Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

DOI: 10.1111/j.1471-0528.2011.03170.x www.bjog.

org

Urogynaecology

The effects of vault drainage on postoperative morbidity after vaginal hysterectomy for benign gynaecological disease: a randomised controlled trial
A Dua,a A Galimberti,a M Subramaniam,b G Popli,c S Radleya
Royal Hallamshire Hospital, Shefeld Teaching Hospitals NHS Trust, Shefeld b Royal United Hospitals, Bath, UK c Department of Economics, University of Shefeld, Shefeld Correspondence: Dr A Dua, Royal Hallamshire Hospital, Shefeld Teaching Hospitals NHS Trust, Shefeld S10 2SF, UK. Email anupreet.dua@sth.nhs.uk Accepted 1 September 2011. Published Online 18 October 2011.
a

Objective To evaluate the efcacy of vault drainage in reducing

Main outcome measures The primary outcome measure was

the immediate postoperative morbidity associated with vaginal hysterectomy carried out for benign gynaecological conditions.
Design Randomised controlled trial. Setting A tertiary referral gynaecology centre in UK. Population A total of 272 women who underwent vaginal

reduction in postoperative febrile morbidity. Secondary outcome measures were hospital readmission rate, blood transfusion, change in postoperative haemoglobin and length of stay.
Results In all, 135 women were randomised to have a drain and 137 to no drain. There were no differences in the incidence of febrile morbidity, length of stay, change in haemoglobin or need for postoperative blood transfusion between the two groups. Conclusions The routine use of vault drain at vaginal

hysterectomy for benign conditions between March 2005 and June 2010.
Methods The 272 women were randomised to have a drain

inserted or not inserted, drain or no drain, respectively, before vault closure during vaginal hysterectomy, using a sealed envelope technique. The surgical procedures were performed using the surgeons standard technique and postoperative care was delivered according to the units protocol.

hysterectomy for benign disorders has no signicant effect on postoperative morbidity. The use of vault drain in this context is not recommended.
Keywords Drains, febrile morbidity, vaginal hysterectomy, vault

haematoma.

Please cite this paper as: Dua A, Galimberti A, Subramaniam M, Popli G, Radley S. The effects of vault drainage on postoperative morbidity after vaginal hysterectomy for benign gynaecological disease: a randomised controlled trial. BJOG 2012;119:348353.

Introduction
Research into the morbidity associated with hysterectomy to date, has predominantly focused on outcomes of the vaginal approach compared with abdominal and or laparoscopic approaches. Vaginal hysterectomy (VH) is associated with lower morbidity, faster recovery and shorter stay compared with abdominal hysterectomy. A recent systematic review of randomised trials concluded that, where possible, VH should be performed in preference to abdominal hysterectomy for benign gynaecological disease.1 Even though VH has lower risks compared with abdominal hysterectomy, it is associated with a signicant risk of vault hae-

matomas (2559%)24. A haematoma represents the most common perioperative complication following VH and is signicantly associated with febrile morbidity, postoperative haemoglobin drop, need for blood transfusion, readmission to hospital and length of hospital stay.2 The focus of research into morbidity related to VH has most commonly concerned the use of antibiotic prophylaxis or the postoperative diagnosis and management of haematoma. The use of prophylactic antibiotics has been clearly shown to have a signicant role in the reduction of infection, and is now standard practice in the UK.5 More recently, the use of a bipolar vessel sealing system (Ligasure, Autosuture, Valleylab, Boulder, CO, USA) during VH

348

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

Role of routine vault drainage at vaginal hysterectomy

was shown to reduce the incidence of perioperative haemorrhagic complications.6 The use of such a measure, however, depends on equipment availability and the surgeons willingness to change technique and is therefore unlikely to become standard practice in the near future. Other prophylactic measures such as different techniques for vaginal cuff closure and preoperative vaginal cleansing have been described but have not been shown to be effective in reducing postoperative morbidity.7,8 Measures taken to reduce haematoma formation such as vault drainage may help to reduce the postoperative complications and morbidity. The role of drains in abdominal surgery is well recognised. However, no formal evaluation of routine drain insertion at VH has been performed. The aim of our study was to evaluate the efcacy of vault drainage in reducing the risk of early postoperative morbidity following VH for benign gynaecological conditions.

