Laparoscopic assisted radical vaginal hysterectomy versus radical abdominal hysterectomy—a randomised phase II trial: perioperative outcomes and surgicopathological measurements

Authors


Mr R Naik, Northern Gynaecological Oncology Centre, Queen Elizabeth Hospital, Sheriff Hill, Gateshead NE9 6SX, UK. Email raj.naik@ghnt.nhs.uk

Abstract

Please cite this paper as: Naik R, Jackson K, Lopes A, Cross P, Henry J. Laparoscopic assisted radical vaginal hysterectomy versus radical abdominal hysterectomy—a randomised phase II trial: perioperative outcomes and surgicopathological measurements. BJOG 2010; DOI: 10.1111/j.1471-0528.2010.02479.x.

Objective  To evaluate perioperative surgical outcomes and resection size for laparoscopically assisted radical vaginal hysterectomy (LARVH) compared with radical abdominal hysterectomy (RAH).

Design  A prospective randomised phase II trial.

Population  Early stage IB cervical cancer requiring radical surgical treatment.

Setting  Northern Gynaecological Oncology Centre, Gateshead, UK.

Methods  Fifteen women were randomised to LARVH and to RAH.

Main outcome measures  Outcomes included requirement in days for bladder catheterisation after surgery, operating time, blood loss, hospital stay, opiate pain relief, complication rate, time to normal activities and resection size of major ligaments and vaginal cuff.

Results  Statistically significant differences were found between LARVH and RAH, respectively: median duration of bladder catheterisation, 4 days versus 21 days (P = 0.003); median operating time, 180 minutes versus 138 minutes (P = 0.05); median blood loss, 400 ml versus 1000 ml (P = 0.05), median hospital stay, 5 days versus 7 days (P = 0.04) and median opiate requirement in the first 36 hours postoperatively, 30 mg versus 53 mg (P = 0.004). The mean resected lengths for LARVH versus RAH, respectively, were: mean resected vaginal cuff, 1.26 cm versus 2.16 cm (P = 0.014); mean resected cardinal ligament length, 1.30 cm versus 2.79 cm (P = 0.013) and mean resected uterosacral ligament length, 1.47 cm versus 4.68 cm (P = 0.034).

Conclusions  This study confirms the short-term surgical benefits of LARVH. In addition, LARVH has been shown to be a less radical procedure than RAH, supporting the need for strict patient selection and to restrict the procedure to small tumours.

Introduction

The technical feasibility of laparoscopically assisted radical vaginal hysterectomy (LARVH) as treatment for early-stage cervical cancer has been well established1–4 through a series of retrospective reports.

These reports suggest that LARVH may have surgical benefits over radical abdominal hysterectomy (RAH) in terms of reductions in blood loss, transfusion requirement and hospital stay, but operating time may be prolonged. That said, the majority of these reports are largely descriptive in nature and lack sufficient comparison with the standard treatment, and the technique has been introduced into clinical practice without the benefit of large-scale cohort studies or prospective randomised controlled trials.

Our institution’s preliminary analysis2 and the work of other authors1,3–5 have shown no difference with respect to recurrence or survival between RAH and LARVH, but this cannot be proved conclusively because of the small numbers of women involved. Surgical trials are notoriously difficult to design and recruit to and resolving issues of recurrence and survival by means of a trial would require a very large sample size. It is unlikely that an adequately powered randomised study will ever be undertaken to test the applicability and efficacy of these concepts because it has been estimated that at least 1400 women would be required.3

By contrast, to evaluate surgical outcomes by means of a prospective trial should be possible with smaller numbers of women. It was therefore proposed to conduct a randomised phase II trial at our institution, based around short-term outcomes and comparing LARVH with RAH for the treatment of early-stage cervical cancer.

