Hysteroscopy in daily practice Dongen, H. van

Hysteroscopy in daily practice

Dongen, H. van

Citation

Dongen, H. van. (2009, February 26). Hysteroscopy in daily practice.

Retrieved from https://hdl.handle.net/1887/13533

Version: Corrected Publisher’s Version

License:

Downloaded from:

Note: To cite this publication please use the final published version (if

applicable).

Chapter 6

A randomised comparison of discomfort of vaginoscopic office hysteroscopy and saline infusion sonography

Heleen van Dongen Cor D. de Kroon Sabrina A.H.M. van den Tilaart Trudy C.M. Trimbos-Kemper Leoni A. Louwé Frank Willem Jansen Adapted from BJOG 2008;115:1232-1237

Chapter 6 58

Introduction

Diagnostic hysteroscopy is widely accepted as the standard method for direct visualisa- tion of the uterine cavity and is well tolerated as an outpatient procedure with a high success rate [Nagele 1996b]. A meta-analysis showed that the diagnostic accuracy of saline infusion sonography is as good as the accuracy of diagnostic hysteroscopy [de Kroon 2003a]. Moreover, a few studies showed that saline infusion sonography causes less discomfort for patients than office hysteroscopy [Rogerson 2002; Widrich 1996].

Nowadays, hysteroscopy has evolved considerably in favour of patient compliance thanks to smaller instruments [Sharma 2005], but perhaps more importantly due to the introduction of the vaginoscopic approach: without the use of a speculum and tenacu- lum [Bettocchi 1997]. The possibility to perform a biopsy or a small surgical interven- tion during the same session is an additional advantage of hysteroscopy compared to saline infusion sonography. Therefore, it would be interesting to know whether vagino- scopic office hysteroscopy is comparable to saline infusion sonography with regard to patient discomfort. For this reason we conducted a randomised controlled trial in order to compare pain scores of patients undergoing saline infusion sonography and office hysteroscopy performed according to a vaginoscopic approach.

Methods

This study was conducted at the outpatient clinic of the department of Gynaecology of the Leiden University Medical Center (Leiden, The Netherlands) from January 2006 to July 2007. Women were eligible to participate if further investigation of the uter- ine cavity was indicated, irrespective of age, parity, general health and menopausal state. Patients were excluded if they had received previously office hysteroscopy or saline infusion sonography, if they had a uterine size greater than a twelve-week preg- nancy at physical examination, or if a contraindication for saline infusion sonography or office hysteroscopy existed. Ethical approval for this study was obtained from the Leiden University Medical Center Ethics Committee, and participants gave a written consent.

Randomisation was based on a computer-generated randomisation table. The sequence was concealed with serial sealed opaque envelopes. Randomisation envelopes were drawn after written consent. Two study groups of equal size were generated: (A) saline infusion sonography and (B) office hysteroscopy.

Two clinically experienced investigators carried out the saline infusion sonography (HvD and LL), and three experienced investigators performed office hysteroscopy (LL, TTK and FWJ). In premenopausal women both investigations were preferably performed in the follicular phase of the menstrual cycle. Patients did not receive any antibiotic prophylaxis or analgesics before or during the procedure.

Discomfort of hysteroscopy and saline infusion sonography 59

Saline infusion sonography was performed as previously described [de Kroon 2003b].

After placing a Cusco-speculum the cervix was cleaned with Povidon-Iodine prior to insertion of a Goldstein sonohysterography catheter® (Cook Ireland ltd, Ireland) through the cervical canal. A 5-7.5 MHz B-mode endovaginal probe was used (Toshiba Medical Systems, Zoetermeer, The Netherlands) and sterile normal saline was injected into the uterine cavity with a 20cc syringe. To differentiate between endometrial polyps and sub- mucous myomas the classification published by Parsons and Lense was used [Parsons 1993].

Hysteroscopy was performed following a standardised procedure in all cases, using a 2.0mm rigid diagnostic hysteroscope (Karl Storz, Tüttlingen, Germany), with a 12º foreoblique lens, an operating sheath (5 French) and an outer sheath of 4.2mm diameter including a continuous flow canal. For the vaginoscopic technique, following biman- ual examination, the hysteroscope was placed into the lower vagina and normal saline was infused at a maximum pressure of 80mmHg pressure. The instrument was directed towards the cervix and upon identifying the external ostium, introduced into the cervi- cal canal and guided into the uterine cavity. The cervical canal and uterine cavity were inspected systematically. A hysteroscopy was considered representative if both of the tubal ostia were visualised. For the diagnosis of polyps and myomas, the diagnostic crite- ria previously published by Hamou and Salat-Baroux were used [Hamou 1984]. During the diagnostic hysteroscopies included in this study no therapeutic procedures were performed.

