The impact of mode of anaesthesia on postoperative recovery from fast-track abdominal hysterectomy: a randomised clinical trial


Dr N Borendal Wodlin, Department of Obstetrics and Gynaecology, University Hospital, S-581 85 Linköping, Sweden. Email


Please cite this paper as: Borendal Wodlin N, Nilsson L, Kjølhede P, for the GASPI study group. The impact of mode of anaesthesia on postoperative recovery from fast-track abdominal hysterectomy: a randomised clinical trial. BJOG 2011;118:299–308.

Objective  To determine whether the duration of hospital stay after abdominal hysterectomy in a fast-track setting differed between women operated under general anaesthesia or in spinal anaesthesia with intrathecal morphine.

Design  An open randomised controlled multicentre study.

Setting  Five hospitals in the south-east of Sweden.

Population  One hundred and eighty women scheduled for benign hysterectomy were randomised: 162 completed the study, 82 were allocated to spinal anaesthesia and 80 were allocated to general anaesthesia.

Methods  Fast-track model comprising no use of sedatives for premedication, pre-emptive anti-emetic therapy, intravenous fluid restriction, analgesics based on non-opioids, early enteral nutrition and mobilisation, and standard criteria for discharge. Spinal anaesthesia with 20 mg hyperbaric bupivacaine and 0.2 mg morphine. General anaesthesia with propofol, fentanyl and rocuronium, and with continuous propofol and ventilation with oxygen-in-air for maintenance of anaesthesia.

Main outcome measures  Hospital stay, consumption of analgesics, vomiting, pruritus and bowel function recovery.

Results  Median hospitalisation did not differ significantly between women who had hysterectomy with spinal or general anaesthesia (46 and 50 hours, respectively). Spinal anaesthesia was associated with a significantly lower use of opioids and a faster recovery of bowel function, although vomiting and pruritus were more prevalent.

Conclusions  In a fast-track model the duration of hospitalisation after abdominal hysterectomy was <50 hours, independent of the mode of anaesthesia. Spinal anaesthesia reduced the need for postoperative morphine compared with general anaesthesia. In order to improve patient recovery after gynaecological surgery further studies based on fast-track programmes are needed.


Hysterectomy is a common gynaecological operation. It is usually performed through laparotomy under general anaesthesia as an in-hospital procedure, with the duration of hospitalisation between 2 and 11 days.1 Medical factors such as postoperative pain, nausea and paralytic ileus may affect the length of the hospital stay. In addition, time to discharge may vary because of different medical and local traditions, as well as geographic and social differences. The concept of ‘fast-track surgery’ has developed with the aim of reducing recovery times, reflected in reduced hospitalisation.2 The fast-track concept comprises providing the patient with preoperative education concerning pre-, peri- and postoperative care, optimising anaesthesia and the management of pain and nausea, including the minimal use of opioids, early postoperative mobilisation and enteral nutrition, with perioperative intravenous fluid restriction.3–5

Optimising anaesthesia and pain management in order to reduce the surgical stress response facilitates postoperative recovery.2 The use of perioperative regional anaesthesia during elective surgery, with or without concomitant general anaesthesia, has been shown to reduce discomfort and the need for opioids postoperatively.6,7 Intrathecally applied opioids in patients undergoing urologic, orthopaedic, haemorrhoid, gynaecologic or obstetric surgery with spinal anaesthesia can further optimise postoperative pain management.8,9

Today there are no published randomised prospective studies with the focus on hospitalisation after gynaecological surgery in a fast-track setting.

The aim of this prospective, randomised multicentre study was to determine whether the duration of hospital stay after benign abdominal hysterectomy in a fast-track setting differed between those carried out under general anaesthesia and those performed under spinal anaesthesia with intrathecal morphine.


An open, prospective, randomised, controlled multicentre study comparing two different anaesthetic techniques in fast-track surgery of women undergoing abdominal hysterectomy for benign gynaecological diseases was undertaken. Ethical approval was obtained from the Regional Ethical Board at Linköping University. The departments of Obstetrics and Gynaecology at five hospitals in the south-east health region of Sweden participated in the study.


