You are here

Laparoscopic-assisted vaginal hysterectomy vs abdominal hysterectomy for benign disease: a meta-analysis of randomized controlled trials

European Journal of Obstetrics & Gynecology and Reproductive Biology, 1, 159, pages 1 - 18

Abstract

The objective of this meta-analysis was to assess whether laparoscopic-assisted vaginal hysterectomy achieves better clinical results compared with abdominal hysterectomy. Medline (PubMed), EMBASE, Web of Science, ProQuest, Cochrane Library and China Biological Medicine Database were searched to identify randomized controlled trials that compared laparoscopic-assisted vaginal hysterectomy with abdominal hysterectomy. Twenty-three trials were studied and the analysis was performed using Review Manager Version 5 and R Version 2.11.1. The results showed that laparoscopic-assisted vaginal hysterectomy was associated with a longer operation time, less blood loss, shorter hospital stay, smaller haemoglobin drop, less postoperative pain, quicker return to normal activities and fewer peri-operative complications. Quality of life is likely to be the key outcome to evaluate the approach for hysterectomy, but further research is needed. For suitable patients and surgeons, laparoscopic-assisted vaginal hysterectomy is a better choice than abdominal hysterectomy.

Keywords: Meta-analysis, Laparoscopic-assisted vaginal hysterectomy, Abdominal hysterectomy, Approach.

1. Introduction

Hysterectomy is one of the most common operations for benign disease [1] . In 2003, 538,722 patients underwent hysterectomy for benign disease [2] . Currently, although many other methods are available to treat benign disease, hysterectomy remains the preferred technique and its ratio also remained stable [2] .

There are several types of hysterectomy: abdominal hysterectomy (AH), supracervical hysterectomy, vaginal hysterectomy, laparoscopic-assisted vaginal hysterectomy (LAVH), total laparoscopic hysterectomy and robotic hysterectomy. Since LAVH was first reported in 1989, it has gained worldwide acceptance [3] . Johnson et al. [4] found that LAVH decreased pain, surgical site infections and hospital stay, and led to a quicker return to normal activities and fewer postoperative adhesions. However, Sculpher et al. [5] found that the efficiency of the laparoscopic route was less than that of the vaginal route, but equal to the abdominal route. Kluivers et al. [6] reported that the quality of life following LAVH was better than that following AH at 6 weeks after surgery, but the difference was not significant after 1 year.

Although previous systematic reviews [4] and [7] have referred to comparisons of LAVH and AH, new research may support or refute the findings. Moreover, to date, no systematic reviews have evaluated these two methods directly. As LAVH and AH are common procedures, further evaluation and comparison is needed. As such, this study aimed to assess which method resulted in the best clinical results, and to provide evidence for decision making regarding the surgical treatment of benign disease.

2. Methods

2.1. Data sources and searches

Electronic databases and hand searching were performed from 20 February 2010 to 2 July 2010 to identify all relevant studies. The electronic databases searched were Medline (PubMed), EMBASE, Web of Science, ProQuest, Cochrane Library and China Biological Medicine Database. The keywords were ‘laparoscopic’, ‘laparoscopically’, ‘vaginal’, ‘assisted’, ‘hysterectomy’, ‘randomized’, ‘randomised’ and ‘randomly’. The ‘related’ function was used to broaden the search. No language restrictions were made. The search strings used in Medline (PubMed) and EMBASE are presented in Table 1 . Other electronic databases were searched similarly.

Table 1 Search strings in Medline (PubMed) and EMBASE.

1. Medline (PubMed)
 #1 laparoscopic[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomized[Title/Abstract]
 #2 laparoscopic[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomised[Title/Abstract]
 #3 laparoscopic[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomly[Title/Abstract]
 #4 laparoscopically[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomized[Title/Abstract]
 #5 laparoscopically[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomized[Title/Abstract]
 #6 laparoscopically[Title/Abstract] AND vaginal[Title/Abstract] AND assisted[Title/abstract] AND hysterectomy[Title/abstract] AND randomized[Title/Abstract]
2. EMBASE
 #1 (laparoscopic OR laparoscopically):ab,ti AND vaginal:ab,ti AND assisted:ab,ti AND hysterectomy:ab,ti AND (randomized OR ramdomised OR randomly):ab,ti

The manual search was performed by screening the references cited in the acquired articles. Published systematic reviews and meta-analyses on the topic [4], [7], [8], and [9] were also searched to identify eligible studies.

2.2. Study selection

Two reviewers independently reviewed the titles and abstracts of all identified potential studies to exclude duplicates. The titles, abstracts and full texts were then scrutinized to assess whether the studies met the eligibility criteria, and to exclude the unqualified studies. The studies were assessed using qualitative synthesis. Reviewers were not blinded to the authors or outcomes. Discussions were held in the case of disagreement between the two reviewers.

2.3. Eligibility criteria

Published randomized controlled trials that compared LAVH with AH for benign disease, clearly documented the case selection criteria for LAVH and AH, and reported on at least one of the outcomes mentioned below were included in the study. Studies were excluded if the laparoscopic technique was not LAVH, the outcomes were not clearly reported, it was impossible to extract data, there was considerable overlap between studies, the study was not a randomized controlled trial, review and retrospective articles and abstracts were not documented, and the case selection criteria were not unified.

2.4. Definitions of interventions and outcomes of interest

2.4.1. Definitions of interventions

In LAVH(a), the uterine vessels were ligated laparoscopically but part of the operation was performed vaginally. In LAVH(b), the laparoscopic component did not involve division of the uterine vessels [10] . AH was defined as removal of the uterus through an incision in the lower abdomen [7] .

2.4.2. Outcomes of interest and definitions

The operative outcomes were: (1) operation time (min), calculated from the first incision to the closure of all wounds; (2) blood loss (ml), estimated by calculating the blood volume of the suction machine during surgery excluding liquid used for intraperitoneal washing, and by weighing swabs; (3) hospital stay (days), defined as the number of days in hospital after surgery excluding the day of surgery; (4) postoperative pain, assessed using a 0–10-point visual analogue scale; (5) haemoglobin drop (g/100 ml), defined as the decrease in haemoglobin from pre-operative level to postoperative level; and (6) return to normal activities (days), defined as the number of days after discharge from hospital until full recovery.

Peri-operative complications were subdivided into major and minor complications [11] , and were defined as one or more of the following criteria adapted from a previous publication [9] —major complications: visceral damage (bladder, ureter or bowel), vaginal vault dehiscence and other life-threatening complications, such as thromboembolic disease; minor complications: pelvic haematoma, febrile morbidity (not specified), need for blood transfusion and other miscellaneous complications, including anaesthetic problems and vault granulation. Overall complications were defined as major plus minor complications.

2.5. Data extraction and quality assessment

Two reviewers extracted the data independently using Microsoft Excel (USA). The data were summarized beyond the analysis if the following attempts failed: (1) contacted corresponding authors via e-mail to access the original data or suitable results, and sought the relevant data in previous published systemic review and meta-analysis articles [4], [7], [8], and [9] when the continuous data were not expressed as means and standard deviations; and (2) the units of outcomes were impossible to convert in a mathematical way when they were not expressed in the predefined form. If conversions from LAVH to AH were reported without specifically reporting that the data were analysed in the AH group, the data were considered to have remained in the LAVH group. A discussion was held if the data of the two reviewers differed, and the final data were agreed by both reviewers.

The quality of individual studies was assessed using the Cochrane methods for assessing the risk of bias [12] . Specifically, study quality was assessed by evaluating the methods of randomization, allocation concealment, blinding, whether incomplete outcome data were addressed, and if the study was free of selective reporting and other bias.

2.6. Statistical analysis

Analysis was performed using Review Manager Version 5 and R Version 2.11.1 [13] in accordance with the recommendations of Cochrane Handbook Version 5.0.2 [12] and the Preferred Reporting Items for Systemic Reviews and Meta-analysis (PRISMA) guidelines [14] and [15]. p ≤ 0.05 was taken to indicate statistical significance.

Heterogeneity was evaluated using χ2-test and I2 [12] . The random effect model was used when heterogeneity was significant; otherwise, the fixed effect model was used. If I2 > 50%, meta-regression analyses with random effects [16], [17], and [18] were performed (R, package meta) to detect the possible sources of heterogeneity according to the prespecified potential effect modifiers, including LAVH categorization, setting (single centre or multicentre), year of publication, power calculation (reported or not) and adequate sequence generation (reported or not). In addition, sensitivity analyses were conducted by alternating analytical models. To further assess the effect of individual studies on the summary estimate of effect, influence analyses were performed to recalculate the pooled estimate by omitting one study at a time (R, package meta). If the results indicated a factor to be the possible source of heterogeneity, subgroup analyses were conducted on the factor.

