Short-term recovery after subtotal and total abdominal hysterectomy—a randomised clinical trial


  • P Persson,

    1. Department of Obstetrics and Gynaecology, University Hospital, Uppsala, Sweden
    Search for more papers by this author
  • J Brynhildsen,

    1. Division of Obstetrics and Gynaecology, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, University Hospital, Linköping, Sweden
    Search for more papers by this author
  • P Kjølhede,

    1. Division of Obstetrics and Gynaecology, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, University Hospital, Linköping, Sweden
    Search for more papers by this author
  • on behalf of the Hysterectomy Multicentre Study Group in South-East Sweden

Dr P Persson, Department of Obstetrics and Gynaecology, University Hospital, 751 85 Uppsala, Sweden. Email


Please cite this paper as: Persson P, Brynhildsen J, Kjølhede P on behalf of the Hysterectomy Multicentre Study Group in South-East Sweden. Short-term recovery after subtotal and total abdominal hysterectomy—a randomised clinical trial. BJOG 2010;117:469–478.

Objective  To determine whether the day-by-day postoperative recovery differs between women undergoing subtotal and total abdominal hysterectomy, and to analyse factors associated with postoperative recovery and sick leave.

Design  A prospective, open, randomised multicentre trial.

Setting  Seven hospitals and one private clinic in the south-east of Sweden.

Population  Two-hundred women scheduled for hysterectomy for benign conditions were enrolled in the study, and 178 women completed the study. Ninety-four women were randomised to subtotal abdominal hysterectomy; 84 women were randomised to total abdominal hysterectomy.

Methods  The day-by-day recovery of general wellbeing was measured on a visual analogue scale in a diary 7 days preoperatively and 35 days postoperatively. Psychometric measurements included depression, anxiety and general psychological wellbeing.

Main outcome measures  Effects of operating method and preoperative wellbeing on the day-by-day recovery and duration of sick leave.

Results  No significant difference was found in the day-by-day recovery between operating methods. The day-by-day recovery of general wellbeing and duration of sick leave was strongly associated with the occurrence of minor complications, but not with major complications. The level of psychological wellbeing preoperatively was strongly associated with the day-by-day recovery of general wellbeing and duration of sick leave.

Conclusions  Day-by-day recovery of general wellbeing is no faster in subtotal versus total abdominal hysterectomy. Independent of operation method there is an interaction between preoperative psychological wellbeing, postoperative recovery of general wellbeing and the duration of sick leave. Postoperative complications and preoperative psychological wellbeing are strong determinants for the duration of sick leave. There is a need for intervention studies with a focus on complications and preoperative wellbeing.


For benign gynaecological disorders subtotal abdominal hysterectomy (SH) has gained popularity compared with total abdominal hysterectomy (TH) because of a shorter operating time, less peroperative bleeding and fewer perioperative complications.1 The introduction of the laparoscopic subtotal hysterectomy has further increased the popularity of the preservation of the cervix.2,3 Observational studies from the 1980s and 1990s showed potential advantages with SH, compared with TH, regarding sexual outcome, and the postoperative general wellbeing seemed to be better after SH compared with TH.4–6 In a meta-analysis of SH and TH, Gimbel stated that conclusions regarding quality of life outcome could not be drawn.1 No difference has been found in the postoperative recovery time between TH and SH.7–9 Whether the day-by-day postoperative recovery of general wellbeing differs between the two modes of hysterectomy has not been investigated.