Women who were admitted to the participating units for benign hysterectomy between March 2007 and June 2009 were asked to participate. Medical inclusion criteria were women between 18 and 60 years of age, who had been admitted for abdominal subtotal or total hysterectomy because of benign gynaecological disorders. At least one ovary was to be preserved in the operation. The women had to speak Swedish fluently and understand it equally well. Exclusion criteria were an American Society of Anesthesiologists (ASA) score of ≥3, former or planned concomitant bilateral oophorectomy, postmenopausal women without hormone therapy (HT), gynaecological malignancy (cervical dysplasia not included), morphine allergy, physically disabled, severe psychiatric or mental disorder and any condition that would exclude the women from undergoing regional anaesthesia in the standardised manner of the study. After having given oral and written informed consent the included patients were randomised to receive either general anaesthesia or spinal anaesthesia including intrathecal morphine.


A computer generated the randomisation sequences into blocks of ten, with an equal number of the two modes of anaesthesia for each of the five participating centres. The participating centres were assigned slightly different numbers of blocks corresponding to the expected number of eligible patients at the hospital. The allocated mode of anaesthesia, written on a label, was sealed in opaque consecutively numbered envelopes. At each centre the envelopes were opened in consecutive number order of patient inclusion in the study. The flow chart of the study population is shown in Figure 1. One hundred and eighty women were randomised and 162 completed the study.

Figure 1.

 Flowchart of the study participants.


One hour before surgery the patient received 2 g of oral paracetamol. No premedication sedatives were given. The patient was allowed to drink clear fluids until 2 hours before surgery. Antibiotic and antithrombotic prophylaxes were administered according to department routine.

General anaesthesia

General anaesthesia (GA) was induced with propofol and fentanyl. Tracheal intubation was facilitated by rocuronium. The anaesthesia was maintained with a continuous intravenous infusion of propofol and oxygen in air. Rocuronium and fentanyl were repeated during anaesthesia according to the attending anaesthetist. Twenty minutes before ending the operation, 5 mg of morphine was given intravenously. All patients had orogastric tubes inserted during surgery.

Spinal anaesthesia

Spinal anaesthesia (SA) was administered with a 25-gauge needle, preferably in the L3/L4 or L2/L3 intervertebral spaces. The anaesthetic consisted of 20 mg of hyperbaric bupivacaine (5 mg/ml) and 0.2 mg of morphine (0.4 mg/ml). Fifteen minutes after administration of anaesthesia the level of the neural blockade was determined with a cold test and registered. Sedation was applied throughout the operation with a continuous intravenous infusion of propofol. If the spinal anaesthesia was insufficient the patient received GA according to the protocol described above.


The decisions about mode of abdominal hysterectomy and skin incision were made prior to randomisation. The surgical technique used in the hysterectomy was left to the surgeon’s discretion. It was permissible for the surgeon to plicate the sacrouterine and cardinal ligaments, but not to anchor the round ligaments to the vaginal cuff or the cervical stump; peritonealisation was not to be carried out. The surgeon injected 40 ml of bupivacaine (2.5 mg/ml) subcutaneously and pre-fascially in the abdominal wall wound before ending the surgery. A transurethral urinary bladder catheter was inserted preoperatively and left until the next morning.

Perioperative parenteral fluid infusions

A restricted and similar regimen of intravenous fluids was used perioperatively in both groups. At the start of anaesthesia 5 ml/kg of lactated Ringer’s solution was administered. A continuous infusion of fenylephrine (40 μg/ml) was used if the mean arterial blood pressure decreased by more than 30% from the baseline. Parenteral fluids composed of lactated Ringer’s solution and glucose 2.5% were used, and the target for the total dose was set at 25 ml/kg per day, provided no complications had occurred perioperatively. Parenteral infusion was terminated on the morning of the day after surgery if the postoperative surveillance was uneventful. Otherwise parenteral infusion was continued on medical indications.

