Correspondence: Professor G. Pardi, Department of Obstetrics and Gynaecology, San Paolo Biomedical Sciences Institute, Via A. Di Rudinì, 8, 20142, Milan, Italy.
Objective To compare laparoscopically assisted vaginal hysterectomy (LAVH) and total abdominal hysterectomy (TAH) in patients with uterine fibroids.
Design A prospective randomised study.
Setting The San Paolo Hospital, Milan.
Population Sixty-two patients, who were not suitable for a vaginal hysterectomy, requiring treatment for uterine fibroids.
Methods Randomisation between LAVH and TAH. Comparison of outcomes on the whole series, patients with uteri ≤ 500 g (Group 1) and patients with uteri > 500 g (Group 2).
Main outcome measures To establish operating time, blood loss, complications, febrile morbidity, analgesics administration and hospital stay for both treatment approaches.
Results Median uterine weight was 400 g in both LAVH and TAH group. Median operating time was longer for LAVH (135 min compared with 120 min for TAH; P= 0.001), but patients undergoing LAVH had less analgesics administration (23% compared with 77%, P < 0.001) and a shorter median hospital stay (3.8 compared with 5.8 days; P < 0.001). LAVH, when compared with TAH in the two weight subgroups, required a significantly longer operating time only in Group 2, significantly reduced analgesics administration only in Group 1, and significantly reduced hospital stay in both groups. Conversions of LAVH to laparotomy were significantly more frequent in Group 2 (3/11) than in Group 1 (0/20) (P= 0.04).
Conclusions Compared with TAH, LAVH has advantages in removing uteri weighing ≤ 500 g, with comparable operating time, less post-operative pain and shorter recovery. Among uteri weighing > 500 g LAVH showed a shorter recovery, but longer operating time than TAH and a 27% rate of conversion to laparotomy.
Despite vaginal hysterectomy being associated with less surgical trauma, less post-operative complications and faster recovery than abdominal hysterectomy1,2, only one-fourth of hysterectomies for benign pathologies are carried out vaginally3. The number of vaginal hysterectomies might be increased considering that: 1. traditional contraindications, such as the presence of a large uterus or adnexal masses, the need for salpingo-oophorectomy and previous pelvic surgery, represent no more than relative contraindications in the hands of well trained vaginal surgeons4,5; and 2. when absolute contraindications to vaginal hysterectomy are present, laparoscopic assistance may allow to convert an abdominal to a vaginal hysterectomy.
Laparoscopically assisted vaginal hysterectomy (LAVH) has shown no advantages over vaginal hysterectomy6, and it should be performed in total abdominal hysterectomy (TAH) candidates only. The best way to assess the possible advantages of LAVH over TAH is through prospective, randomised studies that include patients with a single indication for hysterectomy. Most published studies comparing LAVH and TAH were not randomised, while no studies have considered patients with a single indication for hysterectomy.
Uterine fibroids represent the most frequent indication for hysterectomy3, but the rate of vaginal hysterectomies in these patients has been reported to be as low as 4%7. It is therefore important to assess the role of LAVH as an alternative to TAH among women with uterine fibroids. In this prospective, randomised study, we sought to determine whether, and to what extent in terms of uterine size, LAVH could safely and effectively replace TAH in women due to have an abdominal hysterectomy for the single indication of uterine fibroids. The aim of this study was to compare operating time, blood loss, complications, febrile morbidity, post-operative analgesics requirement and hospital stay between LAVH and TAH on 1. the whole series of patients with uterine fibroids; 2. patients with uteri weighing ≤ 500 g; and 3. patients with uteri weighing > 500 g.
Over a 24 months period, 108 consecutive patients underwent hysterectomy for symptomatic uterine fibroids at the San Paolo Biomedical Sciences Institute, University of Milan, Italy.