Methods
This was a prospective randomised study comparing the use of vault drain with no drain at vaginal hysterectomy. Sample size calculation, was based on an incidence of febrile morbidity of 30% as described in a review article9 and including data from six studies relating to morbidity following VH. To demonstrate a 50% reduction in the treatment group, 135 women were required in each arm of the study (a = 0.05, b = 0.80 1:1 randomisation ratio). The study was approved by the South Shefeld Research Ethics Committee and was undertaken in a tertiary referral gynaecology unit between March 2005 and July 2010. All women undergoing VH (with or without vaginal prolapse repair or oophorectomy) for benign gynaecological disease were eligible for inclusion in the study. We excluded women in whom hysterectomy was performed for malignant disease and also women for whom it was felt by the operating surgeon that the insertion of a surgical drain to the vault would be clinically indicated. The primary outcome measure for the study was reduction in postoperative febrile morbidity. There is no widely accepted standard denition of febrile morbidity. We chose a level of 37.5C, which has been used previously in a number of studies assessing postoperative infectious morbidity.10,11 Temperature was measured 4-hourly on the postoperative ward and recorded in the nursing observation charts. Secondary outcome measures were hospital readmission rate, return to theatre, blood transfusion, change in postoperative haemoglobin and length of hospital stay. Randomisation (drain or no drain) was carried out using a sealed envelope technique. Randomisation envelopes were prepared at the beginning of the study. Envelopes were double sealed to prevent bias. The

randomisation envelope was opened at the time of closure of the vaginal vault. Women were recruited during their visit to the preoperative assessment clinic, usually 24 weeks before surgery. During that visit, women were counselled and written information was provided. If they agreed to participate, a signed written consent was obtained and a randomisation envelope was included in the medical records (simple randomisation). The hysterectomy was performed by each surgeon according to each surgeons standard technique. All women had prophylactic antibiotics and thromboprophylaxis according to standard hospital protocol. Depending on the randomisation either a non-suction surgical drain (size 16 Robinson) was inserted in the vault or the surgeons standard closure was performed with no drain. All surgeons who participated in this trial practice non-closure of the peritoneum with interrupted sutures to the vaginal vault. The use of vaginal pack and urinary catheter was left to the discretion of the operating surgeon. Postoperative monitoring and care were according to standard hospital policy. Pack and drain were removed at the surgeons request. Haemoglobin was measured routinely on the second postoperative day. Women were discharged home when deemed clinically t by the operating surgeon. Case notes were reviewed 3 months postoperatively to collect information regarding recovery and readmission to hospital. Data were anonymised and entered to an excel spreadsheet. Statistical analysis was performed using spss (SPSS version 15; IBM, Chicago, IL, USA). Data were analysed on an intention-to-treat basis. Categorical variables were analysed by Pearsons chi-square test and parametric data were analysed by Students t test. A P value of <0.05 was considered signicant.

Results
A total of 318 women were approached of whom 12 declined to participate in the study. In all, 306 women were recruited between March 2005 and July 2010. Of these, 34 were excluded (either the scheduled operation was cancelled or a VH was not performed as part of the procedure). The remaining 272 women were included in the nal analysis, of whom 137 were randomised to no drain and 135 to drain insertion (Figure 1). Overall, in 79.8% of women, the primary indication of hysterectomy was prolapse. The indications for hysterectomy in the two groups are listed in Table 1. In women with more than one indication for hysterectomy, the primary indication recorded on the surgical notes has been listed. In six of the 135 women in whom a drain was inserted, the drain fell out in the recovery area or during transfer from recovery to the ward. In three women randomised to

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

349

Dua et al.

Table 1. Indications for hysterectomy Indication Drain (n = 135), n (%) 17 (12.6) 3 (2.2) 2 113 11 65 31 6 (1.5) (83.7) (8.1) (48.1) (23) (4.5) No drain (n = 137), n (%) 25 (18.2) 3 (2.2) 5 104 16 61 23 4 (3.6) (76) (11.7) (44.5) (16.8) (3)

Heavy periods Post-menopausal bleeding Pelvic pain Prolapse Stage 1* Stage 2* Stage 3* Stage 4*

*Pelvic Organ Prolapse Quantication System.