Postoperative care

At the conclusion of the operation the patient was transferred to the post-anaesthesia care unit (PACU) for postoperative monitoring of haemodynamic and respiratory stability, degree of sedation, pain, nausea and pruritus. The monitoring was similar after the two modes of anaesthesia. The patient was discharged from PACU to the gynaecological ward when vital signs were stable and the patient was awake. The monitoring was continued in the gynaecological ward. Provided that measures were within clinically normal limits, they were registered once every hour during a period of 12 hours, then once every third hour until 24 hours after commencing surgery. The patient was instructed to drink and eat as soon as possible after surgery. Mobilisation was initiated in the PACU and actively encouraged early in the gynaecological ward.

Postoperative pain management was obtained orally with 1330 mg paracetamol and 50 mg diclofenac three times daily during hospitalisation. Pain intensity was measured on a visual analogue scale (VAS), ranging from 0 to10, with 0 indicating no pain and 10 indicating the worst pain imaginable. The measurements were carried out in association with the monitoring of vital signs. Additional pain relief was offered if the patient requested it or if the VAS score was >3. Opioids were avoided if possible, but were given orally or intravenously if necessary. Pre-emptive anti-emetic therapy was administered using bilateral acupressure wrist bands. These were applied preoperatively and maintained throughout the hospital stay. Rescue anti-emetic treatment was available with single or repeated doses (droperidol 0.6 mg intravenously and 5-HT3 receptor antagonist, in that order) upon patient demand. Treatment of discomfort of pruritus comprised, in order of use, clemastine and naloxone (in low doses given orally or intravenously).

The criteria for discharge of the patient were standardised: the patient was mobilised, tolerated a normal diet, had sufficient pain relief with oral analgesics (VAS ≤ 4), had voided spontaneously with <150 ml residual urine (measured by a portable bladder ultrasound scan) and showed no signs of mechanical bowel obstruction. If the patient had insufficient bladder emptying upon discharge, the patient was given a transurethral bladder catheter for another couple of days. The catheter was then removed in the outpatient clinic, and the residual urine volume was subsequently controlled by means of a portable bladder ultrasound scan. The time of discharge was registered.

Time of first pass of gas and first bowel movement were registered during the hospital stay. At discharge the patient was requested to complete a diary once a day for 35 days postoperatively. The patient was instructed to record experiences of pain and nausea, recovery of bowel function, time of the first pass of gas and the first bowel movement, in case these had not occurred during the hospital stay.


Data are presented as medians and (ranges) or numbers and percentages. All analyses were performed according to intention-to-treat principles. In addition, outcome measures were analysed according to per-protocol principles, as reported in the text. In univariate analysis Mann–Whitney U-test, Yates’ corrected chi-square test and Fishers’ exact test were used, when appropriate, to compare descriptive and clinical data between the two groups. Analysis of covariance (ancova) was used to test differences in continuous outcome measures between the two groups, and logistic regression analysis was used to analyse nominal effect measures. Adjustment was simultaneously performed for mode of hysterectomy, skin incision, body mass index (BMI), smoking habits and occurrence of change in protocol of anaesthesia in the multivariate models, when appropriate. The results of the logistic regression models are presented as odds ratios and 95% confidence intervals. The level of significance was set at 5%. Statistical analyses were carried out with the software StatView® v5.0.1 (SAS Institute Inc., SAS Campus Drive, Cary, NC, USA).

Sample size calculation

With an α = 0.05 and 1 – β = 0.90, the sample size was estimated to be 180 women, including a drop-out of 10% given that the stay in hospital after benign abdominal hysterectomy in GA before commencing the study was 3.7 ± 2.0 days (mean ± 1 SD), and based on the assumption that the duration of stay in hospital after hysterectomy would be reduced by 1 day after SA compared with GA.


Eighty of the women randomised to GA and 82 of those randomised to SA completed the study, and comprise the study population. The demographic and descriptive data of the study groups are presented in Table 1. No statistically significant differences were observed between the groups in any of these data.

