To document normal measurements (length, width, anteroposterior (AP) diameter) and proportions of the non-pregnant uterus according to age and gravidity. We hypothesized that uterine proportions conform to the classical ‘golden ratio’ (1.618).
This was a retrospective study of ultrasonographic measurements of the length, width and AP diameter of non-pregnant uteri recorded in our database between 1 January 2000 and 31 July 2012. All patients for whom abnormal findings were reported were excluded and only the first set of measurements for each patient was retained for analysis. Loess (local regression) analysis was performed using age and gravidity as explanatory variables.
Measurements of 5466 non-pregnant uteri were retrieved for analysis. The mean length was found to increase to 72 mm at the age of 40 and decrease to 42 mm at the age of 80 years. Gravidity was associated with greater uterine length, width and AP diameter. Mean length/width ratio was found to be 1.857 at birth, decreasing to 1.452 at the age of 91 years. At the age of 21 years, the mean ratio was found to be 1.618, i.e. equal to the golden ratio. Increasing gravidity was associated with lower mean length/width ratio.
The normal anatomical appearance of the uterus is well known to gynecologists and can be recognized immediately on ultrasound examination or direct vision. Accordingly, any deviation from normal appearance is usually easily detected. However, very few normative data regarding uterine dimensions have been published.
In human anatomy, the normal appearance of a number of structures has been found to conform to the classical ‘golden ratio’. For example, the golden ratio has been applied to facial features and has also been found to apply when describing the anatomy of the heart. Historically the golden ratio was used in relation to architecture and the arts, as objects with proportions based on the ratio were perceived to be more beautiful. This ratio has also been found to exist within natural structures, such as the seeds of sunflowers. These findings are somewhat controversial, as the distributions of observed proportions only approximate the golden ratio in most cases.
The golden ratio (also known as the golden section, golden mean or divine proportion) is a ratio of two lengths, for which the relative size of the longer of the two compared to the shorter equals their total dimension compared to the longer (denoted by φ = (1 + √5)/2 ≈ 1.618). It can also be derived from the Fibonacci sequence (0, 1, 1, 2, 3, 5, 8, 13, 21, 34 …), in which each number is made up of the sum of the previous two numbers; when dividing a number by the number just before it, the ratio converges to 1.618 as the sequence continues.
Abnormal uterine proportions may have relevance for the detection of disease or physiological dysfunction. For example, uterine myometrial asymmetry is a well recognized marker of the presence of adenomyosis. Furthermore, disproportion of the uterus may reflect myometrial dysfunction and thus be associated with subfertility. Although normative data for uterine length and volume have been described for girls prior to puberty[4, 5], there is a paucity of data in relation to uterine proportions for other age groups. We aimed to define normal values for uterine length, width and anteroposterior (AP) diameter for a wide range of ages. We hypothesized that uterine proportions conform to the classical golden ratio.
We conducted a retrospective study using ultrasonographic measurements of all consecutive non-pregnant uteri recorded in the electronic database of Leuven University Hospital from 1 January 2000 to 31 July 2012. Approval of the Ethics Committee of the University Hospitals Leuven was obtained.
All data were recorded on a computer database (Astraia, Munich, Germany) and subsequently entered onto an Excel (Microsoft Corp., Redmond, WA, USA) spreadsheet for statistical analysis. All patients for whom abnormal findings were reported, e.g. myomas, adenomyosis, adhesions or cysts, were excluded and only measurements from uteri described as ‘normal’ were used. The database was checked for incomplete and false entries. Only the first set of measurements for each patient was retained for statistical analysis. No restrictions were placed on patient age. The study population was predominantly Caucasian.
All scans were performed by gynecologists specialized in gynecological ultrasonography or by gynecologists in training, in which case all scans were supervised by an experienced consultant. Ultrasonographic measurements were carried out with high-end ultrasound equipment using a transabdominal (2.5–5-MHz) or transvaginal (5–9-MHz) transducer for B-mode imaging, with an Acuson Sequoia (Siemens-Acuson Inc., Mountain View, CA, USA), or a Voluson 730, Voluson E8 or Voluson E6 (GE Medical Systems, Zipf, Austria) or an ESAOTE Technos (Esaote, Genova, Italy).
Every uterine assessment started with identification of the bladder and the cervix. Uterine position was noted. The uterus was scanned in the sagittal plane from cornu to cornu and in the (oblique) transverse plane from the cervix to the fundus and back. All measurements were obtained on transvaginal imaging except for patients in whom transvaginal examination was not possible (e.g. in virgin females or in case of patient refusal). Uterine length was measured from the external cervix to the fundus in the sagittal plane. If the angle between the cervical canal and the uterine corpus was more than c. 30°, the cervix and uterine corpus were measured separately and the sum was taken as the uterine length. The AP diameter was also measured in the sagittal plane from the anterior serosa to the posterior serosa at the point at which the uterus appeared at its thickest and perpendicular to the endometrial line. Uterine width was measured from the right to the left side of the uterine corpus in the transverse plane where the width of the corpus appeared at its thickest. The uterine length/width (L/W) ratio was assessed for each patient.
The relationship between length, width, L/W ratio and AP diameter of the uterus and patient age was assessed with Loess smoothing analysis, which is a non-parametric local regression technique. We used locally quadratic functions and a smoothing parameter of 1 to estimate the mean by age, and 95% confidence limits for the estimated mean according to age were calculated. We also drew boundary lines symmetrically around the mean curve such that 95% of the data points were included within these boundaries. This enabled us to identify patients with measurements outside the normal 95% range. Differences in L/W ratio and age between women with and without information on previous pregnancies were assessed using Wilcoxon's test. Statistical analyses were performed using the SAS System version 9.3 (SAS Institute Inc., Cary, NC, USA).