Recovery is not only a matter of treatment of postoperative physical symptoms and complications, but is also likely to depend on preoperative psychological wellbeing and the ability for psychological recovery.10,11 The speed of the postoperative recovery depends on optimal management of postoperative symptoms such as pain, nausea and intestinal paralysis, and on the occurrence of postoperative complications.12 In the randomised clinical trials (RCTs) comparing SH and TH it was only in the study by Thakar et al. that psychiatric symptoms were measured.8,9,13,14 In these studies the main focus was upon surgical outcome, and in secondary analyses sexuological outcome.14–16 No analyses were performed on the impact of preoperative psychological wellbeing on the outcome measures. Some studies have shown that preoperative psychological wellbeing is associated with postoperative overall wellbeing,17,18 and that personality factors such as neuroticism and coping strategies may be of importance.11,19 Whether these factors also have an impact on the postoperative day-by-day recovery has not been investigated, and may be of great interest in order to tailor postoperative periods for patients with low levels of wellbeing by optimising their physical and psychological recoveries.

The aims of this randomised study were to determine whether the day-by-day postoperative recovery differed between SH and TH, and to analyse factors associated with postoperative recovery and sick leave. Our hypothesis was that SH, because of its presumed less invasive character and lower complication rate, would give a faster day-by-day recovery in general wellbeing than TH.


The departments of Obstetrics and Gynaecology at seven hospitals and one private gynaecological clinic in the south-east health regions of Sweden participated in this randomised, controlled, open, multicentre study of SH versus TH. The hospitals comprised three county hospitals, three central hospitals, and one university hospital.

Women admitted for hysterectomy because of benign gynaecological conditions, between March 1998 and April 2004, were eligible for the study. Not all of the departments actively recruited patients during the whole period. Proficiency in Swedish was a prerequisite for inclusion.

Medical inclusion criteria were primarily uterine fibroids with bleeding disturbance or mechanical symptoms, but other benign disorders where hysterectomy was recommended were also included. After the operation at least one ovary was to be preserved. Exclusion criteria were malignancy in the genital organs, previous or present cervical dysplasia, rapidly growing fibroids where malignancy could not be ruled out preoperatively, preoperative treatment with gonadotrophin-releasing hormone analogues, postmenopausal women without hormone therapy (HT), and severe psychiatric disorders.

The calculation of the sample size was based on the results from the Psychological General Well-Being (PGWB) questionnaire from a previous study.20 In order to detect a difference in the PGWB score between the two surgical methods as being statistically significant, we assumed that an increase in the PGWB score of 8 points was clinically relevant. Thus, to demonstrate that this difference was statistically significant in 90% of the women at a 5% level (two-sided testing), i.e. with a 90% power, 166 patients were necessary. We estimated the drop-out rate to be 20%, and thus approximately 200 patients were needed for the study. After written and verbal informed consent the women were randomised to either TH or SH according to a random table. The randomisation was conducted as a block randomisation with eight blocks, one for each participating centre. Each block contained 26 objects with equal numbers of the two surgical methods. The surgical procedure was written on a label sealed in opaque envelopes that were numbered sequentially in accordance with the random table. The envelopes were opened consecutively. The women were assigned approximately 1 week before surgery and were informed about their assignment prior to surgery. If a centre could not recruit the planned number of women to the study within a specified period of time, the remaining randomisation envelopes were transferred to the university hospital, which recruited the remaining patients. Because of a decreased inclusion rate during the last year of the inclusion period the inclusion was stopped when 200 women were randomised. A total of 178 women completed the study: 94 allocated to SH and 84 to TH. These women constitute the study population. The flow chart of women in the study is illustrated in Figure 1.

Figure 1.

 Flowchart of the study participants. The number of women assessed for eligibility corresponds to the total number of hysterectomies for benign indications performed in the participating hospitals during the study period. Hysterectomies carried out in association with operations for benign ovarian tumours or genital prolapse are not included. No systematic information was collected during the study regarding reasons for choosing not to enrol in the study.

The study was approved by the ethics research committees of Linköping University and Örebro Regional Hospital.


The hysterectomy was performed under general anaesthesia through a transverse or low midline skin incision, decided on by the surgeon based on the size of the uterus or previous surgery. The surgical technique of the hysterectomy was left to the surgeon’s discretion. The only prerequisites according to the study protocol were that the uterine cervix was dissected extrafascially and no peritonealisation was made in either SH or TH. The patients should preferably receive prophylactic antibiotics with a single dose of 1.5 g cefuroxim and 1 g metronidazol, intravenously. In the case of allergy to penicillin each hospital used prophylactic antibiotics according to local policy.