Table 1.   Perioperative demographic and descriptive data
CharacteristicsGeneral anaesthesia (n = 80)Spinal morphine anaesthesia (n = 82)
  1. Figures denote medians and (ranges) or numbers and (%). No statistically significant differences were observed between the groups in any of the variables (univariate analyses).

  2. *A women may have had more laparotomies.

  3. **The American Society of Anesthesiologist classification of physical status.

Age (years)45 (33–58)46 (35–58)
Body mass index (kg/m2)25.3 (19.0–41.5)25.4 (18.9–38.0)
BMI ≤ 2539 (48.8%)35 (42.7%)
BMI > 25 and <3024 (30.0%)31 (37.8%)
BMI ≥ 3017 (21.2%)16 (19.5%)
Parity2.0 (0–5)2.0 (0–8)
Smokers16 (20.0%)13 (15.9%)
Previous laparotomy29 (36.3%)29 (35.4%)
Previous anaesthesia*
General anaesthesia21 (72%)21 (70%)
Spinal/epidural anaesthesia8 (28%)9 (30%)
Concomitant diseases
Psychiatric8 (10.0%)6 (7.3%)
Musculoskeletal11 (13.8%)6 (7.3%)
Cardiovascular8 (10.0%)9 (11.0%)
Chronic pulmonary3 (3.8%)7 (8.5%)
Concomitant medication
Analgesics21 (26.3%)11 (13.4%)
Antidepressants8 (10.0%)5 (6.1%)
Indication of hysterectomy
Bleeding disturbances46 (57.5%)46 (56.1%)
Mechanical symptoms27 (33.7%)29 (35.4%)
Cervical dysplasia/endometrial hyperplasia4 (5.0%)5 (6.1%)
Endometriosis/dysmenorrhoea3 (3.8%)2 (2.4%)
Class I59 (73.7%)55 (67.1%)
Class II21 (26.3%)27 (32.9%)
Mode of hysterectomy
Total abdominal55 (68.8%)51 (62.2%)
Subtotal abdominal25 (31.2%)31 (37.8%)
Mode of skin incision
Midline6 (7.5%)7 (8.5%)
Low transverse74 (92.5%)75 (91.5%)

In eight women (9.8%) the SA gave insufficient analgesia, and the women received GA according to the study protocol. In the GA group the general anaesthesia given according to the study protocol could not maintain sufficient depth of anaesthesia in five women (6.3%), and N2O or sevoflurane was added in these cases. The upper level of sensory blockade was median to the level of Th4 (range C3–Th10) in the 74 women who had sufficient analgesia with SA. None of the patients in the SA group needed respiratory assistance because of the high level of motor blockade. Two women in the GA group experienced severe postoperative pain in the PACU that did not respond sufficiently to intravenously administered opioids. The pain was controlled with a single dose of intrathecal sufentanil in one case, and with continuous epidural analgesia using ropivacaine in the other case.

The ANCOVA model showed no significant difference in the duration of stay in hospital from the start of anaesthesia to discharge from the gynaecological ward between the two groups [GA group 50 hours (range 24–100 hours) versus SA group 46 hours (range 22–125 hours); = 0.4004]. However, the proportion of women who were discharged within 48 hours was significantly higher in the SA group than in the GA group: 56% (46/82 women) versus 39% (31/80 women) (P = 0.0401; Yates’ corrected chi-square test; df = 1). In contrast with this, the proportion of women who stayed in hospital for more than 72 hours did not differ significantly between the two groups: 11.3% (9/80) in the GA group versus 11.0% (9/82) in the SA group. The per-protocol analysis revealed similar results concerning the median duration of hospital stay, but the proportion of women who were discharged within 48 hours did not differ significantly between the modes of anaesthesia. Peri- and postoperative clinical data are presented in Table 2. The time in the PACU was significantly shorter for the SA group. The more rapid recovery of bowel function in the SA group was almost exclusively attributed to the quantity of opioids used. When adjustment for the equivalent dose of morphine and fentanyl during day 0 was included in the ancova model, there was no significant association between mode of anaesthesia and time to first bowel movement. A strong association was found between the equivalent dose of morphine and fentanyl during day 0 and the time to first bowel movement (positive correlation, P = 0.0037). The per-protocol analysis was similar in this aspect.