Patients were asked to complete a post-procedure questionnaire immediately following the investigation, scoring their discomfort during the investigation. In addition, they were asked whether they experienced shoulder pain, nausea or dizziness, and whether they would prefer general anaesthesia if they were to undergo the investigation again in the future.

Success of the investigation by the intended technique (defined as adequate inspection of the cervical canal and uterine cavity), need for local anaesthesia, use of instruments (speculum, tenaculum or cervical dilators), procedure time and complications were recorded. Procedure time was defined as the time between the moment that the patient took place in and left the examination chair. Conclusiveness of the investigation, and pain experienced by the patient as assessed by the investigator, were documented as well.

The primary outcome measure was pain during the procedures. This was assessed by the patient and the investigator (blinded from each other) using the visual analogue scale (VAS) of 10cm (0=no pain, 10=unbearable pain) and the present pain intensity (PPI) scale. The PPI scale, derived from the McGill Pain Questionnaire [Melzack 1975], was obtained by grading the patients’ description of pain as: 0=none, 1=mild, 2=discomfort, 3= distressing, 4=horrible, 5=excruciating. Patients were categorised as having signifi- cant pain using a cut-off point of VAS >7 and PPI >2. An adjusted version of the lat- ter scale was also used to score to which extent the patient had experienced shoulder-

Chapter 6 60

pain, nausea and dizziness (0=none, 1=mild, 2=moderate, 3=considerable, 4=heavy, 5=extreme). Additionally, the investigator was asked to complete a VAS depicting how certain he/she was of the diagnosis with 0 as very uncertain and 10 as very certain.

We assumed that the standard deviation of the 10cm VAS pain scores in each group would be close to 1.9cm, based on previously published literature [Rogerson 2002].

It was also assumed that an underlying difference in mean pain score between both techniques of 1.0cm or greater would be of clinical significance [Shankar 2004]. We calculated in NCSS-Pass [Hintze 2001] that a sample size of 50 patients in each group would be needed to prove that both investigations are equal regarding discomfort with a statistical power of 80% at a 5% level of significance.

The received information was collected in the statistical SPSS-program (SPSS, version 14, SPSS Inc., Chicago, IL). Data were analysed with the Student’s t-test for normally distrib- uted continuous variables, the Mann-Whitney U-test for skewed data and the Pearson’s Chi-square test to compare dichotomous data. Confidence intervals (95%-CI) for differ- ence in means were calculated. If continuous variables were not normally distributed, the median and range were given.

To determine whether pain was influenced by other factors a multiple linear regression analysis in a backward stepwise manner was performed, with pain (VAS) as the predict- ing variable and the investigation (saline infusion sonography or hysteroscopy), parity, menopausal state, body-mass-index (BMI), position of the uterus and procedure time as independent variables. A level of p>0.10 was used for removal of variables of the step- wise model. All tests were two-sided and a p-value <0.05 was considered statistically significant.

Results

Of the 121 patients approached, 21 patients were excluded (figure 1). Of these, six patients did not meet the inclusion criteria, and the other fifteen patients refused to participate.

Of the 100 patients randomised, three procedures were not performed; two patients withdrew from the study and one appeared to be pregnant at time of the investiga- tion. Of all patients included 81 (84%) were premenopausal and sixteen (16%) were postmenopausal. Patients’ characteristics of both randomised groups are featured in table 1.

Pain ratings are shown in figures 2 and 3. The level of pain associated with the two techniques differed significantly (p<0.05). The patient’s pain scores on both the VAS and PPI scale were lower for saline infusion sonography compared to office hysteroscopy.

However, among those women experiencing significant pain there was no statistically significant difference (VAS>7 or PPI>2) between both groups. Outcome measures are shown in table 2.

Discomfort of hysteroscopy and saline infusion sonography 61

A multiple linear regression analysis was performed to determine whether pain was influenced by factors other than the investigation itself. Besides saline infusion sono- graphy, multiparity (parity ≥1), a shorter procedure time and an anteflexed position of the uterus appeared to decrease pain scores in the patients studied (table 3).