Continuous and dichotomous variables were analysed using mean difference (MD) with 95% confidence intervals (CI) and the inverse variance method, and odds ratio (OR) with 95% CI and the Mantel–Haenszel method, respectively. When there were no events in one group, this was solved by adding 0.5 to each cell of the 2 × 2 table [19] and [20]. If there were no events for both groups, the study was excluded.

Publication bias was evaluated by funnel plot, and the funnel plot's asymmetry was assessed using Egger's test [21] (R, package meta). p < 0.1 was considered to represent significant asymmetry.

3. Results

The search strategy identified 23 studies [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], and [44] ( Fig. 1 ). Analysis was conducted on 2051 patients (1023 LAVH, 1028 AH). The characteristics of the included studies, and descriptions of study quality and assessment are shown in Table 2, Table 3, and Table 4 and Fig. 2 .

gr1

Fig. 1 Search strategy for randomized controlled trials comparing laparoscopic-assisted vaginal hysterectomy (LAVH) and abdominal hysterectomy (AH).

Table 2 Characteristics of included studies.

Study LAVH categorization LAVH vs AH Inclusion criteria Exclusion criteria Outcomes
Olsson [40] LAVH(a) 71 vs 72 Scheduled for AH attributable to benign disorders, with a maximum uterine width <11 cm as measured by transvaginal ultrasound and not considered suitable for vaginal hysterectomy Vaginal hysterectomy could be performed (adnexa are not to be removed; there is no suspicion of endometriosis or postinflammatory disorders; uterine size is normal or, in the case of uterovaginal prolapse, less than the size of an 8-week pregnancy) Operating time (min); complications; postoperative pain relief (VAS); convalescence (days); hospital stay (days); QOL; economic analysis (cost)
Kunz [22] LAVH(b) 35 vs 35 Scheduled for hysterectomy for non-malignant disease Not stated Operating time (min); pain relief (VAS); uterine size (cm); haemoglobin change; stay in hospital (days); complications
Langebrekke [23] LAVH(a) 46 vs 54 With indications for hysterectomy Proven or suspected malignancies in the pelvic area, suspected intra-abdominal adhesions, uteri enlarged beyond the size of a 12-week pregnancy, serious cardiopulmonary disease or previous colporraphy Operation time (min); hospital stay (days); time elapsed before resuming work (days); postoperative pain; complications
Ellstrom [44] LAVH(a) 20 vs 20 Scheduled for abdominal hysterectomy for benign disorders; maximum width of uterus, measured by transvaginal ultrasound, <11 cm; American Society of Anesthesiologists Grade 1 Not stated Postoperative pain (VAS); pulmonary function; time of anaesthesia (min); operation time (min); complications; difference in erythrocyte volume fraction
Summitt [24] LAVH(a) 34 vs 31 (1) Documented visual diagnosis of pelvic endometriosis Required a concomitant colporrhaphy, urethropexy, vaginal vault suspension or a non-gynaecological major operation; medical conditions requiring in-hospital monitoring (e.g., diabetes mellitus, organic heart disease) or had known cervical or endometrial cancer Operating time (min); estimated blood loss (ml); uterine weight (g); mean haematocrit; hospital stay (days); convalescence (days); hospital charges (US$)
    (2) Documented pelvic adhesions    
    (3) Three or more previous laparotomies (minilaparotomies were excluded)    
    (4) Uterine leiomyomata 12–18 gestational weeks in size    
    (5) Previous tubo-ovarian abscess or two documented episodes of pelvic inflammatory disease requiring intravenous antibiotic therapy    
    (6) Adnexal mass in the presence of an indication for hysterectomy, or the need for total ovariectomy    
    (7) Indicated hysterectomy with lack of mobility and unfavourable vaginal introitus    
Yuen [24] LAVH(a) 20 vs 24 With no major medical diseases requiring hysterectomy for benign disorders Suitable for vaginal hysterectomy or had uterine size >16 weeks of gestation Operation time (min); blood loss (ml); postoperative stay (days); postoperative complications
Falcone [25] LAVH(a) 24 vs 24 Scheduled for abdominal hysterectomy for benign disease Pelvic mass size >2 cm below the umbilicus Operative time (min); blood loss (ml); length of hospital stay (days); uterine weight (g); intra-operative complications; postoperative pain (VAS); return to work/normal activities (days); hospital costs (US$)
      Concomitant incontinence or pelvic reconstructive procedures required  
Marana [26] LAVH(b) 58 vs 58 One or more contra-indications to vaginal hysterectomy: uterine size >280 g and an upper limit of 16 weeks of gestation (700 g); previous pelvic surgery; history of pelvic inflammatory disease; moderate or severe endometriosis; concomitant adnexal mass or indication for adnexectomy; and nulliparity with lack of uterine descent and limited vaginal access Suitable for vaginal hysterectomy Pre-operative haemoglobin (g/100 ml); operating time (min); estimated blood loss (ml); postoperative day 1 haemoglobin drop (g/100 ml); postoperative stay (days); postoperative pain (VAS)
Harkki-Siren [41] LAVH(a) 25 vs 25 Aged between 30 and 70 years scheduled for AH for benign disease Major medical diseases, body mass index >32 kg/m2, uterine size >14 weeks of gestation or uterine width >10 cm by transvaginal ultrasonography, severe adhesions or endometriosis, prolapse and any other contra-indication for laparoscopy Operating time (min); anaesthetic time (min); blood loss (ml); haemoglobin change (g/100 ml); hospital stay (days); sick leave (days); complications
Ferrari [27] LAVH(b) 31 vs 31 With symptomatic uterine fibroids Without a history of severe pelvic disease and with an accessible and mobile uterus and a sonographically estimated uterine volume ≤500 ml; with either a history of severe pelvic disease, a lack of uterine accessibility and mobility or a sonographically estimated uterine volume >1500 ml; patients were not unwilling to receive randomized treatments Operating time (min); uterine weight (g); haemoglobin drop (g/dl); blood transfusions; febrile morbidity; postoperative analgesia; hospital stay (days)
Lumsden [28] Non-categorizable LAVH 95 vs 95 Not suitable for vaginal hysterectomy (uterine size >14 weeks of gestation or a requirement for oophorectomy) With hormone replacement therapy Length of operation (min); total length of stay (days); admission to intensive care unit; re-admissions; blood transfusions; women requiring additional surgery; complications
Ottosen [29] LAVH(b) 40 vs 40 Menorrhagia, leiomyomas <15 cm in diameter, dysplasia, endometrial atypia and pain With ovarian pathology, uterine size >16 weeks of gestation, previously known dense adhesions, narrow vagina or obvious inaccessible uterus Duration of surgery (min); duration of anaesthesia (min); stay in hospital (days); sick leave recovery (days); peri-operative blood loss (ml); complications
Lin [30] LAVH(b) 60 vs 60 With benign disease of myoma, adenomyosis, dysfunctional uterine bleeding, etc. Uterine size >12 weeks of gestation Not stated Uterine weight (g); operation time (g); blood loss (ml); haemoglobin drop (g/100 ml); hospital stay (days); postoperative pain (VAS)
Hwang [31] LAVH(a) 30 vs 30 With a myoma diameter >8 cm and second myoma <5 cm or two myomata, both at least 6 cm in diameter but <8 cm; total number of uterine fibroids could not be more than three Under the indications of adenomyosis, uterine prolapse, chronic pelvic pain, dysfunctional uterine bleeding, cervical dysplasia or pelvic inflammatory disease Operation time (min); blood loss (cc); hospital stay (days); postoperative tenderness scores after 24 h; return to work (days); days of antibiotics used; pathological ultrasound correlation (%); uterine weight (g); complications
Schutz [32] LAVH(a) 28 vs 20 A sonographically estimated uterine weight >200 g Not stated Duration of operation (min); blood loss (ml); blood transfusion; additional procedures; uterine weight (g); postoperative complications; median haemoglobin; median haematocrit; pain index; hospital stay (days); convalescence (days)
Seracchioli [43] LAVH(a) 60 vs 62 Presence of a very large uterus (>14 weeks of gestation) caused by myomas, which made women eligible for AH With uteri projecting above the transverse umbilical line as well as those with other pelvic pathologies, such as prolapse, pelvic floor relaxation, stress incontinence and adnexal masses; medical conditions requiring hospital monitoring (diabetes mellitus, heart disease), if they had undergone previous abdominal surgery requiring longitudinal laparotomy, or if they had absolute contra-indications to operative laparoscopy Operating time (min); laparoconversions; blood loss (ml); haemoglobin drop (g/100 ml); fever (no.); transfusions (no.); hospital stay (days); convalescence (days)
Tsai [33] LAVH(b) 100 vs 100 (1) Estimated uterine upper margin is not beyond the midpoint between umbilicus and pubic symphysis Not stated Parity; body mass index (kg/m2); blood loss (ml); total operating time (min); uterine weight (g); complications; time of hospitalization (days); number of meperidine ampoules
    (2) No pre-existing cardiopulmonary dysfunction or poor control of systemic diseases    
    (3) Bimanual pelvic examination confirmed good mobility of an enlarged uterus    
    (4) No cervical malignancy on colposcopy    
    (5) No indication for conventional vaginal hysterectomy    
Atabekoglu [34] LAVH(b) 23 vs 23 Scheduled for hysterectomy With any major contra-indication to surgery; with any medical condition considered high risk for laparoscopic surgery, with uterine size >14 weeks of gestation; with body mass index >32 kg/m2, considered suitable for vaginal hysterectomy, and with previous history of pelvic inflammatory disease and endometriosis Intra-operative bleeding (ml); previous abdominal operation (no.); duration of surgery (min); duration of hospitalization (days); analgesic requirement
Zhang [35] LAVH(a) 60 vs 60 With benign disease of myoma, adenomyosis, dysfunctional uterine bleeding, etc. and uterine size ≤16 weeks of gestation Uterine weight >280 g; with history of previous pelvic surgery, pelvic infection, moderate or severe endometriosis, adnexal mass with adnexectomy, nullipara, coleostenosis or pelvic adhesion Uterine weight (g); operation time (min); blood loss (ml); postoperative analgesia rate (%); hospital stay (days)
Muzii [36] LAVH(b) 40 vs 41 Had one or more of the generally considered contra-indications to vaginal approach for hysterectomy: uterine size >280 g calculated on pre-operative sonography, previous pelvic surgery (except for caesarean section), history of pelvic inflammatory disease or moderate or severe endometriosis, presence of adnexal disease, nulliparity with limited vaginal access and lack of uterine descent Uterine size >700 g calculated on pre-operative ultrasonography, previous midline laparotomy, and absolute contra-indication to laparoscopy Median operating time (min); conversion to laparotomy (%); postoperative haemoglobin drop (g/dl); postoperative VAS; median time to bowel function (h); median time to discharge (days)
Sesti [37] LAVH(b) 50 vs 50 Presence of symptomatic or rapidly growing myomas, age <55 years, and uterine size ≥12 weeks of gestation Nulliparous women, uterine size ≥16 weeks of gestation, previous uterine surgery, and suspicion of malignant gynaecological disease Operating time (min); blood loss (ml); conversion to laparotomy; intra-operative complications; paralytic ileus time (h); hospital discharge time (h); postoperative complications
Yue [38] LAVH(b) 42 vs 42 With uterine myoma disease With evidence of malignant gynaecological disease Operating time (min); anaesthesia time (min); blood loss volume (ml); postoperative stay (days); interleukin-6 (ng/l); tumour necrosis factor-alpha (ng/l); C-reactive protein (ng/l); T-lymphocyte subpopulations
Zhu [39] LAVH(a) 34 vs 32 With myoma disease, had indication for hysterectomy, uterine volume between 10 and 12 gestational weeks Not stated Operation time (min); blood loss (ml); operation fee (RMB); postoperative pain score (VAS); bowel recovery (h); high temperature (°C); morbidity (%); hospital stay (days)