Epidural analgesic was allowed peri- and postoperatively according to local clinical routine; otherwise, the standard postoperative analgesics were paracetamol [1 g four times per day] and non-steroidal anti-inflammatory drugs administered enterally. Opioids were given parenterally on a need basis.

The patient was discharged from the hospital when she was mobilised and had sufficient pain relief with oral analgesics. At discharge the patient had a sick leave of 21 days. The research nurse contacted the patient regarding general wellbeing approximately 10 days after discharge and then once weekly until the patient considered she had recovered. The sick leave was prolonged with at most 7 days at a time, if necessary, until the patient was able to return to work.

The patient visited the outpatient clinic after 5 weeks. A clinical examination was carried out and data on postoperative complications and the de facto period of sick leave were recorded.

Measurement of general wellbeing

Starting 1 week before surgery and continuing daily until day 35 postoperatively, the women completed a diary concerning their general wellbeing. On a visual analogue scale (VAS) ranging from 0 to 100 they were asked to state how they considered their overall general wellbeing was on average for the past day. Zero represented extremely bad wellbeing and 100 represented feeling extremely well. Furthermore, the women were asked to report their consumption of analgesics after discharge in the diary.

Psychological measures

Psychological functioning was measured approximately 1 week preoperatively by psychometric tests. In order to measure psychological general wellbeing the PGWB and the Women’s Health Questionnaire (WHQ) were used. These are validated instruments for this purpose.21–23 The State-Trait Anxiety Inventory (STAI) and Beck’s Depression Inventory (BDI) were also used for measuring specific psychological functioning. The STAI and BDI tests were chosen so that changes regarding anxiety and depression could be specifically detected. The PGWB index was constructed to measure personal affective states, reflecting wellbeing or distress. It consists of 22 questions referring to anxiety, depression, wellbeing, self-control, health and vitality.21 Each question is rated on a six-point scale from 1 to 6. The total score ranges from 22 to 132: the higher the score, the higher the degree of wellbeing. The PGWB has been validated against various mental health scales. Its Swedish version has also been validated.22

The WHQ is a questionnaire providing a detailed examination of minor psychological and somatic symptoms experienced by peri- and postmenopausal women.23 It consists of 36 questions grouped in nine sections describing somatic symptoms, depressive mood, cognitive difficulties, anxiety, sexual function, vasomotor symptoms, sleep problems, menstrual symptoms and attraction. Each question is rated from 1 to 4, and the total score ranges from 36 to 144: a higher score indicates more distress and dysfunction. The Swedish translation of the WHQ has been validated.24

The STAI is a questionnaire that assesses anxiety in two different forms, representing state and trait. In this study the trait form is used. The trait form (Y-2) consists of 20 statements that evaluate how the respondents feels ‘generally’.25 Individuals respond to each item on a four-point Likert scale, indicating the frequency with which each strategy is used. The total score ranges from 20 to 80, and increases in response to physical danger and psychological stress. The STAI has been widely used in assessing general anxiety in medical, surgical and psychiatric patients, and has been translated to Swedish, but has not been validated for Swedish.

The BDI is a quantitative self-reporting scale for measuring the presence and severity of depression in clinical and normal populations of adults and adolescents.26 It is made up of 21 items, with each item rated on a four-point scale (0–3) in increasing severity, adding up to a total score of 0–63: a high score indicates a more depressive state. The BDI has been translated to, but not validated for, Swedish.


All analyses were performed according to intention-to-treat principles. For the analyses of the descriptive and demographic data, Student’s t tests were used for continuous data demonstrating approximately normal distribution. Yates’ corrected chi-square and Fisher’s exact test were used for nominal data. Analysis of variance (ANOVA) for repeated measurements was used to compare the results of general wellbeing and psychological measurements between the two groups. Statistical significance was accepted at the 5% level.