no drain, a drain was inserted by the operating surgeon because of excessive oozing from the vault. As data were analysed on an intention-to-treat basis, these women were still included in their original randomisation group. The mean age of women with no drain was 56.3 years and of women with drain was 57 years. There was no difference in mean body mass index, parity and estimated blood loss during surgery in the two groups. The number of women with signicant co-morbidity was also similar (Table 2). All women included in the study had a catheter inserted after hysterectomy. Seven women in the drain group and four in the no drain group did not have a vaginal pack postoperatively. A total of 67 (24.6%) women had a temperature of 37.5C during the postoperative period of whom 33 (24.4%) were in the drain group and 34 (24.5%) were in the no drain group (chi-square test P = 1.0). The maximum temperature in the drain group was 38.4C and in the no drain group was 38.5C. There was no difference in the mean temperature of women who had a temperature of 37.5C. A subgroup analysis was also performed for women who had a temperature of 38C and no difference was noted in the number of women and the mean temperature in the two groups. Table 3 shows the results of febrile morbidity in the two groups.
Table 2. Demographics of the two groups Demographic No drain (n = 137) 56.3 27.0 2.4 227.9 60 (56.5) (27) (2) (200) (43.8) Drain (n = 135) 57 28.4 2.5 245 57 (59) (28) (2) (200) (42.2) P value

The mean postoperative haemoglobin dropped by 1.79 g/dl in the drain group, and by 1.74 g/dl in the no drain group (Students t test: P > 0.05). Similarly, the length of stay and incidence of blood transfusion was similar in both groups (Table 4). Seven women in the drain group and ve with no drain had culture proven urinary tract infection during the postoperative period. Two women in the drain group returned to theatre in contrast to the no drain group where none of the women returned to theatre. However, the reasons for return to theatre were not related to vault haematoma. One of these women had ureteric trauma and the other had excessive oozing from the vagina following anterior colporrhaphy. Three women (two in the no drain group and one in the drain group) were readmitted to hospital, all with vaginal infection, which was managed conservatively with antibiotics.

Discussion
This study suggests that the routine use of vault drainage at VH does not inuence immediate postoperative morbidity. This is the rst randomised controlled trial to assess the effect of routine vault drainage during VH on morbidity. Previous studies in the literature have compared the use of prophylactic antibiotics with vault drainage at vaginal hysterectomy. Wijma et al.12 compared suction drainage with perioperative antibiotics in preventing postoperative infections and found signicantly more vaginal cuff abscesses and febrile morbidity in the drain group. However, in their study population, women who had a drain inserted did not receive prophylactic antibiotics so it is difcult to ascertain whether the excess febrile morbidity was attributable to drain use or absence of antibiotic prophylaxis. Similar results were observed by Galle et al.13 and Poulsen et al.14 in studies comparing suction drainage with prophylactic antibiotics. In contrast, Swartz and Tanaree15 in an observational study showed signicantly reduced infection rates with the use of drains compared with a control group and advocated vault drainage instead of antibiotics. They concluded that suction drainage was as effective as prophylactic antibiotics even though no women in this study received any antibiotics. As the use of prophylactic antibiotics is now standard practice, the efcacy of vault drainage warrants further investigation. Shen et al.16 looked at the efcacy of drains after laparoscopic assisted VH in reducing postoperative morbidity. They concluded that prophylactic drainage is not necessary because there was no difference in postoperative infectious morbidity or complications. These results however, may not be generalised to VH, as laparoscopic assisted VH surgeons have the benet of visualising persis-

Mean age (median) Mean BMI (median) Mean parity (median) Mean EBL (ml) (median) Co-morbidities, n (%)

0.64 0.13 0.81 0.32 0.86

BMI, body mass index; EBL, estimated blood loss.

350

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

Role of routine vault drainage at vaginal hysterectomy

Table 3. Febrile morbidity in the two groups Temperature n (%) 37.5C 38.0C *Two-tailed Students t test. 34 (24.5) 15 (11) No drain (n = 137) Mean (SD) 37.8 (0.30) 38.1 (0.14) n (%) 33 (24.4) 17 (12.6) Drain (n = 135) Mean (SD) 37.9 (0.25) 38.1 (0.15) 0.45 0.55 P value*

Table 4. Outcomes in the two groups Outcome No drain (n = 137) )1.74 (1.4) 3.43 (0.95) 2 (1.5) 6 (4.4) 2 (1.5) Drain (n = 135) )1.79 (2.1) 3.44 (1.52) 1 (0.7) 5 (3.7) 0 P value

Haemoglobin change, mean (SD) Length of stay (nights), mean (SD) Readmission to hospital, n (%) Blood transfusion, n (%) Return to theatre, n (%)