LAVH, laparoscopic-assisted vaginal hysterectory; AH, abdominal hysterectomy; LAVH(a), laparoscopic hysterectomy where laparoscopic procedures include uterine artery ligation; LAVH(b), laparoscopic hysterectomy where the laparoscopic component did not involve ligation of the uterine vessels; QOL, quality of life; VAS, visual analogue scale.

Table 3 Quality of included studies.

Study Single centre or multicentre Power calculation (sample size) Method of randomization Method of allocation concealment Blinding Method of blinding Withdrawals and dropouts
Olsson [40] Single centre 140 patients required; primary outcome: complications (n = 143) By opening one of 150 sealed opaque envelopes Sealed opaque envelopes No Not stated No withdrawals or dropouts
Kunz [22] Single centre None (n = 70) Not stated Not stated No Not stated No withdrawals or dropouts
Langebrekke [23] Multicentre (two centres) None (n = 100) Random number table Sealed opaque envelopes No Not stated No withdrawals or dropouts
Ellstrom [44] Single centre None; primary outcomes: postoperative pain, pulmonary function (n = 40) Not stated Not stated No Not stated No withdrawals or dropouts
Summitt [24] Multicentre (three centres) None (n = 200) Computer generated Sealed opaque envelopes No Not stated Two women refused assigned procedure and were excluded from analysis
Yuen [42] Single centre None (n = 50) Computer generated Not stated No Not stated Six dropouts were not analysed (four declined the operation and two refused to participate postoperatively)
Falcone [25] Single centre 44 patients required; primary outcome: operation time (n = 48) Computer generated Not stated No Not stated Four patients withdrew before surgery
Marana [26] Multicentre (four centres) 116 patients required; primary outcome: complications (n = 116) Computer generated Not stated No Not stated No withdrawals or dropouts
Harkki-Siren [41] Single centre 42 patients required; primary outcome: C-reactive protein level (n = 50) Not stated Sealed opaque envelopes No Not stated No withdrawals or dropouts
Ferrari [27] Single centre None (n = 62) Computer generated Sealed opaque envelopes No Not stated Three women refused to sign informed consent and were excluded from further analysis
Lumsden [28] Multicentre (three centres) 240 patients required; primary outcome: complications (n = 200) Computer generated By third party (research nurse) No Not stated 10 dropouts were not analysed
Ottosen [29] Single centre 40 patients required; primary outcome: hospital stay (n = 120) Computer generated Sealed opaque envelopes No Not stated No withdrawals or dropouts
Lin [30] Single centre None (n = 120) Not stated Not stated No Not stated No withdrawals or dropouts
Hwang [31] Single centre Retrospective power calculation performed (n = 90) Computer generated Sealed opaque envelopes No Not stated No withdrawals or dropouts
Schutz [32] Single centre Power calculation performed; primary outcome: hospital stay (n = 48) Computer generated Telephone inquiry No Not stated No withdrawals or dropouts
Seracchioli [43] Single centre None (n = 122) Computer generated Not stated No Not stated No withdrawals or dropouts
Tsai [33] Single centre None (n = 200) Computer generated Not stated No Not stated No withdrawals or dropouts
Atabekoglu [34] Single centre None (n = 100) Computer generated Not stated Yes Not stated Three patients were excluded from the analysis: one patient in the laparoscopy group because of conversion to laparotomy due to severe adhesions, and two patients in the laparotomy group due to bladder injury and postoperative thrombophlebitis
Zhang [35] Single centre None (n = 120) Not stated No stated No Not stated No withdrawals or dropouts
Muzii [36] Multicentre (three centres) Retrospective power calculation; primary outcome: overall complications (n = 81) Computer generated Sealed opaque envelopes No Not stated No withdrawals or dropouts
Sesti [37] Single centre 72 patients required; primary outcome: hospital stay (n = 150) Computer generated Sealed opaque envelopes Yes Sequence was concealed until the interventions were assigned 28 patients refused to participate and were excluded from the analysis
Yue [38] Multicentre (two centres) None (n = 84) Not stated Not stated Yes Not stated No withdrawals or dropouts
Zhu [39] Single centre None (n = 101) Not stated Not stated No Not stated No withdrawals or dropouts

Table 4 Risk of bias of each study.