The correlation of pain experienced by the patient and predicted by the investigator was quite good according to Pearson’s correlation coefficient of 0.715.

In two cases there were complications reported: one vasovagal collapse following saline infusion sonography and one false tract during office hysteroscopy. The success rate for saline infusion sonography was 94.0% compared to 92% for office hysteroscopy (p=0.663). Three saline infusion sonography and four hysteroscopy procedures were regarded inconclusive for the following reasons: two procedures (one saline infusion sonography and one hysteroscopy) had to be aborted prematurely due to extreme pain, in two cases of saline infusion sonography distension of the uterine cavity could not be accomplished, one cervical stenosis hampered hysteroscopy, for one adipose patient Figure 1 | Flow diagram depicting the outcome of patients recruited for the study.

Assessed for eligibility (n=121)

Enrollment

Randomised (n=100)

Allocation

Analysis

Excluded (n=21)

Not meeting inclusion criteria (n=6)

Refused to participate (n=15)

Allocated to saline infusion sonography (n=50)

• Received allocated intervention (n=50)

• Did not receive allocated intervention (n=0)

Allocated hysteroscopy (n=50)

• Received allocated intervention (n=47)

• Did not receive allocated intervention (n=3): one preg- nancy, two withdrawals

Analysed (n=50)

Excluded from analysis (n=0)

Analysed (n=47)

Excluded from analysis (n=3)

Chapter 6 62

Table 1 | Patients’ characteristics.

Outcome SIS

(n=50)

Office hysteroscopy (n=47)

P-value

Mean age in years (95%-CI) 44.8 (42.1-47.5) 44.4 (42.9-46.9) 0.774

Median parity (range) 2 (0-6) 2 (0-5) 0.137

Nulliparous (range) 12 (24%) 14 (30%) 0.339

Menopausal state 0.554

Premenopausal 42 (84%) 39 (83%)

Postmenopausal 8 (16%) 8 (17%)

Mean BMI (95%-CI) 25.4 (23.8-27.0) 25.2 (23.8-26.6) 0.801

Indication for examination 0.589

Abnormal uterine bleeding

43 (86%) 42 (89%)

Infertility 5 (10%) 5 (11%)

Other 2 (4%) -

Table 2 | Outcome measures; comparison of saline infusion sonography and office hysteroscopy.

Outcome SIS

(n=50)

Office hysteroscopy (n=47)

P-value

Mean VAS for pain scored by patient (95%-CI)

2.7 (2.0-3.4) 4.0 (3.4-4.6) 0.010

VAS for pain >7 1 (2%) 2 (4%) 0.521

PPI for pain (range) 1 (0-5) 2 (0-4) 0.004

PPI for pain >2 6 (12%) 10 (20%) 0.219

Shoulder pain* 0 (0%) 0 (0%) -

Nausea/vomiting* 1 (2%) 1 (2%) 0.965

Dizziness/fainting* 1 (2%) 1 (2%) 0.965

Mean VAS for pain scored by examiner (95%-CI)

2.2 (1.7-2.5) 3.4 (2.7-4.1) 0.010

Confidence score (VAS, range) 9 (0-10) 9 (0-10) 0.899

Confidence score >7 39 (78%) 39 (78%) 0.509

Time of procedure in min (95%-CI)

9.3 (8.0-10.6) 10.1 (8.7-11.4) 0.372

Success rate 47 (94%) 43 (92%) 0.633

*Derivative of the PPI was used to score this item.

Discomfort of hysteroscopy and saline infusion sonography 63

0 5 10 15 20 25 30

0 1 2 3 4 5 6 7 8 9 10

Pain (VAS)

Percentage

saline infusion sonography office hysteroscopy

Figure 2 | Patients’ rating of pain on a visual analogue scale (VAS) associated with saline infusion sonography and office hysteroscopy performed according to the vaginoscopic approach.

0 10 20 30 40 50

none mild discomfort distressing horrible excruciating

Pain (PPI)

Percentage

saline infusion sonography office hysteroscopy

Figure 3 | Patients’ rating of pain on the present pain intensity (PPI) scale associated with saline infusion sonography and office hysteroscopy performed according to the vaginoscopic approach.

Chapter 6 64

Table 3 | The results of a multiple linear regression analysis in a stepwise backward manner to determine the influence of several factors on pain scores.