0.84* 0.93* 0.57** 0.78**

*Two-tailed Students t test. **Pearson chi-square test.

tent bleeding from the vault after closure and achieve haemostasis under direct vision. The primary outcome of our study was an objective assessment of immediate postoperative febrile morbidity rather than imaging for vault haematoma. Thomson et al.2 performed transvaginal ultrasound in 223 women following vaginal hysterectomy and found vault haematomas in 25% of women. However, only a small proportion of these women (31%) had a signicant increase in febrile morbidity. Dane et al.17 concluded that sonographic detection of uid collection is common following VH but most haematomas are small in size and do not increase the risk of febrile episodes or require additional treatment. Hence, we did not perform routine postoperative ultrasound to detect vault haematoma; instead, we looked

Enrolment

Assessed for eligibility (n = 318)

Excluded (n = 46) Not meeting inclusion criteria (n = 28) Declined to participate (n = 12) Other reasons (n = 6)

Randomised (n = 272)

Allocation
Allocated to intervention (n = 135) (drain inserted) Received allocated intervention (n = 135) Did not receive allocated intervention (n = 0) Allocated to intervention (n = 137) (no drain inserted) Received allocated intervention (n = 134) Did not receive allocated intervention (n = 3) (drain insertedin view of continued oozing)

Followup
Lost to follow up (n = 0) Discontinued intervention (n = 6) (drain fell out during transfer of patient postoperatively) Lost to follow up (n = 0) Discontinued intervention (n = 0)

Analysis
Analysed (n = 135) Excluded from analysis (n = 0) Analysed (n = 137) Excluded from analysis (n = 0)

Figure 1. CONSORT 2010 ow diagram.

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

351

Dua et al.

at clinical parameters to assess morbidity with the primary outcome measure being febrile morbidity. There is no consensus on the precise denition of febrile morbidity and consequently the reported incidence of postoperative pyrexia is extremely variable (150%).2,1719 The incidence of febrile morbidity is highly dependent on the denition applied. Our sample size calculation was based on an incidence of 30% as described by Thomson and Farquharson9 in a review article. This rate represents a combined incidence of febrile morbidity ranging from 39% (for women with haematoma) to 16% (for women with no haematoma). For drain insertion to be considered an effective intervention, we hypothesised a reduction in febrile morbidity by 50% to 15% or less. This study is potentially underpowered to detect a smaller difference, which may be a limitation. However, this represents a large sample and a lower effect size would put the febrile morbidity within this range (1639%), making our ndings less credible. To specically evaluate the role of vault drainage and to avoid bias, surgeons were invited to perform surgery in a routine fashion. None of the other parameters, including postoperative care and use of packs and catheters, were altered. The study was conducted in a large tertiary referral centre and procedures were performed by ten different surgeons. As a consequence we feel that results can be generalised to centres performing VH in the UK. Most hysterectomies (80%) in our study were performed for uterine prolapse. It may be argued that the incidence of haematoma formation and morbidity is reduced in such women because of easy visualisation of pedicles. The incidence of complications may be higher in women in whom VH is performed for other conditions such as menorrhagia, broids or endometriosis, because these surgeries are more difcult to perform. Traditionally gynaecologists have selectively used vault drain at VH if there is excessive oozing from the vault during closure. Our results do not support routine use of drains; however, there were no adverse outcomes noted with use of drain. Therefore, there may still be a place for selective use of drains in such cases, which needs further evaluation.

Contribution to authorship
AD contributed to recruitment, data collection, data analysis, manuscript writing and submission; AG contributed to protocol writing, ethics approval and project supervision; MS contributed to recruitment; GP contributed to statistical support and advice, and to supervision of data analysis; and SR contributed to project supervision. AG, MS, GP and SR were all responsible for manuscript editing.

Details of ethics approval


The study was granted ethics approval by the South Shefeld Research Ethics Committee (SSREC03/083).

Funding
The study was granted 4100 by the Small Grants Scheme of Shefeld Teaching Hospitals Charitable Trust.