Study Adequate sequence generation (judgement) Allocation concealment (judgement) Blinding (judgement) Incomplete outcome data addressed (judgement) Free of selective reporting (judgement) Free of other bias (judgement)
Olsson [40] By opening one of 150 sealed opaque envelopes (yes) Sealed opaque envelopes (yes) Not stated (no) No dropouts (yes) Primary outcome was complications, power calculation performed (yes) Different surgeons for different procedures (unclear)
Kunz [22] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Langebrekke [23] Random number table (yes) Sealed envelopes (yes) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Different group of surgeons for different procedures (unclear)
Ellstrom [44] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcomes were postoperative pain and pulmonary function (yes) Appeared to be free of other sources of bias (yes)
Summitt [24] Computer generated (yes) Sealed opaque envelopes (yes) Not stated (no) Two women refused assigned procedure and were excluded from analysis (no) Primary outcome not defined, no power calculation reported (unclear) Analysis not according to intention to treat; funding from pharmaceutical or surgical instrumentation company (no)
Yuen [42] Computer generated (yes) Not stated (unclear) Not stated (no) Six dropouts were not analysed (four declined the operation and two refused to participate postoperatively) (no) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Falcone [25] Computer generated (yes) Not stated (unclear) Not stated (no) Four patients withdrew before surgery and data were included where possible (yes) Primary outcome was operation time, power calculation performed (yes) Funding from pharmaceutical or surgical instrumentation company (no)
Marana [26] Computer generated (yes) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome was complications, power calculation performed (yes) Appeared to be free of other sources of bias (yes)
Ferrari [27] Computer generated (yes) Sealed opaque envelopes (yes) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Lumsden [28] Computer generated (yes) By third party (yes) Not stated (no) 10 dropouts were not analysed: seven women did not attend surgery and three records were not available (no) Primary outcome was complications, power calculation performed (yes) Appeared to be free of other sources of bias (yes)
Harkki-Siren [41] By opening one of the 50 sequentially numbered, opaque and sealed envelopes (yes) Sealed opaque envelopes (yes) Not stated (no) No droputs (yes) Primary outcome was C-reactive protein level, power calculation performed (yes) Tissue trauma reported in uncomplicated surgeries only; funding from pharmaceutical or surgical instrumentation company (no)
Ottosen [29] Computer generated (yes) Sealed opaque envelopes (yes) Not stated (no) No dropouts (yes) Primary outcome was hospital stay, power calculation performed (yes) Appeared to be free of other sources of bias (yes)
Lin [30] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Hwang [31] Computer generated (yes) Sealed opaque envelopes (yes) Not stated (no) No dropouts (yes) Retrospective power calculation (yes) Appeared to be free of other sources of bias (yes)
Schutz [32] Computer generated (yes) Telephone inquiry (yes) Not stated (no) No dropouts (yes) Primary outcome was hospital stay, power calculation performed (yes) More residents in training as first surgeons in AH (unclear)
Seracchioli [43] Computer generated (yes) Allocation reported as ‘unknown to surgeons’ (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Tsai [33] Computer generated (yes) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Analysis not according to intention to treat (with two conversions from LAVH to AH) (no)
Atabekoglu [34] Computer generated (yes) Not stated (unclear) Mentioned as blinded but methods not stated (unclear) Three patients were excluded from the analysis: one patient in the laparoscopy group because of conversion to laparotomy due to severe adhesions, and two patients in the laparotomy group due to bladder injury and postoperative thrombophlebitis (no) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Zhang [35] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Muzii [36] Computer generated (yes) Sealed opaque envelopes (yes) Not stated (no) No dropouts (yes) Primary outcome was overall complications, retrospective power calculation performed (yes) Appeared to be free of other sources of bias (yes)
Sesti [37] Computer generated (yes) Sealed opaque envelopes (yes) Sequence was concealed until the interventions were assigned (yes) No dropouts (yes) Primary outcome was hospital stay, power calculation performed (yes) Appeared to be free of other sources of bias (yes)
Yue [38] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Mentioned as blinded but methods not stated (unclear) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)
Zhu [39] Mentioned as randomized but methods not stated (unclear) Not stated (unclear) Not stated (no) No dropouts (yes) Primary outcome not defined, no power calculation reported (unclear) Appeared to be free of other sources of bias (yes)

AH, abdominal hysterectomy; LAVH, laparoscopic-assisted vaginal hysterectomy.

gr2

Fig. 2 Judgements about risk of bias for each item, presented as percentages across all included studies.

3.1. Assessment of study quality

3.1.1. Adequate sequence generation

Six studies [22], [30], [35], [38], [39], and [44] stated that they were randomized but did not report the randomization method. Seventeen studies [23], [24], [25], [26], [27], [28], [29], [31], [32], [33], [34], [36], [37], [40], [41], [42], and [43] reported an adequate randomization method, of which 14 studies [24], [25], [26], [27], [28], [29], [31], [32], [33], [34], [36], [37], [42], and [43] used a computer-generated sequence, one study [23] used a randomized number table, and two studies [40] and [41] used sealed opaque envelopes by opening individually.

3.1.2. Allocation concealment

Eleven studies [23], [24], [27], [28], [29], [31], [32], [36], [37], [40], and [41] reported allocation concealment, of which nine studies [23], [24], [27], [29], [31], [36], [37], [40], and [41] used sealed opaque envelopes, one study [28] was conducted by a third party, and one study [32] used telephone inquiry. Eleven studies [22], [25], [26], [30], [33], [34], [35], [38], [39], [42], and [44] did not report allocation concealment and one study [43] just stated ‘unknown to surgeons’.

3.1.3. Blinding

Twenty studies [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [35], [36], [39], [40], [41], [42], [43], and [44] did not report blinding, and two studies [34] and [38] indicated that blinding was used but did not report the method. Only one study [37] reported an adequate blinding method.

3.1.4. Incomplete outcome data addressed

 

  • Dropouts. Eighteen studies [22], [23], [26], [27], [29], [30], [31], [32], [33], [35], [36], [37], [38], [39], [40], [41], [43], and [44] reported no dropouts, and five studies [24], [25], [28], [34], and [42] reported dropouts. Dropouts were excluded from the analysis in four studies [24], [28], [34], and [42], while one study [25] included the data in the analysis where possible.
  • Intention to treat. Five studies reported dropouts, of which two studies [25] and [28] specifically reported intention to treat, and three studies [24], [34], and [42] did not. One study [33] did not report dropouts, but its analysis was undertaken according to the treatments received rather than intention to treat (two patients converted from LAVH to AH, and were analysed in the AH group).
3.1.5. Free of selective reporting

Twelve studies [22], [23], [24], [27], [30], [33], [34], [35], [38], [39], [42], and [43] did not report their primary outcome or power calculations, so it was unclear whether the outcomes were reported in the predefined way. Eleven studies [25], [26], [28], [29], [31], [32], [36], [37], [40], [41], and [44] reported their sample size calculation and predefined their primary outcome.

Operation time and hospital stay were assessed in all of the studies. Blood loss was assessed in 16 studies [24], [25], [26], [29], [30], [31], [32], [33], [34], [35], [37], [38], [39], [41], [42], and [43]. Postoperative pain was assessed in nine studies [23], [25], [26], [30], [32], [36], [37], [39], and [44], with three studies [23], [25], and [37] presenting graphs. Haemoglobin drop was assessed in nine studies [23], [26], [27], [30], [32], [36], [41], [42], and [43]. Return to normal activities was assessed in nine studies [23], [24], [25], [29], [31], [32], [40], [41], and [43], with one study [25] presenting a graph. Eighteen studies [22], [23], [24], [25], [26], [27], [28], [29], [31], [32], [33], [36], [37], [40], [41], [42], [43], and [44] assessed complications, and one study [28] categorized the complications into major and minor. Twelve studies [24], [25], [27], [28], [29], [33], [34], [36], [40], [41], [42], and [43] reported conversions in surgery.

3.1.6. Free of other bias

Sixteen studies [22], [26], [27], [29], [30], [31], [34], [35], [36], [37], [38], [39], [42], [43], and [44] appeared to be free of other sources of bias, whereas seven studies [23], [24], [25], [32], [33], [40], and [41] had other potential sources of bias. Two studies [23] and [40] reported different surgeons for different procedures, and one study [32] reported more residents in training as first surgeons in AH. Three studies [24], [25], and [41] reported the receipt of funding; for one study [24] , this was not according to intention to treat, and one study [41] only reported tissue trauma in uncomplicated operations. One study [33] reported no funding, but its analysis was not according to intention to treat.

3.2. Meta-analysis results of operative outcomes and complications

Fig. 3 shows the forest plots for operative outcomes and complications. A random effect model was used for operation time, blood loss, hospital stay and return to normal activities because heterogeneity was significant (p < 0.05).

gr3a gr3b gr3c gr3d gr3e

Fig. 3 Forest plots of operative outcomes and complications. Random, random effects model; fixed, fixed effects model; LAVH, laparoscopic-assisted vaginal hysterectomy; AH, abdominal hysterectomy; SD, standard deviation; IV, inverse variance; M–H, Mantel–Haenszel; CI, confidence interval.