Pre-operatively, a careful history was collected from each woman in order to identify previous pelvic pathologies. Severe pelvic disease was defined as an history of tubo-ovarian abscess, moderate or severe endometriosis or three or more previous pelvic surgeries. Subsequently, a clinical examination was carried out to assess uterine accessibility and mobility. Lack of uterine accessibility was defined as a narrow vaginal introitus or narrow vagina, while lack of uterine mobility was defined as the impossibility to pull the uterus down for more than two centimetres by a gentle traction on the cervix with a Martin's forceps. Finally, a sono-graphic examination was performed to assess uterine size and to evaluate size and location of fibroids and concomitant adnexal pathology. Sonographically estimated uterine volume was obtained according to the formula reported by Gohari et al8:
where A was equal to the uterine length (largest sagittal distance between the level of the internal os and the top of the uterine fundus), B to the uterine thickness (the greatest width between the anterior and the posterior uterine walls) and C to the uterine width (largest diameter between the lateral sidewalls). Adnexal masses were evaluated according to a morphological score previously described9.
Women without a history of severe pelvic disease, with an accessible and mobile uterus and a sonographically estimated uterine volume ≤ 500 mL, underwent vaginal hysterectomy (n = 21). Median sonographically estimated uterine volume in these patients was 190 mL (interquartile range 104–295 mL), while median actual uterine weight was 230 g (interquartile range 125–334 g). Six of these patients had previously undergone pelvic surgery.
Women with either a history of severe pelvic disease, a lack of uterine accessibility and mobility or a sonographically estimated uterine volume > 1500 mL were directly scheduled for an abdominal hysterectomy (n= 22). Median sonographically estimated uterine volume in these patients was 653 g (interquartile range 480–896 g), while median actual uterine weight was 780 g (interquartile range 532–1115 g).
The remaining 65 women were considered eligible for random distribution between laparoscopically assisted vaginal hysterectomy and abdominal hysterectomy. Each woman was asked to sign an informed consent before randomisation. Three women refused to sign the consent, were scheduled to have abdominal hysterectomy, and were excluded from further analysis. Randomisation of the remaining 62 patients was done by opening sealed opaque envelopes containing computer-generated randomisation numbers.
Total abdominal hysterectomies with or without salpingo-oophorectomy were performed according to standard technique. The laparoscopic part of laparoscopically assisted vaginal hysterectomies included the visualisation of the pelvis and upper abdomen, the treatment of adhesions or endometriosis when present, and the completion of the upper part of the hysterectomy. Round ligaments, tubes and utero-ovarian ligaments were desiccated and transected when the adnexa were to be preserved, while the round and infundibulopelvic ligaments were desiccated and transected when the adnexa were to be removed. In all cases the broad ligaments were dissected to their lower margin. When the bladder was stretched over the anterior aspect of the uterus due to previous surgery, the bladder flap was developed laparoscopically. The vaginal part of the hysterectomy included colpoceliotomy and bilateral ligation and transection of utero-sacral ligaments, uterine vessels and cardinal ligaments; cervical amputation, corporal emisection, myomectomy and uterine morcellation were performed when necessary (in 25, 23 and 7 cases, respectively).
Data were collected for each patient with regard to uterine weight (as measured within 30 min after hysterectomy), total operating time (from the first incision to the closure of all wounds), pre-operative haemoglobin concentration, haemoglobin concentration on the second postoperative day, complications and hospital stay. Post-operatively, temperature and analgesic requirement were recorded daily. Fever was defined as a temperature greater than 38°C on two consecutive measurements 24 hours after surgery. Patients were discharged when they were apyretic and deambulatory, had no functional impairment of the bladder and did not need analgesics.
Results were evaluated by linear regression analysis, Fisher's exact test and Student's t test for independent samples. Continuous outcome measures, non-normally distributed according to the Skewness/Kurtosis test (sktest) for normality10, were transformed into normally distributed data prior to perform Student's t test. For outcome measures that were significantly different between LAVH and TAH only in one of the two weight Groups, a test of interaction was performed. A P value < 0.05 was considered significant.