On the basis of data from previously published articles concerning psychometric measurements and day-by-day recovery of general wellbeing after hysterectomy,11,20 as well as empiric considerations, we estimated that median preoperative scores of 101, 61 and 33 in the PGWB, WHQ and STAI, respectively, could be discriminatory for a high and low capability of postoperative recovery. These values are in the range of the preoperative mean values quoted in previously published material regarding women undergoing benign hysterectomy, or women who are perimenopausal.11,20,22,25,27 For BDI we used the 75-percentile score of our study group, i.e. 9, which is equivalent to the discriminatory value of BDI for normal and mild depressive states.28


The demographic and clinical data as demonstrated in Table 1 did not differ significantly between the groups. Concerning peroperative data, the operating time was significantly longer and prophylactic antibiotics more frequently used in TH than in SH. Otherwise, no significant differences were observed in the peroperative data, shown in Table 2. One woman, allocated to TH had an SH for technical surgical reasons.

Table 1.   Preoperative demographic and clinical data
 Subtotal abdominal hysterectomy (n = 94)Total abdominal hysterectomy (n = 84)
  1. Data are expressed as means ± 1 standard deviation, or number and (%).

  2. BMI, body mass index; HT, hormone treatment.

Age (years)45.9 ± 5.045.7 ± 5.1
Gravida (n) 2.8 ± 1.4 2.7 ± 1.4
Parity (n) 2.2 ± 1.0 2.0 ± 1.1
Nulliparous (no. women) 8 (8.5) 7 (8.3)
BMI (kg/m2)25.7 ± 3.625.8 ± 4.3
BMI < 25 (no. women)44 (47)40 (47)
BMI ≥ 25 and <30 (no. women)36 (38)30 (36)
BMI ≥ 30 (no. women)14 (15)14 (17)
Smokers (no. women)25 (27)23 (28)
Previous laparotomy in lower abdomen (no. women)40 (43)28 (33)
Present medication
 Antidepressants (no. women) 2 (2.1) 4 (4.8)
 Analgesics (no. women)13 (13.8)12 (14.3)
 HT climacterium (no. women) 7 (7.4) 3 (3.6)
 HT bleeding disturbances  (no. women) 7 (7.4)12 (14.3)
 Hormonal contraceptives (no. women) 9 (9.6)12 (14.3)
Main indication for hysterectomy
 Bleeding disturbances (no. women)65 (6962 (74)
 Mechanical symptoms (no. women)21 (22)15 (18
 Dysmenorrhea/endometriosis/  pelvic pain (no. women) 8 (9) 7 (8)
Table 2.   Peroperative data
 Subtotal abdominal hysterectomy (n = 94)Total abdominal hysterectomy (n = 84)
  1. Data are expressed as means ± 1 standard deviation, or number and (%).

  2. Haemoglobin and erythrocyte volume fraction (EVF) postoperative are measured on postoperative day 2 or prior to blood transfusion.

  3. *P = 0.0085; **P = 0.0015.

  4. ***From day of surgery to the day of discharge.

  5. ****In equipotent morphine dosage.

  6. *****In total of recommended daily dosage (RDD).

