even if route proportion changes from one centre to another. This shows that there is a place for each route, even if the consensus has yet to be reached. References Chapron C, Laforest L, Ansquer Y, Fauconnier A, Fernandez B, Breart G and Dubuisson JB (1999) Hysterectomy techniques used for benign disorders: results of a French multicentre study. Hum Reprod 14, Chauveaud A, de Tayrac R, Gervaise A, Anquetil C and Fernandez H (2002) Total hysterectomy for a nonprolapsed, benign uterus in women without vaginal deliveries. J Reprod Med 47,4 8. Cosson M, Querleu D and Crepin G (1997) Hystérectomies pour pathologies bénignes. In Masson. Williams et Wilkins, Paris, p Daraï E, Soriano D, Kimata P, Laplace C and Lecuru F (2001) Vaginal hysterectomy for enlarged uteri, with or without laparoscopy assistance: randomized study. Obstet Gynecol 97, Davies A, Vizza E, Bournas N, O Connor H and Magos A (1998) How to increase the proportion of hysterectomies performed vaginally. Am J Obstet Gynecol 179, Dorsey JH, Steinberg EP and Holtz PM (1995) Clinical indications for hysterectomy route: patient characteristics or physician preference? Am J Obstet Gynecol 173, Garry R, Fountain J, Brown J, Manca A, Mason S, Sculpher M, Napp V, Bridgman S, Gray J and Lilford R (2004a) Evaluate hysterectomy trial. A multicentre randomised trial comparing abdominal, vaginal and laparoscopy methods of hysterectomy. Health Technol Assess 8, Garry R, Fountain J, Mason S, Hawe J, Napp V, Abbott J, Clayton R, Phillips G, Whittaker M, Lilford R et al. (2004b) The evaluate study: two parallel randomised trials, one comparing laparoscopy with abdominal hysterectomy, the other comparing laparoscopy with vaginal hysterectomy. BMJ 328, 129. Erratum in BMJ (2004) 328,494. Härkki-Siren P, Sjoberg J and Tiitinen A (1998) Urinary tract injuries after hysterectomy. Obstet Gynecol 92, Harris MB and Olive DL (1994) Changing hysterectomy patterns after introduction of laparoscopycally assisted vaginal hysterectomy. Am J Obstet Gynecol 171, Hwang JL, Seow KM, Tsai YL, Huang LW, Hsieh BC and Lee C (2002) Comparative study of vaginal, laparoscopically assisted vaginal and abdominal hysterectomies for uterine myoma larger than 6 cm in diameter or uterus weighing at least 450 g: a prospective randomized study. Acta Obstet Gynecol Scand 81, Johns DA, Carrera B, Jones J, DeLeon F, Vincent R and Safely C (1995) The medical and economic impact of laparoscopically assisted vaginal hysterectomy in a large, metropolitan, not-for-profit hospital. Am J Obstet Gynecol 172, Johnson N, Barlow D, Lethaby A, Tavender E, Curr E and Garry R (2005a) Surgical approach to hysterectomy for benign gynaecological disease. Cochran Database Syst Rev (2): CD Johnson N, Barlow D, Lethaby A, Taverder E, Curr L and Garry R (2005b) Methods of hysterectomy: systematic review and meta-analysis of randomised controlled trials. BMJ 330,1478. Leonard F, Chopin N, Borghese B, Fotso A, Foulot H, Coste J, Mignon A and Chapron C (2005) Total laparoscopy hysterectomy: preoperative risk factor for conversion to laparotomy. J Minim Invasive Gynecol 12, National Centre for Disease Control and Prevention (1997) Hysterectomy surveillance United States CDC surveillance summaries, August. Mabille de Poncheville L (1998) Cœliochirurgie gynécologique en France, instantanée Résultats d une enquête nationale. Thèse de médecine, Tours, France. Moller C, Kehlet H and Ottesen BS (1999) Hospitalization and convalescence after hysterectomy. Open or laparoscopy surgery? Ugeskr Laeger 161, Ribeiro SC, Ribeiro RM, Santos NC and Pinotti JA (2003) A randomized study of total abdominal, vaginal and laparoscopy hysterectomy. Int J Gynaecol Obstet 83, Soriano D, Goldstein A, Lecuru F and Daraï E (2001) Recovery from vaginal hysterectomy compared with laparoscopy-assisted vaginal hysterectomy: a prospective, randomized, multicenter study. Acta Obstet Gynecol Scand 80, University of York (UK) (1991) Centre for Health Economics. The management of menorrhagia. Effective healthcare 1(9). Vessey MP, Villard-Mackintosh L, McPherson K, Coulter A and Yeates D (1992) The epidemiology of hysterectomy: findings in a large cohort study. Br J Obstet Gynaecol 99, Wu SM, Chao Yu YM, Yang CF and Che HL (2005) Decision-making tree for women considering hysterectomy. J Adv Nurs 51, Submitted on February 7, 2006; resubmitted on May 5, 2006, July 19, 2006; accepted on July 24,