Table 2.   Perioperative data
CharacteristicsGeneral anaesthesia (n = 80)Spinal morphine anaesthesia (n = 82)Analysis of covariance P value*
  1. Figures denote medians and (ranges).

  2. *Analysis of covariance with simultaneous adjustments for mode of hysterectomy, skin incision, BMI, smoking habits and change of mode of anaesthesia.

  3. **Time from skin incision to skin closure.

  4. ***Time from start of anaesthesia to extubation (GA) or leaving the operating room (SA).

  5. ****Time from start of anaesthesia to the event occurring postoperatively.

Operating time (minutes)**75 (40–225)70 (40–173)0.1478
Estimated peroperative bleeding volume (ml)150 (10–3000)100 (10–1100)0.3340
Uterus weight (g)298 (41–1655)302 (58–2784)0.2205
Time of anaesthesia (minutes)***124 (86–266)115 (60–210)0.1651
Time in PACU (hours)4.3 (1.8–23.0)3.6 (1.5–14.2)0.0328
Time with urinary catheter (hours)****22 (7–145)22 (9–236)0.8770
Time to first pass of gas (hours)****29 (10–100)27 (5–90)0.0187
Time to first bowel movement (hours)****57 (22–124)50 (11–153)0.0394

The peri- and postoperative use of analgesics is shown in Table 3. The consumption of opioids on day 0 and opioids and non-opioid analgesics on day 1 was significantly lower in the SA group than in the GA group. After day 2 the consumption of opioids was very low in both groups. The data are not shown for days 4 and 5. These results also held true with the per-protocol analysis, except that the consumption of non-opioid analgesics on day 1 did not differ between the two groups.

Table 3.   Perioperative anaesthetic drugs and postoperative analgesics during hospital stay
CharacteristicsGeneral anaesthesiaSpinal morphine anaesthesiaAnalysis of covariance*P value
  1. Figures denote medians and (ranges); RDD, sum of recommended daily dosage.

  2. *Analysis of covariance with adjustments for mode of hysterectomy, skin incision, BMI, smoking habits and change of mode of anaesthesia.

  3. **The dosage given perioperatively is not included.

  4. ***Opioids given postoperatively have been converted into the equivalent intravenous dose of morphine.

Perioperativelyn = 80n = 82 
Propofol (mg)1150 (101–3050)384 (54–1494)<0.0001
Fentanyl (mg)0.3 (0.0–0.7)0.0 (0.0–0.5)<0.0001
Fenylephrine (μg)0 (0–1500)625 (0–3400)<0.0001
Morphine (mg)5.0 (0–10)0 (0–7.5)<0.0001
Day 0 (i.e. day of surgery)n = 80n = 82 
Equivalent morphine dose (mg)**,***18.5 (5.0–49.5)0.4 (0.0–35.8)<0.0001
Non-opioid analgesics (RDD)2.1 (1.1–2.8)1.8 (1.2–2.8)0.8074
Day 1n = 80n = 82 
Equivalent morphine dose (mg)0.0 (0.0–42.7)0.0 (0.0–40.8)0.0462 
Non-opioid analgesics (RDD)2.0 (0.0–2.7)1.2 (0.0–2.7)0.0400
Day 2n = 56n = 47 
Equivalent morphine dose (mg)0.0 (0.0–15.0)0.0 (0.0–26.7)0.6913
Non-opioid analgesics (RDD)0.0 (0.0–2.0)0.0 (0.0–2.3)0.6407
Day 3n = 12n = 9 
Equivalent morphine dose (mg)0.0 (0.0–6.7)0.0 (0.0–16.6)0.2918
Non-opioid analgesics (RDD)0.5 (0.0–2.0)0.7 (0.0–2.2)0.8834