Methods

All participants had histologically confirmed carcinoma of the cervix, International Federation of Obstetrics and Gynecology (FIGO) stage IB1, less than 2 cm in size, requiring radical hysterectomy and potentially suitable for surgery via a laparoscopic vaginal route. Unusual high-risk histological subtypes, e.g. glassy cell, clear cell, were excluded (but adenocarcinoma was not excluded). Pregnant women and those with concomitant or previous malignancy likely to interfere with treatment or comparison were also excluded. Written informed consent was obtained before randomisation. All women had tumours confirmed by diagnostic loop biopsies of the cervix. Women may also have required further investigations to determine their eligibility for the trial, i.e. physical examination under anaesthesia and magnetic resonance imaging of the pelvis and abdomen.

In specific and selected women, fertility-sparing treatment, such as radical vaginal trachelectomy (RVT), or conservative surgery, such as simple hysterectomy or cone biopsy and pelvic node dissection, was offered or requested by women in place of radical hysterectomy. These women were therefore not approached for the trial. The participants underwent a 1:1 randomisation to either laparoscopic or standard open surgery using computer-generated block randomisation. The block size was four and this information was not known to the researchers until the trial had closed. Entry required completion of a password-controlled online randomisation form developed by the Centre for Health Services Research at the University of Newcastle upon Tyne, UK.

LARVH was not offered outside the trial, i.e. women who considered but declined entry to the trial were allocated standard open radical surgery. A record of these women was kept in accordance with the CONSORT guidelines6 for the reporting of randomised controlled trials. The women underwent a baseline assessment after randomisation and before surgery. LARVH was performed as described by Dargent et al.7 and RAH as described by Lopes et al.8 All women received intraoperative broad-spectrum antibiotic coverage for the duration of their surgery. Deep vein thrombosis prophylaxis was via perioperative subcutaneous low-molecular-weight heparin and intraoperative pneumatic calf compression garments. All women received morphine (1 mg/ml) via a patient-controlled analgesia system for the first 24 hours postoperatively and oral analgesia thereafter.

Outcome measures included the following: need for urinary catheterisation (continuous or intermittent) in days, operating time in minutes, estimated blood loss, duration of hospital stay, opiate pain relief requirements in first 36 hours after surgery, intraoperative complications, nodal yield, resection size and postoperative complications. Additional bladder and bowel function outcomes were also assessed, but these are not discussed in this paper.

Blood loss was measured as the sum of suctioned fluids and weighed sponges and drapes at completion of surgery. The upper abdomen was suctioned at LARVH. Operative duration was measured from the time of skin incision to the closure of the skin or vagina. Intraoperative complications were defined as cardiopulmonary events, cystotomy, enterotomy, ureteric injury and injury of a great vessel or nerve. Postoperative complications were defined as cardiac events (cardiac failure, cardiac arrest), deep vein thrombosis, pulmonary embolism, haematoma (radiologically confirmed), urinary tract infection, chest infection and wound infection (microbiologically confirmed).

A transurethral (Foley) catheter was inserted at the time of surgery. This remained on free drainage until day five after RAH, at which time it was withdrawn. For logistical reasons, catheters were withdrawn on day three for the LARVH group because they were generally ready for discharge by this stage.

Frequency/volumes and residual voiding volumes were measured after removal of the catheter. If residual volumes were elevated over 100 ml, the participant was offered either further continuous bladder drainage with fortnightly reassessment or to learn intermittent self-catheterisation if she preferred.9 This continued until residual voiding volumes were satisfactory (minimum 200 ml of urine voided, at least 4-hourly during the day, with residual urine volumes of less than 100 ml).

Fresh measurements of specimens were taken before fixation by the reporting pathologist. They did not have knowledge of the surgical technique employed. The average vaginal cuff length from all four quadrants (measuring from the anterior, posterior, left and right fornices to the edge of the surgical specimen when taut), the average measurement of the two cardinal ligament segments resected (measuring from the left and right cervical edges to the respective lateral edge of the surgical specimen when taut) and the average measurement of the two uterosacral ligament segments resected (measuring from the left and right posterosuperior vaginal edges to the respective lateral edge of the surgical specimen when taut).