Suitable data were reported and included in the analysis as follows: 15 studies [22], [24], [26], [29], [30], [33], [34], [35], [37], [38], [39], [41], [43], and [44] on operation time, 16 studies [24], [25], [26], [29], [30], [31], [32], [33], [34], [35], [37], [38], [39], [41], [42], and [43] on blood loss, 13 studies [22], [24], [26], [29], [30], [31], [33], [34], [35], [37], [38], [39], and [41] on hospital stay, three studies [26], [30], and [39] on postoperative pain, 4 studies [26], [30], [41], and [43] on haemoglobin drop, 6 studies [24], [29], [31], [40], [41], and [43] on return to normal activities, 8 studies [22], [23], [26], [28], [29], [31], [33], and [40] on major complications and 17 studies [23], [24], [25], [26], [27], [28], [29], [31], [32], [33], [36], [37], [40], [41], [42], [43], and [44] on minor complications. Compared with AH, LAVH was shown to have a longer operation time (MD 13.62 min, 95% CI 4.60–22.65 min; p = 0.003), less blood loss (MD −47.92 ml, 95% CI −77.79 to −18.06 ml; p = 0.002), shorter hospital stay (MD −2.11 days, 95% CI −2.63 to −1.59 days; p < 0.00001), less postoperative pain (day 1: MD −1.48, 95% CI −1.95 to −1.01, p < 0.00001; day 2: MD −2.07, 95% CI −2.49 to −1.66, p < 0.00001; day 3: MD −1.81, 95% CI −2.25 to −1.37, p < 0.00001), smaller haemoglobin drop (MD −0.52 g/100 ml, 95% CI −0.73 to −0.31 g/100 ml; p < 0.00001), quicker return to normal activities (MD −13.32 days, 95% CI −16.77 to −9.88 days; p < 0.00001), more major complications (OR 2.54, 95% CI 1.13–5.70; p = 0.02), fewer minor complications (OR 0.50, 95% CI 0.36–0.70; p = 0.0001) and fewer overall complications (OR 0.60, 95% CI 0.44 to −0.81; p = 0.0008); these differences were significant.

The data from eight studies [23], [25], [27], [28], [31], [32], [36], [40], and [42] on operation time, three studies [25], [32], and [42] on blood loss, nine studies [23], [25], [27], [28], [32], [36], [40], [42], and [43] on hospital stay, six studies [23], [25], [32], [36], [37], and [44] on postoperative pain, five studies [23], [27], [32], [36], and [42] on haemoglobin drop and three studies [23], [25], and [32] on return to normal activities were summarized beyond the analysis ( Table 5 ) as they were expressed in a form unsuitable for meta-analysis. These data presented similar results.

Table 5 Descriptive data of excluded studies.

Study No. LAVH vs no. AH Data form LAVH AH p-Value
Operation time (min)
Langebrekke [23] 46 vs 54 Median (range) 100 (50 to 153) 60.5 (22 to 105) NS
Falcone [25] 23 vs 21 Median (quartiles) 180 (139 to 225) 130 (97 to 155) <0.01
Ferrari [27] 31 vs 31 Median (IQR) 135 (115 to 173) 120 (98 to 123) 0.001
Lumsden [28] 95 vs 95 Median (mean) 80 (81.9) 45 (47.3) Significantly less for AH
Hwang [31] 30 vs 30 Mean (SD) With second procedure: 119 (20) (n = 13) With second procedure: 117 (32) (n = 8) With second procedure: 0.12
      Without second procedure: 109 (22) (n = 17) Without second procedure: 98 (16) (n = 22) Without second procedure: <0.001
Schutz [32] 28 vs 20 Median (25–75th percentile) 133 (120 to 160) 132 (121 to 145) NS
Muzii [36] 40 vs 41 Median (range) 86 (60 to 120) 58 (45 to 75) <0.05
Yuen [42] 20 vs 24 Median (range) 95 (79 to 143) 105 (86 to 120) NS
Blood loss (ml)
Falcone [25] 23 vs 21 Median (quartiles) 450 (250 to 700) 250 (150, 300) 0.003
Schutz [32] 28 vs 20 Median (IQR) 200 (150 to 280) 600 (400 to 1225) <0.05
Yuen [42] 20 vs 24 Median (range) 200 (150 to 350) 450 (300 to 800) <0.01
Hospital stay (days)
Langebrekke [23] 46 vs 54 Median (range) 2 (0 to 5) 5 (3 to 12) <0.001
Falcone [25] 23 vs 21 Median (quartiles) 1.5 (1.0 to 2.3) 2.5 (1.5 to 2.5) 0.038
Ferrari [27] 31 vs 31 Median (IQR) 3.8 (3.8 to 4.0) 5.8 (5.3 to 6.3) <0.001
Lumsden [28] 95 vs 95 Median (mean) 4.0 (4.0) 6.0 (5.7) Significantly less for LAVH
Schutz [32] 28 vs 20 Median (25–75th percentile) 6.5 (5 to 7) 10 (8.25 to 11) <0.05
Muzii [36] 40 vs 41 Median (range) 2 (1 to 3) 3 (1 to 5) 0.53
Olsson [40] 71 vs 72 Median (range) 2.0 (1 to 12) 4.0 (2 to 28) <0.001
Yuen [42] 20 vs 24 Median (range) 4 (4 to 5) 6 (5 to 9) <0.01
Seracchioli [43] 60 vs 62 Mean (SD) 76.4 (30.4) (h) 121.8 (41.8) (h) <0.001
Postoperative pain (VAS)
Langebrekke [23] 46 vs 54 Graph Shown in graph with no data Shown in graph with no data Not tested
Ellstrom [44] 20 vs 20 Mean (SD) Day 0 8.00 p.m. Day 0 8.00 p.m. Day 0 8.00 p.m.
      At rest 22 (16) At rest 36 (26) At rest NS
      Coughing 29 (20) Coughing 48 (30) Coughing NS
      Day 1 10.00 a.m. Day 1 10.00 a.m. Day 1 10.00 a.m.
      At rest 17 (16) At rest 30 (24) At rest NS
      Coughing 32 (19) Coughing 53 (30) Coughing <0.05
      Day 1 6.00 p.m. Day 1 6.00 p.m. Day 1 6.00 p.m.
      At rest 24 (20) At rest 28 (24) At rest NS
      Coughing 31 (25) Coughing 52 (28) Coughing NS
      Day 2 10.00 a.m. Day 2 10.00 a.m. Day 2 10.00 a.m.
      At rest 10 (10) At rest 20 (22) At rest NS
      Coughing 15 (14) Coughing 47 (31) Coughing <0.01
Falcone [25] 23 vs 21 Graph Shown in figure with no data Shown in figure with no data NS
Schutz [32] 28 vs 20 Median (25–75th percentile) 0 (0 to 1.75) 5 (4 to 6) <0.05
Sesti [37] 50 vs 50 Graph Shown in graph with no data Shown in graph with no data Not tested separately
Muzii [36] 40 vs 41 Median (range) Day 1: 2.8 (0 to 6) Day 1: 4.4 (2 to 6.2) Day 1: <0.05
      Day 2: 0.8 (0 to 3.7) Day 2: 2.9 (2 to 5.5) Day 2: <0.05
Haemoglobin drop (g/100 ml)
Langebrekke [23] 46 vs 54 Median 2 (g/l) 1.9 (g/l) Not tested
Ferrari [27] 31 vs 31 Median (IQR) 1.1 (0.8 to 1.9) 1.8 (0.7 to 2.5) NS
Schutz [32] 28 vs 20 Median (25–75th percentile) −0.6 (−1.25 to −0.20) −1.55 (−2.67 to −0.50) <0.05
Muzii [36] 40 vs 41 Median (range) 1.7 (1.2 to 2) 1.4 (0.4 to 2.1) 0.10
Yuen [42] 20 vs 24 Median (range) 1.2 (0.8 to 2.3) 1.7 (0.5 to 2.8) NS
Return to normal activities
Langebrekke [23] 46 vs 54 Median (range) 19.5 (0 to 140) 36.5 (23 to 259) <0.001
Falcone [25] 23 vs 21 Graph Shown in graph with no data Shown in graph with no data NS
Schutz [32] 28 vs 20 Median 42 42 Not tested

AH, abdominal hysterectomy; LAVH, laparoscopic-assisted vaginal hysterectomy; IQR, interquartile range; SD, standard deviation; VAS, visual analogue scale; NS, not significant.

3.3. Publication bias

Fig. 4 shows the funnel plot for operation time. The result of Egger's test showed no clear evidence of asymmetry (p = 0.2173), indicating no significant publication bias.

gr4

Fig. 4 Funnel plot of operation time. Studies are marked with squares. SE, summary estimate of effect; MD, mean difference. Egger's test indicated no significant asymmetry (p = 0.2173).

3.4. Meta-regression, sensitivity and subgroup analyses

As the results of operation time, blood loss, hospital stay and return to normal activities presented I2 > 50%, meta-regression analyses ( Table 6 ) indicated that power calculations (p = 0.0122) and adequate sequence generations (p = 0.0243) were the possible sources of heterogeneity in the analysis of hospital stay.

