The standard examination of the pelvis includes estimating the overall uterine size. Although the normal size of the outer uterine contour has been established for both premenopausal and postmenopausal women,1,2 the ultrasonographic width of the uterine cavity itself is an important parameter not yet explored.
Recently, our laboratory reported on a series of patients with intrauterine devices (IUDs) in situ and demonstrated that 18% have the arms of the IUD malpositioned or embedded within the myometrium or cervix.3 This previous study described a significantly higher incidence of pain and or bleeding in patients with embedded IUDs compared with those whose IUDs were in a normal position in the uterine cavity. In that study, we showed that the mean width of the uterine cavity, measured using a reconstructed three-dimensional coronal view of the uterus, was significantly narrower (25 mm) in patients whose IUDs were embedded as compared with those whose IUDs were normally positioned (32 mm) (Shipp TD, Bromley B, Benacerraf BR. The width of the uterine cavity is narrower in patients with an embedded IUD as compared to a normally positioned IUD. J Ultrasound Med. In press).
The aim of the current study is to estimate the width of the normal uterine cavity at the fundus and correlate this width with parity, gravidity, age, prior cesarean delivery, and uterine volume to generate sonographic data for the normal measurement of the uterine cavity.
MATERIALS AND METHODS
Institutional Review Board approval was obtained from Brigham and Women's Hospital for this study involving the review of medical records. Our protocol for all patients undergoing gynecologic ultrasonography includes three-dimensional volume acquisitions of the uterus, acquired both longitudinally and transversely, and saved on our picture archiving and communication system.
From September 1, 2009, to November 10, 2009, 221 consecutive premenopausal patients aged 50 years or younger who were referred for pelvic ultrasonography and who had a normal uterus on two-dimensional standard ultrasonography were included in this retrospective study. Patients with any finding such as fibroids, polyps, presence of an IUD, or uterine anomaly were excluded. Standard measurements of the outer uterine length, width, and anteroposterior diameter as well as standard endometrial width were performed transvaginally at the time of the scan. Information such as gravidity, parity, current oral contraceptive use, and previous cesarean delivery was available to the sonographer. After the sonographer performed the standard gynecologic sonographic examination, two three-dimensional volumes of the uterus were acquired using a three-dimensional, transvaginal, 5- to 8-MHz probe and archived on the picture archiving and communication system (ViewPoint; GE Medical, Milwaukee, WI). One volume was taken longitudinally, acquiring the volume from one side to the other of the uterus, with the entire length of the uterus visible during the acquisition. A second volume was taken in transverse section across the uterus, acquiring the volume from the cervix to the fundus. For both volumes, the settings were such that the entire uterus was included on each acquisition. A single sonologist (B.R.B.) with at least 10 years of experience with three-dimensional ultrasonography reviewed each volume (two per patient) retrospectively on the picture archiving and communication system (ViewPoint; GE Medical) to obtain the reconstructed coronal view of the uterine cavity. The planes were manipulated to obtain the reconstructed coronal view of the uterine cavity at the widest point according to the z-plane technique described by Abuhamad et al.4 The sonologist then measured the width of the uterine cavity from cornu to cornu on each coronal view (Fig. 1). An attempt was made to obtain the widest possible measurement across the top of the uterine cavity. Having accomplished this measurement on the longitudinal acquisition, the same process was repeated on the transverse acquisition of the uterus for each patient (see video online at http://links.lww.com/AOG/A182). Because not all patients had adequate volumes to reconstruct, some patients had to be excluded from the study and others only had one satisfactory volume to use rather than both. A single sonologist performed all the measurements; however, a second sonologist (B.B.) with at least 8 years of experience with three-dimensional imaging independently manipulated the volumes again for the first 30 patients and remeasured the width of the uterine cavity so as to evaluate interobserver variability. The original sonologist (B.R.B.) also independently manipulated the volumes of these first 30 patients once again and remeasured the width of the uterine cavity to estimate the intraobserver variability.
The volume of the outer contour of the uterus was calculated for each patient based on the formula for a prolate ellipsoid (volume=[4/3] [π] [1/2 uterine length] [1/2 anteroposterior diameter] [1/2 transverse diameter]) and using the standard three uterine measurements obtained transvaginally at the time of the original scan.1,2
Analysis of the data was done to compare the two methods of measuring the width of the uterine cavity using the two different volume acquisitions at right angles to each other. Means and standard deviations were calculated for the continuous variables (measured width of the endometrial cavity, uterine volume, endometrial thickness, and age); frequencies were calculated for the patient gravidity, parity, and previous cesarean delivery. These variables were further compared by gravidity and parity category and evaluated using F tests to compare means and χ2 tests to test distributions by gravidity and parity groups. Pearson correlation coefficients were calculated for all variables. Univariable regression analysis was also performed for the noted variables and the width of the endometrial cavity but restricted to women who had never been pregnant to reduce confounding by gravidity and parity. Both intraobserver and interobserver variability were calculated for the assessment of the uterine width. P<.05 was considered statistically significant.