All analyses were performed by intention-to-treat. Statistical analysis was performed using nonparametric tests and were processed using SPSS software (version 10; SPSS Inc., Chicago, IL, USA). The study was approved by the Gateshead Local Research Ethics Committee and funded by the Department of Education, Research & Development, Gateshead Health NHS Foundation Trust, Gateshead, UK.

Results

Fifteen women were recruited to the study over a 20-month period. The trial was stopped at the end of the funding period because extension would not increase the precision of the estimated differences. During the study period, 50 women with FIGO stage IB1 cervical cancer were registered at our institution. Twenty were not eligible for the trial. Reasons for not meeting the inclusion criteria were, 1. tumour greater than 2 cm in diameter (13 women), 2. unsuitable for a laparovaginal procedure (seven women). This group included women with limited vaginal access (four women), cervical defects caused by previous LLETZ biopsies (two women), and one woman who had undergone previous major abdominal surgery in the form of a liver transplant. A further 11 women requested and were considered suitable for fertility-sparing treatment. The trial was discussed with the remaining 19 eligible participants, four of whom declined entry. The remaining 15 women were recruited and randomised as previously described (Figure.1).

Figure 1.

 CONSORT diagram.

There were two exclusions after randomisation. One participant in the LARVH arm declined her allocation, exited the trial and defaulted to standard treatment. One woman in the RAH arm had her surgery abandoned before hysterectomy could be performed because a synchronous bowel tumour with multiple liver metastases was discovered intraoperatively.

In the group assigned to laparoscopic surgery, there was one intraoperative conversion to the open procedure (14%) because of an enterotomy. After examination under anaesthetic, another woman was found to have a cervical defect (because of previous conisation) potentially compromising the vaginal part of the procedure and the decision for open surgery was made.

Characteristics of the participating woman are summarised in Table 1. The two study groups were well balanced for a range of characteristics including age, parity and body mass index.

Table 1.   Characteristic of participants
 LARVH (n = 7)RAH (n = 6)P
  1. IQR, interquartile range; LARVH, laparoscopically assisted radical vaginal hysterectomy; LLETZ, large loop excision of transformation zone; LVSI, lymph–vascular space involvement; RAH, radical abdominal hysterectomy.

Age, median (IQR)38.5 (33.5–53.5)37 (29.5–46)NS
Body mass index, mean (SD)24.8 (1.3)25.0 (1.8)NS
Parity, median (IQR)2.0 (1.0–3.0)2.0 (1.0–2.0)NS
Smoker32NS
Diagnostic LLETZ66NS
Adenocarcinoma11NS
Squamous cell carcinoma65NS
LVSI11NS
Lateral tumour size (cm), mean (SD)1.2 (0.3)1.4 (0.2)NS
Depth of invasion (mm), mean (SD)5.6 (1.2)5.8 (1.1)NS

Statistically significant differences (P < 0.05) were found for the following when LARVH was compared with RAH: the median duration of catheterisation, median operating time, median blood loss, median hospital stay and median opiate requirement in the first 36 hours postoperatively. These findings are shown in Table 2.

Table 2.   Surgical data
 LARVH (n = 7)RAH (n = 6)P
  1. IQR, interquartile range; LARVH, laparoscopically assisted radical vaginal hysterectomy; RAH, radical abdominal hysterectomy.

Duration of catheterisation in days, median, IQR4 (3–9)21 (5–28)0.003
Operation length (min), median (IQR)180 (171–220)138 (120–137.5)0.05
Hospital stay (days), median (range)5 (4–7)7 (6–7)0.04
Blood loss (ml), median (IQR)400 (325–1050)1000 (800–1025)0.05
Opiate analgesic requirement (ml morphine), median (range)30 (20–46)53 (43–210)0.004
Chemoradiotherapy21NS
Nodal yield, median (range)14 (12–19)12 (11–14)NS
Time to normal activities (days), median (IQR)35 (28–49)35 (30–42)NS

In terms of nodal yield, number of women requiring radiotherapy and number of complete tumour excisions, there were no statistically significant differences (Table 2).