Skin incision
Midline (n)17 (18)17 (20)
Low transverse (n)77 (82)67 (80)
Operating time (minutes)70 ± 2380 ± 28*
Uterine weight (g)456 ± 332440 ± 300
Blood loss (ml)222 ± 236243 ± 201
Blood loss per unit time (ml/h)175 ± 138173 ± 128
Blood transfusion (no. women)5 (5.3)3 (3.6)
Haemoglobin preoperative (g/l)128 ± 13126 ± 15
Haemoglobin postoperative (g/l)110 ± 17108 ± 16
EVF preoperative (%)39 ± 438 ± 4
EVF postoperative (%)33 ± 532 ± 5
Intravenous fluids peri/postoperatively (l)3.9 ± 1.54.0 ± 1.3
Prophylactic antibiotics (no. women)81 (86)83 (99)**
Duration of hospital stay (days)***3.4 ± 1.23.4 ± 1.1
Duration of sick leave (days)32.5 ± 9.433.6 ± 11.3
Epidural analgesia (no. women)24 (26)24 (29)
Parenteral and enteral opioids (mg)****
 Postop day 115.8 ± 14.217.9 ± 16.7
 Postop day 27.7 ± 9.59.2 ± 13.3
 Postop day 33.9 ± 5.24.2 ± 7.9
 Postop day 42.6 ± 3.82.2 ± 3.9
 Postop day 51.8 ± 3.11.9 ± 3.4
Enteral non-opioid analgesics (RDD)*****
 Postop day 11.0 ± 0.41.1 ± 0.5
 Postop day 21.0 ± 0.41.0 ± 0.5
 Postop day 30.8 ± 0.50.9 ± 0.5
 Postop day 40.7 ± 0.50.7 ± 0.5
 Postop day 50.6 ± 0.50.6 ± 0.5

The use of analgesics is shown in Table 2. No significant differences were observed between the groups in use of parenterally or enterally administered analgesics, even when adjusted for the use of epidural analgesia. None of the women received opioids parenterally after postoperative day 4.

Association between day-by-day recovery of general wellbeing and mode of hysterectomy

An illustration of the recovery of day-by-day general wellbeing is shown in Figure 2. The ANOVA for repeated measurements revealed that day-by-day general wellbeing did not differ significantly between the hysterectomy groups in the preoperative and postoperative periods (= 0.9010 for the difference between the groups in the preoperative period, and P = 0.5048 for the postoperative period). There were highly significant main effects over time (P < 0.0001) for the preoperative as well as the postoperative periods. In the preoperative period the general wellbeing deteriorated day-by-day, and in the postoperative period it improved. For the postoperative period a significant interaction effect was observed (P = 0.0029). As illustrated in Figure 2, the speed of recovery was initially higher for women in the SH group than for the TH group, but later on in the postoperative period it was vice versa. For the postoperative period this holds true even when adjusted for analgesics and complications, but in women with complications the interaction effect disappeared. The women had regained their self-rated general wellbeing equivalent to the mean value of the entire study population of preoperative day –7 on postoperative day 22 in the SH group, compared with day 19 in the TH group.

Figure 2.

 Illustration of self-rated general wellbeing according to the diary 1 week prior to surgery and 35 days postoperatively. Bars represent means; error bars represent ± SD.

The complications are described in Table 3. Minor complications included: abdominal wound infection or seroma; vaginal cuff infection; cervicitis; lower urinary tract infection; prolonged period with pain; fatigue and other conditions occurring in the postoperative period causing substantial, but not severe, discomfort to the patient. Major complications consisted of: peri- or postoperative intra-abdominal bleeding exceeding 1000 ml; blood transfusion; respiratory depression; re-laparotomy and re-admittance for subileus.

Table 3.   Peri- and postoperative complications
 Subtotal abdominal hysterectomy (n = 94)Total abdominal hysterectomy (n = 84)
  1. Data are expressed as means ± 1 standard deviation, or number and (%).

  2. A woman may have more than one complication.

  3. *Other complications consist of unspecified neuralgia, diarrhoea and sore throat.