Measurements from 6880 scans of 5482 women and girls were retrieved from the database. A total of 5466 sets of measurements were retained, as only the first scan from each woman in the database was used and 16 outliers were identified for which measurements were probably wrongly entered into the database (e.g. uterine length of 712 mm). The age range of patients included was < 1–91 years.
Raw data for observed values of L/W ratio in relation to age, along with the mean, are shown in Figure 1a. The mean ratio at birth was c 1.86 and decreased over time to 1.45 after menopause. Due to the large number of patients, the 95% confidence limits are very narrow around the estimated mean, except at the extremities of the curve. The mean L/W ratio in relation to age and the upper and lower boundaries for the 95% range of observations are shown in Figure 1b. By the age of 21 years, the mean uterine L/W ratio was 1.628 (95% confidence limits, 1.609–1.644) and by the age of 22 years the mean ratio was 1.615 (95% confidence limits, 1.559–1.631), i.e. at the age of 21 years the mean L/W ratio crossed 1.618 (Table 1). A schematic representation of uteri with different ratios is shown in Figure 2.
Table 1. Mean uterine length/width (L/W) ratio at ages 20–23 years with 95% lower and upper confidence limits derived from Loess analysis
Mean L/W ratio (95% confidence limits)
n, number of women per age group.
Information about gravidity was available in fewer than half of the patients; among those for whom this information was recorded, 682 patients were nulligravid, 420 patients had one previous pregnancy and 1134 patients had two or more previous pregnancies. There was no statistically significant difference between patients with and without information on gravidity regarding uterine L/W ratio (median, 1.523 vs 1.535; P = 0.97), but patients with information on gravidity were older (median, 39 vs 36; P < 0.001).
When stratified by gravidity, we observed a marked decrease in the L/W ratio as the number of pregnancies increased, indicating that the uterus becomes more flat and rounded. Numbers were only comparable after the age of 25 years, because very few women had experienced one or more pregnancies before that age (Figure 3a). Curves for mean L/W ratio according to gravidity converge by the age of 40 years regardless of the number of pregnancies, and diverge again after menopause. Curves for mean uterine length, width and AP diameter by gravidity and in relation to age are shown in Figures 3b–d. Mean uterine length increased up to the age of 40 years, from 3 cm at birth to 7.5 cm and became shorter again afterwards, shrinking to 3 cm by the age of 90 (Figure 4).
This is the first study to describe normative data for uterine length, width and AP diameter, extracted from a large population of women and girls with an apparently normal uterus. Furthermore, we analyzed the data in relation to age and gravidity. Our results show that the uterus undergoes changes in the size of all three dimensions during a woman's lifespan. The length increases to a mean of 7 cm and decreases again after the age of 40, suggesting that mechanisms other than menopausal estrogen depletion account for this shrinkage. Increasing gravidity is associated with greater uterine length, and this difference persists throughout life.
The uterus becomes more round with age, with the mean L/W ratio decreasing from 1.857 to 1.452. This observation is not true for women who have never conceived, as the mean curves according to gravidity diverge after menopause, with a ratio of 1.618 by the age of 58 years in this group. This may suggest that some other mechanism takes place in the uterus, making it narrower rather than shorter in nulligravid women. Collagen composition has been reported to increase with age and decrease with parity. Furthermore, collagen stiffness has been found to increase with age. These factors may be responsible for the altered pattern in curves of the ratio for women with differences in gravidity.
A strength of our study is that only one set of measurements per woman was used for the analysis, thus eliminating possible codependency. Furthermore, the study population was large enough to allow stratification of the data according to gravidity. A weakness of the study is that it was retrospective. However, throughout the study period, measurements from the uterus were recorded in a standardized way in all cases using an electronic reporting system incorporating a structured database.
A limitation of the study is that we have not been able to examine the impact of factors other than age and gravidity that could influence uterine dimensions, such as hormonal therapy, uterine prolapse or ethnicity. These data were not collected with sufficient consistency or in enough detail to enable any useful analysis. Whilst the population attending the clinic was mainly Caucasian, we do not routinely collect data on the ethnicity of patients.
As the mean L/W ratio was found to be 1.618 at the age of 21, it is tempting to speculate that there is a relationship between uterine dimensions and optimal fertility. Fecundity is known to be related to age, decreasing rapidly after the age of 30[8-10]. In an historical series of 209 Hutterite women, amongst whom first conception occurred by the age of 22 in most cases, the infertility rate was only 2.4% . As the golden ratio is observed at the age at which women are most fertile, one can question whether an optimally fertile uterus at any age would demonstrate proportions with a ratio of 1.618. This study did not investigate whether young women with subfertility have uterine dimensions that deviate from the golden ratio, but it is interesting to observe that at the time of peak fertility the uterus conforms to this classical concept of ‘perfection’.
Adolescent girls do not have a uterus in conformance with the golden ratio. Similarly, in the field of facial esthetics, no significant relationship with the golden ratio was found in adolescents as compared to adults. It seems that, anatomically, there is maturation in the proportion of structures throughout adolescence.
Gynecologists may feel capable of recognizing a normal uterus without resorting to measurements or further investigations. It is possible that the presence of the idealized golden ratio hidden in normal uterine anatomy is a factor contributing to this. The ubiquitous presence of the golden ratio in art and culture may have primed our ability to recognize ‘ideal’ proportions and thus conclude that an organ is normal.
Normative curves for uterine dimensions throughout life are important to establish normality and to detect aberrant uterine morphology and size. Our findings could stimulate further research on uterine dimensions and proportions as anatomical, functional and potentially prognostic markers of fertility.
T.B. is supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Imperial College Healthcare NHS Trust and Imperial College London. The views expressed here are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.