There were two intraoperative complications in the LARVH group—one enterotomy requiring conversion to the open procedure and one cystotomy, which was repaired vaginally. Neither woman suffered any long-term sequelae. Postoperative complications encountered were minor (Table 3). One woman was readmitted to hospital after 32 days with severe constipation after RAH.

Table 3.   Perioperative complications
 LARVH (n = 7)RAH (n = 6)
  1. LARVH, laparoscopically assisted radical vaginal hysterectomy; RAH, radical abdominal hysterectomy.

Urinary tract infection03
Wound infection11
Transient leg paraesthesia10
Cystotomy10
Enterotomy10
Readmission within 60 days01

Two women in the laparoscopic group required chemoradiotherapy—both had pelvic nodal metastases and one had additional margin involvement. One woman in the RAH group had positive pelvic nodes requiring chemoradiotherapy. All the women undergoing RAH had clear margins.

Measurements from freshly resected paravaginal and parametrial tissue showed that the mean length of the vaginal cuff and supporting ligaments removed at LARVH was significantly less than for RAH, P < 0.05. (Table 4).

Table 4.   Comparison of measurement of resected paravaginal and parametrial tissue
 LARVHRAHP
  1. LARVH, laparoscopically assisted radical vaginal hysterectomy; RAH, radical abdominal hysterectomy.

Vaginal cuff (cm), mean (SD)1.26 (0.25)2.16 (0.66)0.014
Cardinal ligament (cm), mean (SD)1.30 (0.68)2.79 (1.0)0.013
Uterosacral ligament (cm), mean (SD)1.47 (0.68)4.68 (1.25)0.034

Discussion

To our knowledge, this is the first attempt at a randomised controlled trial to evaluate short-term surgical outcomes after LARVH. It confirms the previously suggested benefits of laparoscopic surgery in terms of a reduced duration of bladder catheterisation, reduced blood loss, reduced analgesic requirement and shorter hospital stay as described in our preliminary matched–controlled study.2

In addition, pathological assessment of the specimens has shown that LARVH is a less radical procedure compared with RAH, and this may explain some of the differences in surgical outcomes. To date, although several authors have claimed a similar extent of radicality for total laparoscopic radical hysterectomy10 and extended laparoscopic radical hysterectomy11 compared with the open procedure, there are no such data for a ‘standard’ LARVH, and its radicality has been questioned.12 Our data support the suggestion that it is a less radical procedure.

Although outside the scope of this trial, there are no current data to suggest that this reduction in radicality compromises survival, so long as it is restricted to small tumours. Our matched–controlled study2 showed a higher rate of positive margins after LARVH compared with RAH (10% versus 2%), and raised suspicions that this was related to a less radical resection. Hence, we advised caution with LARVH, together with restriction to tumours less than 2 cm in size. Despite this restriction, the occurrence of ‘cut-through’ procedures and close margins must be highlighted in this trial. One woman in the laparoscopic group was found to have tumour involvement at the left and right parametrial resection margins despite a preoperative magnetic resonance imaging scan suggesting small volume disease and negative pelvic lymph nodes. Clearance at the anterior vaginal margin was just 3 mm. She also had positive pelvic lymph nodes and therefore required dual therapy with adjuvant chemoradiation regardless of the margins. One other woman who underwent LARVH had a left parametrial margin of just 3 mm, but she also had positive pelvic lymph nodes and required chemoradiotherapy. On the contrary, there were no cases of close or incomplete excision margins in the RAH group.

Complication rates for LARVH and RAH appeared similar, but those undergoing LARVH were more likely to have complications related to difficulty with the operation itself. The prolonged operating time and the conversion rate for the procedure also merit consideration.

Developing the skills to perform LARVH involves a substantial amount of time and effort for the surgeon and requires a high degree of skill in terms of both laparoscopic and vaginal surgery. Even once the technique is mastered, the complication rate does not appear to reduce2,3 and technical factors such as nulliparity, lack of uterine descent and the bony structure of the pelvis continue to make vaginal surgery challenging.13 Previous case series have reported particular difficulty with the radical vaginal part of LARVH,3 and lower urinary tract injuries are especially common.1–4 It has been suggested that total laparoscopic radical hysterectomy may avoid some of the difficulties associated with the vaginal part of LARVH.