Complications during hospital stay
Peroperative bleeding exceeding 1000 ml3 (3.2)0 (0)
Re-operation, postoperative intra-abdominal bleeding1 (1.1)0 (0)
Postoperative bleeding requiring blood transfusion4 (4.3)3 (3.6)
Postoperative bleeding treated conservatively0 (0)2 (2.4)
Unspecified infections treated with antibiotics2 (2.1)2 (2.4)
Respiratory depression (side effect of opioids)0 (0)1 (1.2)
Complications within 5 weeks of discharge from hospital
Wound infection/seroma8 (8.5)6 (7.1)
Vaginal cuff infection/cervicitis1 (1.1)1 (1.2)
Lower urinary tract infection5 (5.3)3 (3.6)
Abdominal/pelvic pain4 (4.3)3 (3.6)
Fatigue2 (2.1)2 (2.4)
Subileus0 (0)1 (1.2)
Other complications*3 (3.2)2 (2.4)

The surgery and postoperative period was uncomplicated in 68 women (72%) in the SH group, and in 64 (76%) in the TH group. Major complications occurred in seven women (7.4%) in the SH group and in four (4.8%) in the TH group. The corresponding figures for minor complications were 19 (20.2%) and 17 (20.2%), respectively. None of these differences reached statistical significance.

Factors associated with postoperative recovery of day-by-day general wellbeing

The postoperative day-by-day general wellbeing was strongly negatively associated with the occurrence of complications (ANOVA for repeated measurements: P = 0.0004 for the main effect between groups; P < 0.0001 for the main effect over time and no significant interaction). This was mainly attributed to minor complications, whereas no significant association was observed for major complications. The self-rated general wellbeing equivalent to the mean value of preoperative day –7 of the entire study population was regained on postoperative day 28 when complications had occurred, compared with day 18 when no complications occurred.

The category outcome measures of the psychometric tests (low versus high) were all strongly associated with the recovery of postoperative day-by-day general wellbeing. However, there were also strong associations with preoperative general wellbeing for all four psychometric tests. The results of the ANOVA tests for repeated measurements of pre- and postoperative general wellbeing, and the psychometric tests including estimation of time of regained general wellbeing equivalent to the self-rated general wellbeing on preoperative day –7, are presented in Table 4.

Table 4.   Associations between day-by-day general wellbeing pre- and postoperatively, and the four psychometric measurements in the entire study population
Outcome measureComparison between groupsTimingRepeated measures ANOVAPostoperative day when self-rated general wellbeing was regained*
Main difference between groups (P)Main difference over time (P)Interaction (P)CategoryDayCategoryDay
  1. *The self-rated general wellbeing equivalent to the mean value of preoperative day –7 of the entire study population.

  2. PGBW low versus high, total score of PGBW ≤ or > than 101; WHQ high versus low, total score of WHQ ≥ or < than 62; BDI high versus low, total score of BDI> or ≤ than 9; STAI high versus low, total score of STAI > or ≤ than 33.

General wellbeingPGWB low versus highPreoperative days −7 to −1<0.00010.00010.6911    
  Postoperative days 1 to 350.0003<0.00010.8503PGWB high17PGBW low28
General wellbeingWHQ high versus lowPreoperative days −7 to −1<0.00010.00010.2778    
  Postoperative days 1 to 350.0004<0.00010.9879WHQ low17WHQ high28
General wellbeingBDI high versus lowPreoperative days −7 to −1<0.00010.01990.1860    
  Postoperative days 1 to 350.0010<0.00010.8597BDI low18BDI high31
General wellbeingSTAI high versus lowPreoperative days −7 to −10.00010.00020.0819    
  Postoperative days 1 to 350.0050<0.00010.9974STAI low18STAI high26

Sick leave

The 25 percentile of duration of sick leave was 27 days. The duration of sick leave was also strongly associated with recovery in terms of day-by-day general wellbeing postoperatively (ANOVA for repeated measurements: P < 0.0001 for the main effect of the ANOVA between the groups with sick leave of <28 days and those taking 28 or more days; P < 0.0001 for the main effect over time; and P = 0.0440 for an interaction effect), which means that those with a shorter sick leave rated their day-by-day general wellbeing higher, and thus had a faster recovery than those with a longer sick leave.