Among our 221 patients, 11 had volumes that were deemed inadequate to measure the width of the endometrial cavity as a result of very thin endometrium or high body mass index. Table 1 shows the demographics of the remaining 210 patients who formed the study cohort. Table 2 shows the indications for the ultrasonographies. The transverse and longitudinal acquisitions were adequate for reconstruction of the coronal image on 184 (87.6%) and 187 (89.0%) of our 210 patients, respectively, although all 210 patients had at least one volume that provided an adequate coronal view of the uterus. One hundred fifty-four patients had measurements done both on transverse and longitudinal volumes and these were compared (Fig. 2). There was no difference in measurement when the width of the uterine cavity was measured using a transverse compared with a longitudinal volume acquisition (P=.904, using the 154 patients who had both measurements). The first 30 patients had their volumes (acquired in transverse orientation) remanipulated and the width of the uterine cavity remeasured by a second observer and then again by the original observer. Based on these 30 patients, the interobserver and intraobserver variability were 0.96 (95% confidence interval 0.91–0.98) and 0.53 (95% confidence interval 0.21–0.74) respectively. As seen in Tables 3 and 4, the width of the uterine cavity and the uterine volume increase significantly with both increasing gravidity and parity. Both the longitudinal and transversely obtained uterine volume measurements were highly correlated (Table 5). There was no statistically significant relationship between prior cesarean delivery and width of the uterine cavity. As expected, there was a strong correlation between the width of the uterine cavity and the overall uterine volume as well as width of the endometrial echo.
Patient age varied significantly with gravidity and parity as well as with uterine measurements (uterine cavity width and overall volume). We evaluated the relationship between patient age and uterine cavity width by excluding patients with previous pregnancies from the linear regression model. Among patients who were never pregnant, age did not predict uterine cavity width (P=.384). It is likely that the relationship between patient age and uterine cavity width initially seen was the result of increasing number of pregnancies rather than age itself.
There were 29 (13.8%) patients on oral contraceptives. In the absence of other variables, the patients on oral contraceptives had significantly smaller uterine cavities by transverse measurement (mean 25.52, standard deviation 5.09) compared with those not on hormones (mean 29.24, standard deviation 7.08) (P=.016). Because a higher proportion of nulliparous women were taking oral contraceptives (χ2=4.34, P=.037), the observed difference in mean uterine size could be a result of parity instead of contraceptive use alone.
Investigators have studied the size of the outer dimensions of the uterus, both in linear and volume measurements, and have shown that these vary with parity.1,2,5 Although our data support this, our study differs in that we studied the width of the uterine cavity rather than the outer dimensions of the uterus. Our data demonstrate that the width of the uterine cavity varies with gravidity and parity, although not with prior cesarean delivery or age. The width of the uterine cavity ranged from a mean of 27 mm in nulliparous women to 32 mm in those with more than one pregnancy.
These numbers compare favorably with those of our prior study, which involved a different group of women who had an IUD in place (Shipp TD, et al. J Ultrasound Med. In press).4 In that prior study, the patients with an IUD in the proper location had a mean uterine cavity width of 32 mm at the fundus compared with 25 mm for those who had embedded IUDs. The prior study suggests that the malpositioning of the IUD within the uterus may be responsible for the higher incidence of pain and bleeding in these patients compared with those with normally positioned IUDs. These findings prompted us to evaluate the normal uterine cavity width and how this varies with obstetric history. The width of the standard, commercially available IUDs (Mirena [Bayer Health Care Pharmaceuticals, Finland] and ParaGard [Duramed Pharmaceuticals, Pomona, NY]) inserted routinely is 32 mm. Our data shows that not all uteri are large enough to accommodate the standard size IUD and it may be necessary to provide more than one size of IUD. Most IUDs are inserted blindly with only a rough estimate of the size of the uterus judged by bimanual examination but with no knowledge of the uterine cavity width or shape. Our data show that performing volume sonography to measure the size of the uterine cavity may be beneficial before inserting an IUD.
Our report had several limitations. As a result of the retrospective nature of the study, we were not able to obtain reliable information about the menstrual cycle from most of these women and therefore measured the uterine cavities regardless of the menstrual cycle. This may have had an effect on the endometrial echo measurements, which in turn seemed to vary with the width of the uterine cavity. We also did not perform a formal justification for the number of patients that we included in this primarily descriptive study. Perhaps a larger study can be done to confirm our observations. Also, a single practitioner performed all the measurements of the uterine cavity. However, to address this potential bias, we had a second practitioner independently remanipulate a subset of the volumes to determine the interobserver variability. Last, we relied on an established formula to calculate the uterus volume based on the three standard two-dimensional measurements obtained during the transvaginal scan. We did not test the validity of this calculation against an actual volume of the uterus as determined from the three-dimensional volumes because this was not the aim of our study.
Our data establish the width of the normal uterus and show that standard IUDs manufactured today are wider when open than the majority of uterine cavities in nulliparous patients, suggesting that physicians should consider ultrasonography to measure the uterine cavity before inserting an IUD. Further study is needed to determine whether a three-dimensional volume measurement of the width of the uterine cavity should be performed on patients, particularly those nulliparous, before inserting an IUD and whether smaller IUDs should be made available for these women. These standard measurements of the uterine cavity may also be helpful in preparation for other uterine procedures such as endometrial ablation and insertion of other contraceptive devices.
1. Merz E, Miric-Tesanic D, Bahlmann F, Weber G, Wellek S. Sonographic size of the uterus and ovaries in pre- and postmenopausal women. Ultrasound Obstet Gynecol 1996;7:38–42.
2. Kung FT, Chang SY. The relationship between ultrasonic volume and actual weight of pathologic uterus. Gynecol Obstet Invest 1996;42:35–8.
3. Benacerraf BR, Shipp TD, Bromley B. Three-dimensional ultrasound detection of abnormally located intrauterine contraceptive devices which are a source of pelvic pain and abnormal bleeding. Ultrasound Obstet Gynecol 2009;34:110–5.
4. Abuhamad AZ, Singleton S, Zhao Y, Bocca S. The Z technique: an easy approach to the display of the mid-coronal plane of the uterus in volume sonography. J Ultrasound Med 2006;25:607–12.
5. Bartoli JM, Moulin G, Delannoy L, Chagnaud C, Kasbarian M. The normal uterus on magnetic resonance imaging and variations associated with the hormonal state. Surg Radiol Anat 1991;13:213–20.