The results of this study have significant implications for the procedure of RVT. It should be considered to be a less radical procedure than LARVH in that the uterine vessels are not transected at origin and there is less resection of the parametrial tissues. By contrast, radical abdominal trachelectomy (RAT) has been shown to provide a wider parametrial resection, including contiguous parametrial nodes compared with the RVT.14 Based on current information on radicality and surgical margins, our centre’s current practice in relation to fertility-sparing surgery is to offer loop cone plus laparoscopic bilateral pelvic lymph node dissection for small volume stage IB1 disease15 and to reserve RAT for larger tumours, making the RVT procedure obsolete. Possible future developments include the radical laparoscopic trachelectomy, which may provide radicality and surgical margins similar to RAT.

In conclusion, this study confirms the short-term surgical benefits of LARVH. In addition, LARVH has been shown to be a less radical procedure than RAH supporting the need for strict patient selection and to restrict the procedure to small tumours.

Disclosure of interests

The authors confirm there is no conflict of interest.

Contribution to authorship

R.N. devised the study and developed the protocol, obtained funding, Research & Development approval and ethics approval, supervised the study and contributed to the manuscript. K.S.J. contributed to the protocol, performed the study, analysed the results and contributed to the manuscript. A.L. contributed to the study design and the manuscript. P.A.C. and J.A.H. performed the pathological measurements.

Details of ethics approval

Ethics approval was obtained from Gateshead & South Tyneside LREC on the 4 November 2004 (Ref: 06-01). The trial was registered with the Education, Research & Development Department at Gateshead Health NHS Foundation Trust in November 2004 (Ref: 2005/GON/01). Clinical Trial Registration was not required as this was a randomised phase II trial.

Funding

Funding for the study was obtained from the Education, Research & Development Department at Gateshead Health NHS Foundation Trust.

Journal club

Discussion points

1. Background: How common is advanced cervical cancer in your country? How often are radical abdominal hysterectomies performed in your unit?

2. Methods—clinical trials: The International Committee of Medical Journal Editors (ICMJE) mandates prospective registration for any ‘research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes’ (Laine et al.,N Engl J Med 2007; 356:2734–6). Why is this mandatory for all completed trials? Why is this mandatory and useful for trials in progress? Do you know if this requirement includes exploratory and pilot studies?

Do all clinical trials need approval by ethics committees or Institutional Review Boards? What about pilot studies?

The Consolidated Standards for Reporting of Trials (CONSORT) statement is one of several guidelines to help authors improve the reporting of their studies, and is specific to randomised clinical trials (Moher et al.,Lancet 2001;357:1191–4). Can you identify the relevant sections in the manuscript that cover essential items of the CONSORT checklist? What is the benefit of covering all these items? Are you aware of any other reporting guidelines? Could these guidelines be helpful for any other purposes?

3. Descriptive data: Table 1 shows that there was not a clinical or statistical difference (P > 0.05) in the mean depth of invasion between the two arms of the study. Is this reassuring? Would your answer be different if the mean depth differed by 100% but the P value was still >0.05 because of the small sample? Would your answer be different if the mean lateral size was 6.5 mm in one arm and 7.5 mm in the other [consider International Federation of Gynecology and Obstetrics (FIGO) staging]? Are the mean and standard deviation the best statistics to present continuous variables in small samples?

4. Results: Summarise the results and the implications for clinical practice. Were there any disadvantages of the laparoscopic approach? Do you think the two approaches might differ in terms of cost-effectiveness or patient satisfaction?

5. Implications for research: One of the main objectives of small trials is to inform power and sample size calculations for larger, often multicentre, studies. Which factors could introduce bias to a multicentre study?

D Siassakos

Southmead Hospital, Bristol, UK

Email jsiasakos@gmail.com

Ancillary