The randomisation process assigned 31 women to the LAVH group and 31 to the TAH group. The two groups were comparable according to median age, sonographically estimated uterine volume, nulliparity, prevalence of adnexal masses and previous pelvic surgery (Table 1). In three women (10%) in the LAVH group, the procedure was converted to a TAH. In all cases the decision was made during the laparoscopic part of the procedure. The first patient, during broad ligament dissection, developed heavy bleeding from a uterine vein that could not be controlled laparoscopically; in the other two patients, the round and infundibulo-pelvic ligaments were inaccessible by laparoscopy due to adhesions and large fibroids. Uterine weights in these patients were 580 g, 600 g and 700 g, respectively, mean hospital stay was six days and no complications were observed. For the purpose of data analysis, these patients remained in the LAVH group.
Table 1. Pre-operative patients' characteristics. Values are given as median (IQR) or n [%]. LAVH = laparoscopically assisted vaginal hysterectomy; TAH = total abdominal hysterectomy
LAVH (n= 31)
TAH (n= 31)
Sonographic uterine volume (mL)
Previous pelvic surgery
Figure 1 shows a comparison of sonographically estimated uterine volumes with actual uterine weights: the correlation is highly significant (R2= 0.92, P < 0.001). The equation for this relationship is:
As shown in Table 2, LAVH took significantly longer than TAH, was associated with a significantly lower administration of analgesics after the first 24 post-operative hours and was associated with a significantly shorter hospital stay. Uterine weight, post-operative haemoglobin decrease, blood transfusions and postoperative febrile morbidity were comparable in the two groups. No major complications occurred in both groups. Bilateral salpingo-oophorectomy was performed in 19 women (61%) in the LAVH group and in 21 women (68%) in the TAH group. All patients had pathologically confirmed uterine leiomyomas; the eight adnexal masses diagnosed as benign at preoperative sonography were six functional ovarian cysts and two endometrioma.
Table 2. Intra- and post-operative results. Values are given as median (IQR) or n[%]. LAVH = laparoscopically assisted vaginal hysterectomy; TAH = total abdominal hysterectomy
LAVH (n= 31)
*Patients who required analgesia for more than 24 hours after surgery.
Operating time (min)
Uterine weight (g)
Haemoglobin decrease (g/dL.)
Hospital stay (days)
Forty-one women (66%) had uteri weighing ≤ 500 g (Group 1), while 21 (34%) had uteri weighing > 500 g (Group 2). Table 3 compares LAVH and TAH in these two groups. LAVH operating time was significantly longer than TAH operating time only in patients of Group 2; analgesics administration was significantly reduced for LAVH only in patients of Group 1; and hospital stay was significantly reduced for LAVH in both groups. Test of interaction was not significant for both operating time (P= 0.12) and analgesics administration (P= 0.06).
Table 3. Intra- and post-operative results. Values are given as median (IQR) or n [%]. Group 1 = uterine weight <500 g; Group 2 = uterine weight ≥ 500 g; LAVH = laparoscopically assisted vaginal hysterectomy; TAH = total abdominal hysterectomy
LAVH (n= 20)
TAH (n= 21)
LAVH (n= 11)
TAH (n= 10)
*Patients who required analgesia for more than 24 hours after surgery
Operating time (min)
Uterine weight (9)
Haemoglobin decrease (g/dL)
Hospital stay (days)
Had the three procedures that were converted from a LAVH to a TAH been excluded from intra- and post-operative data analysis, no significant differences in the data reported in Table 2 and 3 would have been observed.
Among patients undergoing LAVH, conversions to laparotomy were significantly more frequent in Group 2 (three conversions out of 11 patients versus no conversions out of 20 patients; P= 0.04); laparoscopic operating time was comparable between Group 1 and Group 2 (Group 1: median 48 minutes, interquartile range 39–56 minutes; Group 2: median 50 minutes, interquartile range 44–78 minutes; P= 0.5); mean vaginal operating time was significantly longer in Group 2 (median 120 minutes, interquartile range 98–131 minutes, excluding the three conversions to TAH) when compared with Group 1 (median 80 minutes, interquartile range 60–104 minutes) (P= 0.04).