The sick leave was strongly associated with the occurrence of complications (40 ± 13 days with complications versus 31 ± 8 days without complications; < 0.001). This was in particular attributed to the occurrence of minor complications (42 ± 14 days with minor complications versus 31 ± 8 days without minor complications), whereas the occurrence of major complications did not influence the duration of sick leave significantly (36 ± 11 days with major complications versus 33 ± 10 days without major complications).


This study showed no difference in the day-by-day recovery of self-reported general wellbeing between women undergoing subtotal or total abdominal hysterectomy. No major differences were observed in peroperative clinical data or complications, except for a statistically significant longer operation time for TH. The day-by-day recovery of general wellbeing and the duration of sick leave were strongly associated with the occurrence of minor complications and preoperative psychological wellbeing.

The present study was randomised, which eliminates most selection bias. However, women who agree to participate in RCTs per se constitute a selected group. They may have specific reasons for participation that can constitute a selection bias and thus influence the results. The reasons for participation, or not, have not been investigated in this study. This should be kept in mind when the results are generalised. The inclusion period in the study was 6 years. This was somewhat longer than the 2.5–4 years inclusion periods in the other randomised trial of SH and TH.7–9,13 However, the proportion of enrolled women versus eligible women in this study was similar to that in other randomised studies of SH and TH.8,9

The use of a VAS to measure general wellbeing may seem to be a crude method. However, the method has been used in several studies of hysterectomy concerning quality-of-life assessments and general wellbeing,20,29,30 and was found to be useful. A similar VAS scale is used in the The EuroQol Group questionnaire (EQ 5-D), a standardised instrument for use as a measure of health outcome, which is a well-accepted standardised instrument for use as a measure of health outcome.31 In order to detect changes in psychological wellbeing we used four validated tests: two generic and two specific. These tests have been used in several clinical trials and have been found to be useful in describing changes in psychological wellbeing after treatment. Normative clinical values have been proposed for the specific tests, i.e. BDI and STAI,25,28 whereas no such values exist for PQBW and WHQ. In order to discriminate between low and high levels of psychological wellbeing we therefore used these established values to discern between normal and pathological states for the BDI and STAI. For PGWB and WHQ we used the median total score values of the study population. These scores are similar to the mean total score values of other populations in studies of benign hysterectomy or perimenopausal women.20,22,32,33 We excluded women with severe psychiatric disorders in our study to avoid the bias of psychiatric disease on psychological recovery after hysterectomy. The study population thus represents a ‘psychologically’ healthy group of women, and it may therefore be reasonable to establish ‘cut-off’ limits at the median value of the sum scores of the generic tests to distinguish between low and high levels of psychological wellbeing.

The duration of sick leave is not usually strictly determined by medical decisions: social welfare benefits, local traditions, personal reasons, workload and the attitudes of professionals in the healthcare system may also have an influence. In the present study the women had an initial sick leave of 21 days. The research nurse had telephone contact with the patient weekly, and sick leave was prolonged by up to 7 days per patient request if she did not feel well enough to return to work. If the sick leave exceeded 35 days the patient was appointed to visit the doctor in order to be examined and evaluated for further prolongation of the sick leave, and to receive treatment if necessary. In this way the duration of sick leave was more likely determined by the woman’s experience of recovery.

The randomisation worked well, with no significant differences in preoperative data between the two groups. The operation time was statistically significantly longer in the TH group, which is in accordance with other studies.1,34 However, the difference was only 10 minutes, which makes the clinical importance limited. Contrary to Gimbel’s meta-analysis1 and the Cochrane Review,34 no difference was found in peroperative bleeding between SH and TH. Prophylactic antibiotics were used significantly less often in the SH group. This might influence the risk of postoperative infections,35 but no such increase was observed in this study.