The proportions of women who used the acupressure wrist bands in the two groups and the use of rescue anti-emetics did not differ significantly between the groups during the hospital stay (Table 4). With per-protocol analysis a significant difference indicating a risk of vomiting on day 0 after SA occurred (OR 2.08; 95% CI 1.08–4.02). No significant difference in use of rescue anti-emetics was seen in the GA group between those with and without vomiting episodes, whereas a significantly higher proportion of those with vomiting episodes in the SA group received rescue anti-emetics (86 versus 30%; OR 4.56; 95% CI 4.56–44.06). The dose of opioids used on day 0 or day 1 was not associated with an occurrence of vomiting (data not shown). No vomiting episodes were observed in any of the women after day 2. The volume of intravenous crystalloids and colloid expanders (hydroxyethyl starch) given during the period in hospital did not differ significantly between the groups [1888 ml (400–6000 ml) and 0 ml (0–1500 ml) in GA versus 1775 ml (1000–4300 ml) and 0 ml (0–1000 ml) in SA, respectively].

Table 4.   Vomiting and use of anti-emetics and antipruritics postoperatively
CharacteristicsGeneral anaesthesiaSpinal morphine anaesthesiaOR (95% CI)*
  1. Figures indicate number and (%).

  2. *Logistic regression analysis with adjustments for mode of hysterectomy, skin incision, BMI, smoking habits and change of mode of anaesthesia.

Day 0 (i.e. day of surgery)n = 80n = 82 
Vomiting (no. of women)24 (30.4%)36 (43.9%)1.90 (0.98–3.59)
Anti-emetics (no. of women):
 Acupressure wrist bands only33 (41.2%)36 (43.9%)reference
 Acupressure wrist bands + anti-emetics40 (50.0%)42 (51.3%)1.02 (0.52–1.98)
 Anti-emetics only5 (6.3%)2 (2.4%)0.41 (0.07–2.28)
 No acupressure wrist bands or anti-emetics2 (2.5%)2 (2.4%)1.03 (0.13–7.98)
Antipruritics (no. of women)2 (2.5%)31 (37.8%)25.26 (5.75–111.04)
Day 1n = 80n = 82 
Vomiting (no. of women)11 (13.9%)10 (12.2%)0.76 (0.28–2.12)
Anti-emetics (no. of women):
 Acupressure wrist bands only45 (56.3%)44 (53.7%)reference
 Acupressure wrist bands + anti-emetics15 (18.7%)11 (13.3%)0.70 (0.28–1.79)
 Anti-emetics only4 (5.0%)3 (3.7%)0.74 (0.15–3.75)
 No acupressure wrist bands or anti-emetics16 (20.0%)24 (29.3%)1.59 (0.74–3.41)
Antipruritics (no. of women)3 (3.8%)8 (9.8%)3.16 (0.79–12.64)
Day 2n = 56n = 47 
Vomiting (no. of women)1 (1.8%)1 (2.1%)NA
Anti-emetics (no. of women):
 Acupressure wrist bands only32 (57.1%)24 (51.1%)reference
 Acupressure wrist bands + anti-emetics3 (5.4%)2 (4.2%)0.63 (0.05–7.92)
 Anti-emetics only1 (1.8%)0 (0%)NA
 No acupressure wrist bands or anti-emetics20 (35.7%) 21 (44.7%)1.18 (0.49–2.85)
Antipruritics (no. of women)1 (1.8%)2 (4.2%)2.30 (0.19–27.32)

The number of complications during the hospital stay was low in both groups (Table 5). None of the patients were readmitted to PACU, and no untoward cardiovascular or respiratory adverse effects, or other unexpected complications associated with anaesthetic drugs, were observed in the gynaecological wards during the hospital stay. Two complications were categorised as severe (life-threatening): a heavy perioperative bleed (3000 ml) deriving from a laceration of the uterine artery from the internal iliac artery, and a pulmonary arterial embolism diagnosed on day 1. Both complications were treated according to clinical practice, and the patients recovered uneventfully and were discharged on days 4 and 6, respectively. The median time to removal of transurethral catheters was 22 hours after insertion in both groups, and two women in each group were discharged with the catheter in place because of urinary retention or prolonged catheterisation as a result of bladder injury.