Table 6 Results of meta-regression based on centre setting, year of publication, power calculation, adequate sequence generation and laparoscopic-assisted vaginal hysterectomy (LAVH) categorization in the analysis of operation time, blood loss, hospital stay and return to normal activities.

  Estimate SE t-Value p-Value
Operation time
Centre setting −2.6830 15.1650 −0.177 0.8620
Year of publication −2.1170 1.2510 −1.692 0.1140
Power calculation 11.4110 11.906 0.958 0.3550
Adequate sequence generation 0.3223 11.7744 0.027 0.9790
LAVH categorization −18.6000 10.6700 −1.743 0.1049
Blood loss
Centre setting −11.7000 47.2100 −0.249 0.8080
Year of publication 1.2250 5.27000 0.232 0.8200
Power calculation −21.110 38.8300 −0.544 0.5970
Adequate sequence generation −48.7300 35.1600 −1.386 0.1930
LAVH categorization −45.9000 35.1900 −1.304 0.2190
Hospital stay
Centre setting 0.1874 1.0669 0.176 0.8638
Year of publication 0.07998 0.09127 0.876 0.4000
Power calculation 1.5983 0.6126 2.609 0.0243
Adequate sequence generation 1.7360 0.5800 2.993 0.0122
LAVH categorization −0.1521 0.7801 −0.195 0.8490
Return to normal activities
Centre setting 3.8730 4.9370 0.784 0.4767
Year of publication 0.6206 0.8280 0.750 0.4950
Power calculation −1.610 3.8840 −0.414 0.6998
Adequate sequence generation NA NA NA NA
LAVH categorization 6.0620 3.6440 1.663 0.1716

NA, all the included studies (for return to normal activities) reported adequate sequence generation; SE, standard error.

Sensitivity analyses showed that similar results were obtained with fixed and random effects models, except for shorter operation time (MD 4.52 min, 95% CI 3.07–5.97 min; p < 0.00001), more blood loss (MD −37.81 ml, 95% CI −41.99 to −33.63 ml; p < 0.00001), shorter hospital stay (MD −1.35 days, 95% CI −1.42 to −1.28 days; p < 0.00001) and quicker return to normal activities (MD −13.63 days, 95% CI −15.42 to −11.84 ml; p < 0.00001) when the fixed effect model was used. Also, influence analyses ( Table 7 ) indicated that no study had an overwhelming influence on the overall MD estimates, since no point estimate of any one study's omitted analysis lay outside the CI of the combined analysis, and no statistical significance changed, with the exception of a decrease in I2 from 71.2% to 47.9% in return to normal activities when Ottosen et al.’s study [29] was omitted; the difference was found to be due to LAVH categorization.

Table 7 Influence analysis (random effects model) of operation time, blood loss, hospital stay and return to normal activities.

Omitting study Mean difference 95% CI p-Value τ 2 I2 (%)
Operation time
Omitting Atabekoglu [34] 12.5697 3.2821 to 21.8573 0.0080 277.2988 96.5
Omitting Ellstrom [44] 11.7814 2.6356 to 20.9272 0.0116 273.2113 96.5
Omitting Harkki-Siren [41] 12.4667 3.2806 to 21.6528 0.0078 268.7156 96.2
Omitting Kunz [22] 15.1008 5.6327 to 24.5688 0.0018 288.6014 96.6
Omitting Lin [30] 13.5802 4.1110 to 23.0494 0.0049 288.7224 96.6
Omitting Marana [26] 14.7172 5.2151 to 24.2194 0.0024 290.9409 96.6
Omitting Olsson [40] 10.2945 1.8932 to 18.6957 0.0163 220.7663 95.6
Omitting Ottosen [29] 12.1240 2.9369 to 21.3111 0.0097 270.5866 96.4
Omitting Seracchioli [43] 14.1706 4.6798 to 23.6614 0.0034 290.4327 96.6
Omitting Sesti [37] 15.4586 5.6988 to 25.2183 0.0019 306.5779 94.7
Omitting Summitt [24] 12.7813 3.5820 to 21.9806 0.0065 280.0787 96.6
Omitting Tsai [33] 16.4945 7.6896 to 25.2995 0.0002 244.1573 95.8
Omitting Yue [38] 14.3135 2.8006 to 25.8265 0.0148 440.6873 96.1
Omitting Zhang [35] 14.5162 5.0113 to 24.0211 0.0028 291.2003 96.6
Omitting Zhu [39] 14.2575 4.6385 to 23.8766 0.0037 298.1149 96.6
Pooled estimate 13.6211 4.5958 to 22.6463 0.0031 280.8366 96.3
Blood loss
Omitting Atabekoglu [34] −41.1282 −71.6544 to −10.6020 0.0083 1881.6541 91.5
Omitting Harkki-Siren [41] −42.8528 −73.8205 to −11.8851 0.0067 1934.4421 91.7
Omitting Hwang [31] −53.0768 −83.6613 to −22.4923 0.0007 1936.5736 91.8
Omitting Lin [30] −45.4068 −76.4530 to −14.3605 0.0041 1973.5274 91.9
Omitting Marana [26] −45.1723 −76.1505 to −14.1941 0.0043 1968.3945 91.9
Omitting Ottosen [29] −54.2881 −84.5936 to −23.9827 0.0004 1901.7436 91.7
Omitting Seracchioli [43] −46.6245 −77.8701 to −15.3788 0.0034 1992.1574 91.9
Omitting Sesti [37] −35.7281 −61.0766 to −10.3796 0.0057 1016.2516 81.3
Omitting Summitt [24] −47.3599 −77.5088 to −17.2109 0.0021 1962.4733 92.0
Omitting Tsai [33] −49.2381 −81.4815 to −16.9948 0.0028 2085.1029 92.0
Omitting Yue [38] −49.1119 −93.8520 to −4.3718 0.0314 4746.9893 91.9
Omitting Zhang [35] −54.4439 −89.0277 to −19.8600 0.0020 2458.5485 90.0
Omitting Zhu [39] −55.5477 −87.5273 to −23.5681 0.0007 2006.1354 91.0
Pooled estimate −47.9242 −77.7915 to −18.0569 0.0017 1946.5817 91.3
Hospital stay
Omitting Kunz [22] −1.8447 −2.3126 to −1.3769 <0.0001 0.6075 95.2
Omitting Summitt [24] −2.1233 −2.6696 to −1.5769 <0.0001 0.8467 96.5
Omitting Marana [26] −2.1325 −2.6809 to −1.5842 <0.0001 0.8524 96.5
Omitting Harkki-Siren [41] −2.1935 −2.7861 to −1.6009 <0.0001 1.0010 96.5
Omitting Ottosen [29] −2.2438 −2.7940 to −1.6936 <0.0001 0.8553 96.4
Omitting Lin [30] −2.0741 −2.6133 to −1.5349 <0.0001 0.8205 96.4
Omitting Hwang [31] −2.2068 −2.7491 to −1.6645 <0.0001 0.8405 96.5
Omitting Tsai [33] −2.0992 −2.6548 to −1.5436 <0.0001 0.8689 96.0
Omitting Zhang [35] −2.0206 −2.5417 to −1.4995 <0.0001 0.7598 96.0
Omitting Atabekoglu [34] −2.1572 −2.7076 to −1.6068 <0.0001 0.8590 96.5
Omitting Sesti [37] −2.2110 −2.7305 to −1.6914 <0.0001 0.7466 92.9
Omitting Zhu [39] −2.0660 −2.5938 to −1.5383 <0.0001 0.7759 95.7
Omitting Yue [38] −2.1228 −2.6859 to −1.5596 <0.0001 0.8963 96.3
Pooled estimate −2.1141 −2.6344 to −1.5938 <0.0001 0.8270 96.2
Return to normal activities
Omitting Harkki-Siren [41] −12.4107 −16.1600 to −8.6613 <0.0001 11.8986 66.8
Omitting Hwang [31] −13.6461 −17.5194 to −9.7729 <0.0001 14.6117 76.1
Omitting Olsson [40] −12.3367 −15.9664 to −8.7070 <0.0001 11.3326 68.6
Omitting Ottosen [29] −14.6998 −17.6310 to −11.7687 <0.0001 5.2147 47.9
Omitting Seracchioli [43] −13.1282 −17.4342 to −8.8222 <0.0001 17.9752 76.9
Omitting Summitt [24] −13.8669 −17.7066 to −10.0272 <0.0001 13.926 74.6
Pooled estimate −13.3219 −16.7682 to −9.8757 <0.0001 12.7797 71.2

Method: inverse variance method; CI, confidence interval.