The use of postoperative analgesics was similar in the two groups. Provided that the demand for pain relief was equivalent, this indicates that postoperative pain might be equal after SH and TH. It seems reasonable to assume this as the day-by-day wellbeing did not differ between the two groups. The complications and complication rates did not differ between the modes of hysterectomy. According to Gimbel’s meta-analysis,1 SH was encumbered with fewer complications. This was strongly attributed to the lower frequency of pyrexia, as reported in the study by Thakar et al.7 In our study we did not register pyrexia as a complication per se. It was only registered as a complication when associated with infectious morbidity. The rates of complications in the present study are in line with those described in the other RCT of SH and TH.1

The day-by-day recovery of general wellbeing did not differ between the two groups, even when adjusted for use of analgesics. Although it has been a common belief among women and gynaecologists favouring SH that it should be more favourable for the woman, even concerning recovery with less complications and pelvic organ dysfunction, there is a lack of information about this in the literature.36,37 The results of this study may indicate that the surgical trauma and stress is similar for the two methods, and thus SH cannot be considered as less invasive or more lenient to the woman than TH.

Regarding the association between sick leave and complications, these results are in the line with a previous study from our group.11 Major complications, often bleeding or a re-operation, arise in the operating room or in the immediate postoperative period. If adequately managed they seldom have an impact after 2 or 3 weeks postoperatively. This is in contrast with minor complications, as for instance a wound infection arising in the postoperative period when the woman is back in her domestic environment trying to return to daily life.

The psychometric measurements showed strong associations with postoperative recovery, i.e. women who scored low in these tests recovered more slowly. Psychological wellbeing is of importance for the length of the recovery period, and it seems important to preoperatively identify women with low psychological wellbeing in order to give them extra support, so as to minimise the period of recovery and improve wellbeing. Our findings support the view that consequences of hysterectomies are more likely to be predicted on the basis of psychological traits that existed before the operation.16,38 Whether this has its roots in different premorbid personalities or is caused by preoperative stress, or by symptoms that indicate the hysterectomy, is still unclear. There are some data indicating that personality factors such as neuroticism and stress-coping ability seems to be of importance.10,11,39 Work to create psychological models to understand the mechanisms behind the link between gynaecological and psychological symptoms has been undertaken,19 but the answer remains unclear. Since the rise of the RCT era most studies have compared the outcomes of different modes of hysterectomy, but maybe the solution to the enigma of the link between gynaecology and psychological symptoms lies more profoundly in the person that we call the patient.

In conclusion, the day-by-day recovery of general wellbeing after SH and TH is equal. The occurrence of postoperative complications and a low preoperative level of psychological wellbeing impair the postoperative recovery significantly, and prolong the duration of sick leave. Further studies should be encouraged to determine the impact of intervention strategies on the recovery of women with low psychological capacity after hysterectomy. Also, efforts should be made to reduce postoperative complications, especially minor complications, in order to improve recovery.

Disclosure of interests

None of the authors or the study group reports any conflict of interest concerning this study.

Contribution to authorship

PP and PK planned the study, and collaborated in the surgical procedures and clinical follow up. PP, PK and JB analysed the data. PP was the primary author of the manuscript. All of the authors revised the paper and approved the final version.

Details of ethics approval

The study was approved by the ethics research committees of Linköping University and Örebro Regional Hospital.


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


In addition to the main authors of the manuscript, the Hysterectomy Multicentre Study Group in South-East Sweden includes the following gynaecologists: Drs Raymond Lenrick and Laila Falknäs, Ryhov Central Hospital, Jönköping; Drs Urban Jürgensen and Johan Holmberg, the County Hospital, Eksjö; Dr Christina Gunnervik, the County Hospital, Värnamo; Dr Lisbeth Angelsiöö, the Central Hospital, Kalmar; Dr Kenneth Krohn, Vrinnevi Central Hospital, Norrköping; Drs Leif Hidmark and Anna Saland, the County Hospital, Motala; Drs Fatma Bäckman and Björn Holm, Capio Läkargruppen, Örebro; and Drs Lars Fåhraeus and Mats Hammar, and psychologist Dr Klaas Wijma, the University Hospital, Linköping. The physicians and research nurses in the multicentre study group are thanked for their invaluable work and support during the study.