Table 5.   Peri- and postoperative complications during hospital stay
ComplicationsGeneral anaesthesia (n = 80)Spinal morphine anaesthesia (n = 82)
  1. Figures denote numbers and (%).

  2. No statistically significant differences were observed in any of the variables between the groups (univariate analysis, Fisher’s exact tests).

Heavy bleeding, exceeding ≥1000 ml2 (2.5%)2 (2.4%)
Blood transfusion (no. of women)3 (3.8%)0 (0%)
Bladder injury1 (1.3%)2 (2.4%)
Reoperation due to bladder injury0 (0%)1 (1.2%)
Reoperation due to retained surgical towels1 (1.3%)0 (0%)
Retroperitoneal and subcutaneous emphysema1 (1.3%)0 (0%)
Pulmonary embolism0 (0%)1 (1.2%)
UTI1 (1.3%)1 (1.2%)
Cholecystolithiasis1 (1.3%)0 (0%)
Post-dural puncture headache0 (0%)1 (1.2%)
Urticaria1 (1.3%)1 (1.2%)
Urinary catheter at discharge (no. of women)2 (2.5%)2 (2.4%)


This study showed that by using a fast-track model it was possible to achieve a median duration of stay in hospital after abdominal hysterectomy for benign conditions of <50 hours. The length of stay did not differ significantly whether spinal anaesthesia including morphine intrathecally or general anaesthesia was used. The quantity of opioid analgesics used postoperatively was significantly lower after spinal anaesthesia. This had a positive impact on the speed of recovery of bowel function. The negative effects of the spinal anaesthesia with morphine intrathecally compared with GA were a higher prevalence of pruritus and postoperative vomiting.

The study was randomised and performed in a multicentre setting, which strengthens the generalisability of the results. Systematic information about all women admitted to the participating hospitals for benign hysterectomy during the study period has been obtained. More than 38% of the women who were eligible for the study participated. This is a significantly higher proportion than the 5.5–18% reported in other randomised studies of abdominal hysterectomy.10–13

The study was primarily designed to compare the duration of hospital stay and patient recovery following a standard customary procedure for abdominal hysterectomy in GA, with the same measures following hysterectomy in SA. Blinding and/or placebo control was not possible in this study. The temporary paralysis of the lower extremities after SA would, for obvious reasons, be observed immediately by the patient, as well as by the staff. The lack of blinding may pose a risk of bias. In order to reduce such potential bias the women were informed and monitored in a standardised fashion, and the mode of incision and type of abdominal hysterectomy were decided prior to randomisation. The surgical technique of the hysterectomy had only a few limitations according to the study protocol in order to represent the daily practice of abdominal hysterectomy. The surgeon’s competence, individual surgical technique and experience with surgery in spinal anaesthesia may theoretically influence postoperative recovery, and thus presents a bias. However, all surgeons were familiar with gynaecological surgery under spinal anaesthesia, and the operations were performed or supervised by consultants in obstetrics and gynaecology, thus reflecting the daily practice in the clinics.

We failed to demonstrate any difference in hospital stay related to the two modes of anaesthesia, but it is striking that rapid postoperative discharge is achievable by adopting the fast-track recovery ethos. In addition, it is possible that the short duration of hospitalisation post-surgery reflects the clinical trial setting and an information bias effect. All women were carefully informed preoperatively about the discharge criteria, and were actively encouraged by the staff and the research nurses to achieve early mobilisation postoperatively. All staff members were well informed about the fast-track concept, and that discharge was possible when the criteria were met. Compared with the results in published randomised studies concerning abdominal hysterectomy, this study reports a considerably shorter duration of hospital stay.1,11,13–15 However, in most of these studies the criteria to define duration of hospital stay was not clear. In our study the duration of hospital stay was measured as the time from the start of anaesthesia to the time of discharge from hospital. We consider this to be a useful and generalisable method for the calculation of hospital stay in surgery as it eliminates preoperative practices that are often not standardised.