Three subgroup analyses were performed for power calculation, adequate sequence generation and LAVH categorization ( Fig. 5 ). The results showed that I2 values were decreased, but also indicated that these three factors were not the main sources of heterogeneity as high or moderate heterogeneity still existed. However, LAVH continued to lead to shorter hospital stay and quicker return to normal activities compared with AH.

gr5a gr5b

Fig. 5 Subgroup forest plot of hospital stay based on power calculation and adequate sequence generation, and return to normal activities based on laparoscopic-assisted vaginal hysterectomy (LAVH) categorization. Random, random effects model; AH, abdominal hysterectomy; LAVH(a), laparoscopic hysterectomy where laparoscopic procedures include uterine artery ligation; LAVH(b), laparoscopic hysterectomy where the laparoscopic component did not involve ligation of the uterine vessels; SD, standard deviation; IV, inverse variance; CI, confidence interval.

4. Discussion

This study performed a meta-analysis to evaluate all randomized controlled trials that compared LAVH with AH for benign disease, according to eligibility criteria. The search strategy was broad and had no language limitation to reduce any potential bias. According to the quality assessment, most studies had a high risk of bias due to blinding and a low risk of other forms of bias.

Meta-regression and subgroup analyses indicated that power calculation, adequate sequence generation and LAVH categorization may be possible sources of heterogeneity, but these three factors were not the main sources of heterogeneity as high or moderate heterogeneity still existed. Sensitivity analysis indicated sufficient stability of analyses results. As such, this meta-analysis was potentially limited by clinical rather than statistical heterogeneity, due to various factors in the treatment of benign disease by hysterectomy, such as different types of disease (myoma, menorrhagia, adenomyosis, etc.), and variations in surgical procedures and skill level of surgeons.

Although vaginal hysterectomy offers fewer postoperative complications, less blood loss, shorter hospital stay and convalescence, and lower costs compared with AH [45] and [46], many surgeons do not feel comfortable when confronted with traditional contra-indications of the vaginal route [36] . LAVH, which combines the advantages of laparoscopy and the vaginal route, has many advantages compared with AH, although it also has some disadvantages, such as longer operation time and more operative bleeding, compared with vaginal hysterectomy [47] .

The results show that LAVH has a significantly longer operation time than AH. As most time was spent in the laparoscopic phase of the operation, this may be due to the learning curve for laparoscopy. Laparoscopy is a complex technique requiring a high level of skill and good hand–eye co-ordination as the surgeon receives limited tactile feedback. However, comprehensive training of surgeons and the development of surgical instruments may lead to a decrease in the operation time for LAVH in the future. Another contributing factor was the volume of the myoma or uterus; if the target volume was too large, more procedures were required, and this was time consuming. Even experienced surgeons would reduce their speed to guarantee safety in these circumstances. Nevertheless, with more research into methods of volume reduction, LAVH will become easier and faster.

The study found that blood loss was significantly lower for LAVH than AH. This may be due to more precise and accurate anatomical location of bleeding points, better sight of deep vascular structures and less tissue trauma. In the laparoscopic phase of LAVH, clear visualization and the zoom function makes it easier to locate the bleeding points. The vaginal phase of LAVH offers familiar sights and senses to control bleeding, as for AH [48] . Furthermore, as LAVH is performed through three trocar ports and AH involves abdominal incision, LAVH leads to less tissue trauma. However, it should be noted that the experience of surgeons will affect the control of bleeding.

This study found that hospital stay was significantly shorter following LAVH compared with AH. As LAVH leads to less tissue and visceral trauma, and less inflammation [49] and [50], faster recovery is possible, thus decreasing the length of hospital stay.

Postoperative pain and haemoglobin drop were reduced significantly, and return to normal activities was significantly quicker following LAVH compared with AH. However, this meta-analysis was not sufficiently powerful to draw conclusions as few studies reported these outcomes; more data are needed for evaluation. Nevertheless, the results might provide evidence for less blood loss and shorter hospital stay following LAVH, although the conclusions need further review.

This study found that LAVH led to significantly more major complications, and fewer minor and overall complications than AH. Most major complications were bladder and urinary tract injuries, probably caused by the resident surgeons. These complications would be less common among more experienced surgeons [24] . The feasibility and safety of LAVH have been accepted worldwide, and the finding of fewer minor and overall complications following LAVH compared with AH confirms this concept. Since LAVH leads to fewer tissue and visceral injuries, there is less need for blood transfusions, and fewer cases of urinary and vaginal cuff infection, haematomas, etc. Surgery with a low complication rate implies a shorter hospital stay [24] , so these results may explain the shorter hospital stay following LAVH. However, the greatest benefit from LAVH is achieved with appropriate selection of patients and surgeons.

Although various outcomes of hysterectomy have been reported worldwide, there has been no discussion regarding which outcomes are of key importance [7] . Quality of life is likely to be the most key outcome because it reflects patients’ experiences of treatment and takes into account the effects of complications [6] and [51]. Questionnaires such as Short Form-36/Short Form-12 or a quality-of-life scale, such as quality-adjusted life years, could be used to assess quality of life. However, few studies have reported this outcome and further evaluation is needed.

There is no doubt that individual surgeons have their own indications when choosing the approach for hysterectomy. However, the advantages and disadvantages of each approach should be taken into account when making a decision.

5. Conclusion

This study showed that LAVH had a longer operation time, less blood loss, shorter hospital stay, smaller haemoglobin drop, less postoperative pain, quicker return to normal activities and fewer peri-operative complications compared with AH. For suitable patients and surgeons, LAVH is a better choice than AH. However, more studies and more data, particularly regarding quality of life, are needed for further evaluation.