Previous published studies designed to assess the role of LAVH did not identify the proper indications for this procedure11–15, mainly because heterogeneous indications for hysterectomy were reported. Our study was designed according to the principles: 1. LAVH must be carried out in abdominal hysterectomy candidates only; 2. the role of LAVH as an alternative to TAH can be assessed only by prospective randomised studies in patients with a single indication for hysterectomy. We selected uterine fibroids because of the high prevalence of this indication for hysterectomy with a low rate of direct vaginal procedures. We therefore report the results of a study in which patients considered not suitable for a vaginal hysterectomy, all affected by uterine fibroids, were randomly assigned either to LAVH or TAH.
Pre-operative assessment of uterine size is of critical importance in the choice of the route for hysterectomy in patients with uterine fibroids. The advantages of sonographic estimation of uterine volume over bimanual examination are still controversial16. However, in the present series, pre-operative ultrasound was very accurate in the prediction of uterine weight with a correlation coefficient of 0.92. Furthermore, it allowed us to assess not only the size, but also the shape of the uterus and the location of the fibroids, thus potentially being useful in planning the hysterectomy. Finally, it allowed the evaluation of the concomitant adnexal masses9. We believe that a pre-operative ultrasound should be conducted routinely in patients undergoing hysterectomy for uterine fibroids.
In our department we consider uteri up to a sonographically estimated volume of 500 cc, in absence of contraindications to a vaginal hysterectomy (such as lack of uterine accessibility and mobility, or severe pelvic disease) suitable for a direct vaginal hysterectomy. We consider candidates for a direct abdominal hysterectomy patients with either a sonographically estimated uterine volume of > 1500 cc or lack of uterine accessibility and mobility or severe pelvic disease. As reported in the present study, these criteria allowed us to perform a direct vaginal hysterectomy in 20% of women with uterine fibroids and to consider another 60% of women eligible for randomisation between LAVH and TAH, while only 20% of women were scheduled for a direct abdominal hysterectomy.
The overall results of our study confirm, among patients with uterine fibroids only, what was already known from previous studies11–15. Laparoscopically assisted vaginal hysterectomy allows many abdominal hysterectomies to be converted to the vaginal route with the disadvantage of longer operating time, but the advantages of shorter hospital stay and less post-operative pain. Complications were minimal in both groups, and they were not significantly different.
Median operating time in our series was 135 minutes for LAVH and 120 minutes for TAH, longer than that reported by previous studies12–17. This may be due to the homogeneous indication of uterine fibroids, which often led us to deal with very enlarged uteri. Another contributing factor may be that, in order to assess the possible benefits of the introduction of LAVH in general practice rather than in referring hospitals, all procedures were performed by different surgeons, including residents, with a variable experience in gynaecological surgery. Finally, we chose to avoid the use of laparo-scopic staplers because, in agreement with other authors18, we believe that the additional costs and the increased risk of incisional hernia from a 10 mm port, related to the use of these devices, overweight the benefit of a shorter operating time.
The lack of a long term measure of recovery is a limitation of the present study. However, we believe that the peculiar characteristic of our study, the presence of very enlarged uteri, is more likely to influence intra-operative and short term outcome variables than long term recovery. Previous papers that have compared long term recovery between LAVH and TAH are in agreement with each other and indicate that convalescence time is shorter for LAVH12–15,17,19.
Three LAVH had to be converted to a TAH. In one case, early in our experience, bleeding from a uterine vein could not be controlled by bipolar coagulation. In the other two cases, as previously experienced by other authors15,20, we found the upper pedicles inaccessible by laparoscopy. This was due to a large subserous lateral fibroid in one case and to thick adhesions between omentum, bowel loops and a large uterine fundal fibroid in the other case. Both patients had uteri weighing > 500 g and a severe distortion of the pelvic anatomy; in both cases, conversion to laparotomy was decided immediately after diagnostic laparoscopy, thus not increasing operating time. In our experience, conversion to laparotomy was never necessary during the vaginal part of LAVH, confirming the effectiveness of the techniques for reducing uterine volume during vaginal hysterectomy.