The consumption of opioids was significantly lower in the SA group during the first 2 days postoperatively, indicating a sustained postoperative analgesic effect of the intrathecal morphine. This is in accordance with the results presented by Massicotte et al.16 Because of the lower requirement of opioids by the women in the SA group, they had a faster recovery of bowel function. However, this did not affect the need for the treatment of nausea, nor did it affect the duration of hospital stay. These results seem to be in accordance with those of Hansen et al.,17 who investigated oral osmotic laxatives as a means of hastening bowel recovery after abdominal hysterectomy.

As the risks of postoperative nausea and vomiting, according to the Cochrane Review by Lee et al.,18 are reported to be similar after acupoint P6 stimulation and anti-emetic drugs, we chose to use the acupressure wrist bands as a preventive measure in order to avoid undesired side effects of anti-emetic drugs, which could potentially influence the duration of hospitalisation.19 Both groups also received propofol either for induction and maintenance of the general anaesthesia or as sedation in association with spinal anaesthesia. Propofol possesses direct anti-emetic effects.20 The need for rescue anti-emetics did not differ between the groups, irrespective of the quantity of consumed opioids in the present study. Significantly more women vomited postoperatively in the group who solely received SA, even though there was no significant difference in the overall use of rescue anti-emetics between the groups. It is a clinical observation that vomiting is not always preceded or accompanied by nausea. Such vomiting might be triggered by the gastric paralysis after intra-abdominal surgery. Only the women who had general anaesthesia received an orogastric tube and gastric emptying during the hysterectomy. This might contribute to explaining why the prevalence of women who vomited was higher in the group that had SA and did not need conversion to GA.

Pruritus is a known and common adverse effect of intrathecal morphine occurring in about 37%.21 In our study, 38% of the SA group received antipruretics, which was significantly higher than the 2.5% in the GA group. This significant difference in use of antipruretics was only seen on the day of surgery.

The advantages with SA compared with GA in a fast-track hysterectomy model were that significantly less opioids were needed postoperatively, and consequently there was faster recovery of gastrointestinal function. However, the clinical significance of this may be limited given that the median reduction in time to pass flatus and first bowel movement was just 2 and 7 hours, respectively, in favour of SA. Despite these advantages no difference was found in hospital stay between the modes of anaesthesia. This may reflect the disadvantages of SA, namely the increased risks of vomiting and pruritus needing treatment.

Hysterectomy is encumbered with a substantial number of annoying postoperative side effects, which, if not life-threatening, do affect the patient’s comfort, and consequently the character of postoperative care. However, the adoption of a fast-track concept of recovery encompassing the education of staff and patients, in conjunction with a supportive infrastructure of health care, can result in rapid postoperative recovery and discharge from hospital, regardless of the mode of anaesthesia. Future studies of fast-track programmes in gynaecological surgery encompassing improved anaesthesia, with an emphasis on patient comfort and safety, are needed in order to further enhance postoperative recovery.

Disclosure of interests

None of the authors or GASPI study group members has any conflicts of interest to declare.

Contribution to authorship

NBW, LN and PK planned and conducted the study. The data were processed and analysed by NBW, LN and PK. NBW was the primary author of the manuscript. All authors contributed to the elaboration of the manuscript, revised the paper and approved the final version.

Details of ethics approval

The study was registered on the protocol registration system (NCT00527332) with initial release 7 September 2007. The Regional Ethical Board at Linköping University (registration no. M159-06, approval date 15 November 2006; amendment 1, registration no. T83-07, approval date 3 October 2007; amendment 2, registration no. T19-08, approval date 4 March 2008); and the Swedish Medical Products Agency (registration no. EudraCT nr 2006-002520-41) approved the study.


The study was supported financially by grants from the Medical Research Council of south-east Sweden, Linköping University and the County Council of Östergötland.


We are indebted to Associate Professor Sigga Kalman for valuable discussions in the prelude to the study. The physicians and research nurses in the multicentre study group are thanked for their invaluable work and support during the study.