References

  • [1] L.S. Wilcox, L.M. Koonin, R. Pokras, L.T. Strauss, Z. Xia, H.B. Peterson. Hysterectomy in the United States 1988–1990. Obstet Gynecol. 1994;83:555-594
  • [2] T. Falcone, M. Walters. Hysterectomy for benign disease. Obstet Gynecol. 2008;111:753-767 Crossref.
  • [3] H. Reich, J. DeCaprio, F. McGlynn. Laparoscopic hysterectomy. J Gynecol Surg. 1989;5:213-216 Crossref.
  • [4] N. Johnson, D. Barlow, A. Lethaby, E. Tavender, L. Curr, R. Garry. Methods of hysterectomy: systematic review and meta-analysis of randomised controlled trials. BMJ. 2005;330:1478 Crossref.
  • [5] M. Sculpher, A. Manca, J. Abbott, J. Fountain, S. Mason, R. Garry. Cost effectiveness analysis of laparoscopic hysterectomy compared with standard hysterectomy. BMJ. 2004;328:134-137
  • [6] K. Kluivers, N. Johnson, P. Chien, M. Vierhout, M. Bongers, B. Mol. Comparison of laparoscopic and abdominal hysterectomy in terms of quality of life: a systematic review. Eur J Obstet Gynecol Reprod Biol. 2008;136:3-8 Crossref.
  • [7] T.E. Nieboer, N. Johnson, A. Lethaby, et al. Surgical approach to hysterectomy for benign gynaecological disease. Cochrane Database Syst Rev. 2009;3:CD003677
  • [8] Z.H. Huang, R.Q. Tu, L.J. Wu. Comparison among minor invasive surgical approaches to hysterectomy for benign gynecological diseases: a systematic review and meta-analysis. Chin J Evid Based Med. 2009;10:323-338
  • [9] C. Walsh, S. Walsh, T. Tang, M. Slack. Total abdominal hysterectomy versus total laparoscopic hysterectomy for benign disease: a meta-analysis. Eur J Obstet Gynecol Reprod Biol. 2009;144:1-2
  • [10] H. Reich, L. Roberts. Laparoscopic hysterectomy in current gynecological practice. Rev Gynaecol Pract. 2003;3:32-40 Crossref.
  • [11] C. Chapron, A. Fauconnier, F. Goffinet, G. Bréart, J. Dubuisson. Laparoscopic surgery is not inherently dangerous for patients presenting with benign gynaecologic pathology. Hum Reprod. 2000;17:1334-1342
  • [12] J.P.T. Higgins, S. Green. Cochrane handbook for systematic reviews of interventions. Version 5.0.2. (The Cochrane Collaboration, 2009)
  • [13] R Development Core Team. R: a language and environment for statistical computing. (R Foundation for Statistical Computing, Vienna, 2010) http://www.R-project.org Available at:
  • [14] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535 Crossref.
  • [15] A. Liberati, D.G. Altman, J. Tetzlaff, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b2700 Crossref.
  • [16] S.E. Brian, H. Torsten. A handbook of statistical analyses using R. (CRC Press, 2006)
  • [17] H.C. van Houwelingen, L.R. Arends, T. Stijnen. Advanced methods in meta-analysis: multivariate approach and meta-regression. Stat Med. 2002;21:589-624 Crossref.
  • [18] C.S. Berkey, D.C. Hoaglin, F. Mosteller, G.A. Colditz. A random-effects regression model for meta-analysis. Stat Med. 1995;14:395-411 Crossref.
  • [19] N. Mantel, W. Haenszel. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22:719-748
  • [20] S. Yusuf, R. Peto, J. Lewis, R. Collins, P. Sleight. Beta blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis. 1985;27:335-371 Crossref.
  • [21] M. Egger, S.G. Davey, M. Schneider, C. Minder. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315:629-634 Crossref.
  • [22] G. Kunz, T. Plath, G. Leyendecker. Comparison between laparoscopic assisted vaginal hysterectomy (LAVH) and abdominal hysterectomy. Geburtsh Frauenheilk. 1996;56:453-457 Crossref.
  • [23] A. Langebrekke, R. Eraker, B.I. Nesheim, A. Urnes, B. Busund, G. Sponland. Abdominal hysterectomy should not be considered as a primary method for uterine removal. A prospective randomised study of 100 patients referred to hysterectomy. Acta Obstet Gynecol Scand. 1996;75:404-407 Crossref.
  • [24] R.L. Summitt, T.G. Stovall, J.F. Steege, G.H. Lipscomb. A multicenter randomized comparison of laparoscopically assisted vaginal hysterectomy and abdominal hysterectomy in abdominal hysterectomy candidates. Obstet Gynecol. 1998;92:321-326
  • [25] T. Falcone, M.F.R. Paraiso, E. Mascha. Prospective randomized clinical trial of laparoscopically assisted vaginal hysterectomy versus total abdominal hysterectomy. Am J Obstet Gynecol. 1999;180:955-962 Crossref.
  • [26] R. Marana, M. Busacca, E. Zupi, N. Garcea, P. Paparella, G.F. Catalano. Laparoscopically assisted vaginal hysterectomy versus total abdominal hysterectomy: a prospective, randomized, multicenter study. Am J Obstet Gynecol. 1999;180:270-275 Crossref.
  • [27] M.M. Ferrari, N. Berlanda, R. Mezzopane, G. Ragusa, M. Cavallo, G. Pardi. Identifying the indications for laparoscopically assisted vaginal hysterectomy: a prospective, randomised comparison with abdominal hysterectomy in patients with symptomatic uterine fibroids. BJOG. 2000;107:620-625 Crossref.
  • [28] M.A. Lumsden, S. Twaddle, R. Hawthorn, et al. A randomised comparison and economic evaluation of laparoscopic-assisted hysterectomy and abdominal hysterectomy. BJOG. 2000;107:1386-1391 Crossref.
  • [29] C. Ottosen, G. Lingman, L. Ottosen. Three methods for hysterectomy: a randomised, prospective study of short term outcome. BJOG. 2000;107:1380-1385
  • [30] S.Q. Lin, J. Bai, W. Felix. Clinical study of laparoscopically assisted vaginal hysterectomy versus total abdominal hysterectomy. Clin Med J China. 2001;8:483-484
  • [31] J.L. Hwang, K.M. Seow, Y.L. Tsai, L.W. Huang, B.C. Hsieh, C. Lee. 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. 2002;81:1132-1138
  • [32] K. Schutz, M. Possover, A. Merker, W. Michels, A. Schneider. Prospective randomized comparison of laparoscopic-assisted vaginal hysterectomy (LAVH) with abdominal hysterectomy (AH) for the treatment of the uterus weighing >200 g. Surg Endosc. 2002;16:121-125
  • [33] E.M. Tsai, H.S. Chen, C.Y. Long, et al. Laparoscopically assisted vaginal hysterectomy versus total abdominal hysterectomy: a study of 100 cases on light-endorsed transvaginal section. Gynecol Obstet Invest. 2003;55:105-109 Crossref.
  • [34] C. Atabekoglu, M. Sonmezer, M. Gungor, R. Aytac, F. Ortac, C. Unlu. Tissue trauma in abdominal and laparoscopic-assisted vaginal hysterectomy. J Am Assoc Gynecol Laparosc. 2004;11:467-472 Crossref.
  • [35] P. Zhang, F. Ling, Y. Zeng. A prospective randomized study of laparoscopic assisted vaginal hysterectomy versus abdominal hysterectomy. ShangHai Med J. 2004;27:736-738
  • [36] L. Muzii, S. Basile, E. Zupi, et al. Laparoscopic-assisted vaginal hysterectomy versus minilaparotomy hysterectomy: a prospective, randomized, multicenter study. J Minim Invasive Gynecol. 2007;14:610-615 Crossref.
  • [37] F. Sesti, F. Calonzi, V. Ruggeri, A. Pietropolli, E. Piccione. A comparison of vaginal, laparoscopic-assisted vaginal, and minilaparotomy hysterectomies for enlarged myomatous uteri. Int J Gynecol Obstet. 2008;103:227-231 Crossref.
  • [38] Q. Yue, R. Ma, D.W. Mao, et al. Effects of laparoscopically-assisted vaginal hysterectomy compared with abdominal hysterectomy on immune function. J Int Med Res. 2009;37:855-861 Crossref.
  • [39] L. Zhu, J.H. Lang, C.Y. Liu, H.H. Shi, Z.J. Sun, R. Fan. Clinical assessment for three routes of hysterectomy. Chin Med J (Engl). 2009;122:377-380
  • [40] J.H. Olsson, M. Ellstrom, M. Hahlin. A randomised prospective trial comparing laparoscopic and abdominal hysterectomy. BJOG. 1996;103:345-350 Crossref.
  • [41] P. Härkki-Sirén, J. Sjöberg, J. Toivonen, A. Tiitinen. Clinical outcome and tissue trauma after laparoscopic and abdominal hysterectomy: a randomized controlled study. Acta Obstet Gynecol Scand. 2000;79:866-871
  • [42] P.M. Yuen, T.W. Mak, S.F. Yim, et al. Metabolic and inflammatory responses after laparoscopic and abdominal hysterectomy. Am J Obstet Gynecol. 1998;179:1-5 Crossref.
  • [43] R. Seracchioli, S. Venturoli, F. Vianello, et al. Total laparoscopic hysterectomy compared with abdominal hysterectomy in the presence of a large uterus. J Am Assoc Gynecol Laparosc. 2002;9:333-338 Crossref.
  • [44] M. Ellström, M.F. Olsén, J.H. Olsson, G. Nordberg, A. Bengtsson, M. Hahlin. Pain and pulmonary function following laparoscopic and abdominal hysterectomy: a randomized study. Acta Obstet Gynecol Scand. 1998;77:923-928
  • [45] R.C. Dicker, J.R. Greenspan, L.T. Strauss, et al. Complications of abdominal and vaginal hysterectomy among women of reproductive age in the United States. The Collaborative Review of Sterilization. Am J Obstet Gynecol. 1982;144:841-848
  • [46] J.H. Dorsey, P.M. Holtz, R.I. Griffiths, M.M. McGrath, E.P. Steinberg. Costs and charges associated with three alternative techniques of hysterectomy. N Engl J Med. 1996;335:476-482 Crossref.
  • [47] J. Drahonovsky, L. Haakova, M. Otcenasek, L. Krofta, E. Kucera, J. Feyereisl. A prospective randomized comparison of vaginal hysterectomy, laparoscopically assisted vaginal hysterectomy, and total laparoscopic hysterectomy in women with benign uterine disease. Eur J Obstet Gynecol Reprod Biol. 2010;148:172-176 Crossref.
  • [48] A.R. Twijnstra, R.N.A. Kianmanesh, M.J. Smeets, J.F. Admiraal, F.W. Jansen. Twenty-first century laparoscopic hysterectomy: should we not leave the vaginal step out?. Gynecol Surg. 2009;6:311-316 Crossref.
  • [49] C.S. Ou, E. Beadle, J. Presthus, M. Smith. A multicenter review of 839 laparoscopic-assisted vaginal hysterectomies. J Am Assoc Gynecol Laparosc. 1994;1:417-422 Crossref.
  • [50] C.C. Shen, M.P. Wu, C.H. Lu, et al. Short- and long-term clinical results of laparoscopic-assisted vaginal hysterectomy and total abdominal hysterectomy. J Am Assoc Gynecol Laparosc. 2003;10:49-54 Crossref.
  • [51] P. Chien, K. Khan, B.W. Mol. How to interpret the findings of the eVALuate study. BJOG. 2005;112:391-393 Crossref.

Footnotes

First Department of Gynaecology, Renmin Hospital of Wuhan University, Wuhan 430060, Hubei Province, China

lowast Corresponding author. Tel.: +86 13035106467.