In women with uteri weighing ≤ 500 g, LAVH had a longer mean operating time than TAH, but the difference was not statistically significant. This is in disagreement with previous studies14 that showed a significantly longer operating time for LAVH than for TAH. An explanation may be that, in our series, we converted every LAVH to a vaginal procedure as soon as we felt feasible and safe to do so. In fact, to perform more steps of the hysterectomy laparoscopically instead of vaginally results in longer operating times4. In this group of patients with uteri weighing ≤ 500 g, LAVH was associated with less analgesics administration and shorter hospital stay; this is consistent with previous reports14,18. Therefore, in these patients, which represent two-thirds of randomised patients, laparoscopic assistance made possible to perform the hysterectomy vaginally instead of abdominally, with comparable operating time and the advantages of less surgical trauma and faster recovery.
This is the first randomised study that compares LAVH and TAH in patients with uteri weighing > 500 g. As expected, in these patients, we observed longer operating time for LAVH than for TAH. This was related to the significantly longer vaginal part of LAVH in these patients than in patients with uteri weighing ≤ 500 g, while laparoscopic part of LAVH and TAH times were comparable in the two groups. We did not observe any difference between LAVH and TAH with regard to analgesics administration: this may suggest that surgical trauma to pelvic tissues for LAVH was greater in these patients than in patients with uteri weighing ≤ 500 g, either for the length of surgery or the manoeuvres applied for removing large uteri. However, the benefit of a reduced hospital stay for LAVH, observed for patients with uteri weighing ≤ 500 g, was maintained also in this group. In our experience, therefore, LAVH showed a longer operating time, a comparable post-operative pain and a shorter hospital stay when compared with TAH in women with uteri weighing > 500 g.
The need for a conversion to laparotomy was significantly higher in patients with uteri weighing > 500 g (3 out of 11 patients: 27%) when compared with patients with uteri weighing ≤ 500 g (0 out of 20 patients). In comparison Pelosi and Kadar21 had reported the need for conversion to laparotomy for uteri weighing ≥ 500 g in two out of 22 (9%) patients. However, in the present series, the severity of the distortion of the pelvic anatomy rather than the uterine size itself was the main reason for conversion to laparotomy. As a demonstration, when the uterus was symmetrically enlarged and there was a good accessibility to the upper pedicles, we were able to remove by LAVH uteri ≥ 1300 g. From this standpoint, pre-operative sonographic assessment of uterine shape may be of great importance in the pre-operative prediction of the feasibility of a LAVH.
In the present study, we found LAVH safe and effective and advantageous over TAH in removing uteri weighing ≤ 500 g, thus potentially allowing us to increase our rate of vaginal hysterectomy for symptomatic uterine fibroids from 20% to 59%. More importantly, we believe that the use of the simple criteria described in the present paper, based mostly on an accurate selection of patients and on the use of laparoscopy only for the resection of the upper pedicles, may allow most gynaecological units to reproduce these figures. On the contrary, LAVH showed some advantages, but also important limitations when compared with TAH in patients with uteri weighing > 500 g. Moreover, our experience in this group of patients cannot be considered definitive since it is based on a small population. Therefore, we believe that the role of LAVH in patients with very large uteri should still be objective of investigation.
In conclusion, LAVH is advantageous over TAH in removing uteri weighing ≤ 500 g, with comparable operating time, less post-operative pain and shorter recovery. Among uteri weighing > 500 g LAVH showed a shorter recovery, but longer operating time than TAH and a 27% rate of conversion to laparotomy; other prospective studies are needed to assess the relative role of LAVH and TAH in patients with very large uterine fibroids.
The authors would like to thank Professor A. Morabito, from the Institute of Medical Statistics and Biometrics of the University of Milan, for his careful statistical advice.