Acknowledgements
The authors gratefully acknowledge the help of Mr A Farkas, Miss D Fothergill and Dr S Jha for allowing their patients to take part in the study, recruiting and performing the vaginal hysterectomy. The authors would also like to thank all the women who participated in the study. j

References
1 Nieboer TE, Johnson N, Lethaby A, Tavender E, Curr E, Garry R, et al. Surgical approach to hysterectomy for benign gynaecological disease. Cochrane Database of Systematic Reviews 2009;3:CD003677. 2 Thomson AJM, Sproston AR, Farquharson RG. Ultrasound detection of vault haematoma following vaginal hysterectomy. Br J Obstet Gynaecol 1998;105:2115. 3 Wood C, Maher P, Hill D. Bleeding associated with vaginal hysterectomy. Aust NZ J Obstet Gynaecol 1997;37:45761. 4 Slavotinek J, Berman L, Burch D, Keefe B. The incidence and signicance of acute post-hysterectomy pelvic uid collections. Clin Radiol 1995;50:3226. 5 Dicker RC, Greenspan JR, Strauss LT, Cowart MR, Scally MJ, Peterson HB, et al. Complications of abdominal and vaginal hysterectomy among women of reproductive age in the United States. Am J Obstet Gynecol 1982;144:8418. 6 Hefni MA, Bhaumik J, El-Toukhy T, Kho P, Wong I, Abdel-Razik T, et al. Safety and efcacy of using the Ligasure vessel sealing system for securing the pedicles in vaginal hysterectomy: randomised controlled trial. BJOG 2005;112:32933. 7 Cruikshank SH. Methods of vaginal cuff closure during vaginal hysterectomy. South Med J 1988;81:13758. 8 Kjolhede P, Halili S, Lofgren M. Vaginal cleansing and postoperative infectious morbidity in vaginal hysterectomy. A register study from the Swedish National register for Gynecological surgery. Acta Obstet Gynecol Scand 2011;90:6371. 9 Thomson AJ, Farquharson RG. Vault haematoma and febrile morbidty after vaginal hysterectomy. Hosp Med 2000;61:5358. 10 Wen K-C, Chen Y-J, Sung P-L, Wang P-H. Comparing uterine broids treated by myomectomy through traditional laparotomy and two modied approaches: ultraminilaparotomy and laparoscopically assisted ultraminilaparotomy. Am J Obstet Gynecol 2010;202: 144e18.

Conclusion
The results from our study do not support the routine use of prophylactic drains during VH. Further comparative studies are recommended to assess the differences in morbidity after VH for prolapse and other benign conditions and use of drains in relation to indication of hysterectomy and level of difculty of the procedure.

Disclosure of interests
None.

352

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

Role of routine vault drainage at vaginal hysterectomy

11 Kolben M, Mandoki E, Ulm K, Freitag K. Randomised trial of cefotiam prophylaxis in the prevention of postoperative infectious morbidity after elective caesarean section. Eur J Clin Microbiol Infect Dis 2001;20:402. 12 Wijma J, Kauer FM, van Saene HK, van de Wiel HB, Janssens J. Antibiotics and suction drainage as prophylaxis in vaginal and abdominal hysterectomy. Obtet Gynecol 1987;70:3848. 13 Galle PC, Urban RB, Homesely HD, Jobson VW, Wheeler AS. Single dose carbenicillin versus T-tube drainage in patients undergoing vaginal hysterectomy. Surg Gynecol Obstet. 1981;153:3512. 14 Poulsen HK, Borel J, Olsen H. Prophylactic metronidazole or suction drainage in abdominal hysterectomy. Obstet Gynecol 1984;63:291 4. 15 Swartz WH, Tanaree P. Suction drainage as an alternative to prophylactic antibiotics for hysterectomy. Obstet Gynecol 1975;45:30510.

16 Shen CC, Huang FJ, Hsu TY, Weng HH, Chang HW, Young S. A prospective, randomised study of closed-suction drainage after laparoscopic assisted vaginal hysterectomy. J Am Assoc Gynecol Laparosc 2002;9:34652. 17 Dane C, Dane B, Cetin A, Yayla M. Sonographically diagnosed vault hematomas following vaginal hysterectomy and its correlation with postoperative morbidity. Infect Dis Obstet Gynecol 2007;2007: 91708. 18 David-Monteore E, Rouzier R, Chapron C, Darai E. Surgical routes and complications of hysterectomy for benign disorders: a prospective observational study in French university hospitals. Hum Reprod 2007;22:2605. 19 Makinen J, Johansson J, Tomas C, Tomas E, Heinonen PK, Laatikainen T, et al. Morbidity of 10110 hysterectomies by type of approach. Hum Reprod 2001;16:14738.

2011 The Authors BJOG An International Journal of Obstetrics and Gynaecology 2011 RCOG

353

You might also like