Factor Coefficient 95%-CI p-value

Saline infusion sonography -1.082 -1.938;-0.225 0.014

Multiparity -0.963 -1.924; 0.003 0.049

Anteflexed uterus -1.164 -2.132; -0.196 0.019

Procedure time 0.204 0.107;0.300 <0.001

Constant 3.488 2.020; 4.955 <0.001

Table 4 | Subsequent therapy following saline infusion sonography and office hysteroscopy.

Therapy Number of patients

Expectative management 36 (37%)

Medical therapy 11 (11%)

Levonorgestrel intrauterine device 13 (13%)

Polyp removal 23 (24%)

Myoma removal 6 (6%)

Vaginal hysterectomy 2 (2%)

Abdominal hysterectomy 2 (2%)

Total laparoscopic hysterectomy 2 (2%)

Hysterectomy and bilateral salpingo-oophorectomy 1 (1%)

(BMI 32 kg/m²) the hysteroscope was too short, and in one case of hysteroscopy a false tract was made.

Twenty-three (46%) of the patients who received saline infusion sonography had no intracavitary abnormalities, whereas sixteen (32%) were diagnosed with endometrial polyps, seven (14%) with submucous myomas, and one (2%) with an intrauterine sep- tum. Twenty-four (51%) patients who received diagnostic office hysteroscopy were diagnosed with no intracavitary abnormalities, thirteen (28%) with endometrial polyps and six (13%) with submucous myomas. In two patients a malignancy was found; one patient was diagnosed with an endometrial polyp, which turned out to be a well dif- ferentiated endometrial carcinoma after hysteroscopic polyp removal. Because of her young age and her desire to have children she was treated conservatively with progesta- gens. The other patient underwent endometrial sampling under general anaesthesia after a failed hysteroscopy and was also diagnosed with an endometrial carcinoma. She was postmenopausal and underwent a hysterectomy with bilateral salpingo-oophorectomy and subsequent radiotherapy. Subsequent therapy of all patients included in the study is depicted in table 4.

Discomfort of hysteroscopy and saline infusion sonography 65

Discussion

This is the first randomised trial comparing saline infusion sonography to office hys- teroscopy performed by a vaginoscopic approach. Our findings confirm that both saline infusion sonography and office hysteroscopy are successful procedures, and are well tolerated by patients. Nevertheless, our study also demonstrates that patients experience significantly less discomfort with saline infusion sonography than with office hystero- scopy. In addition multiparity, a shorter procedure time and an anteflexed uterus influ- enced pain scores in a positive way. Surprisingly, a postmenopausal state was not found of statistical significance. This could be due to the small number of postmenopausal patients included in this study or due to the fact that more than 90% of the postmeno- pausal women were multiparous, which was found to decrease pain scores.

Although office hysteroscopy is more painful, it offers the option of treatment in the same session. Recently a report was published on saline infusion sonography-guided treatment of intrauterine pathology [Wei 2006], but to date only a series of twenty cases has been described, of which five endometrial polyps could be removed successfully.

Since this treatment method is still in its infancy and a proper comparative study is non- existent, it cannot yet compete with the operative palette of office hysteroscopy. Future research will be needed to prove its feasibility and cost-effectiveness.

A limitation of our study is that neither the patients nor the investigators were blinded.

This is unfortunately inherent to research comparing different techniques in an office setting. Another drawback is that different investigators performed saline infusion sono- graphy and office hysteroscopy. However, since these investigations are completely different we do not expect this to influence our results significantly. This might even decrease the influence of our lack of blinding.

The advantage of our study design is that we used the smallest rigid hysteroscopes with continuous flow and an incorporated working channel, which gave us the opportunity to interpret our results from a therapeutic hysteroscopy’s point of view as well. Although not investigated, it is likely that performing biopsies or small surgical interventions will prolong procedure time and increase the number of movements resulting in higher pain scores.

Another advantage of our study is that all patients with an indication for further investi- gation of the uterine cavity were eligible to participate. This gives us a realistic reflection of the actual population.

In addition, the success rates of saline infusion sonography and office hysteroscopy obtained in this study are in line with previously published literature [chapter 5; de Kroon 2003a].

Chapter 6 66

In conclusion, despite the fact that both the introduction of the no touch technique and smaller instruments have made office hysteroscopy a well tolerated investigation [Garbin 2006; Sagiv 2006], saline infusion sonography causes less discomfort. Since still a great deal of clinics do not offer operative hysteroscopy facilities in an outpatient setting [Clark 2002a], saline infusion sonography should remain significant in diagnos- ing intrauterine pathology.