Predictors of Ovarian Torsion: Clinical and Sonographic Characteristics

Abstract

Objective:

Acute pelvic pain is one of the leading reasons for an emergency room visit. Ovarian torsion, in which patients experience pain, is the fifth-most prevalent gynecologic emergency. The condition is difficult to diagnose and requires surgical intervention. This study was performed to assess clinical and sonographic variables and their association with ovarian torsion.

Materials and Methods:

A retrospective cohort of 73 patients with acute pelvic pain had 35 patients with surgically confirmed ovarian torsion and 38 patients with surgically confirmed not ovarian torsion. Multivariate logistic regression was used to examine predictors according to demographic, medical-condition, medical-history, pain, and ovarian-cyst variables for their potential associations with ovarian torsion.

Results:

Variables associated with increased ovarian torsion confirmed by surgery were: intermittent pain (odds ratio [OR]: 3.90; 95% confidence interval [CI]: 1.23–12.35; p = 0.02); pain lasting less than 8 hours (OR: 6.74; 95% CI: 1.94–23.37; p = 0.003); and ovarian cyst ≥ 5 cm (OR: 8.11; 95% CI: 1.21–54.34; p = 0.03). A simple cyst seen on imaging was associated with decreased ovarian torsion (OR: 0.17; 95% CI: 0.04–0.72; p = 0.02). No demographic, medical-condition, and medical-history variables were associated with ovarian torsion.

Conclusion:

Intermittent pain, pain lasting less than 8 hours, and ovarian cyst ≥ 5 cm are frequently seen clinical and sonographic indicators of ovarian torsion. Clinicians who see patients presenting with these symptoms with work-ups that strongly indicate ovarian torsion should bring such patients in for surgical treatment.

Introduction

Acute pelvic pain is one of the leading reasons for emergency visits to gynecology departments, and ovarian torsion, in which patients experience pain, is the fifth-most prevalent gynecologic emergency.¹ Ovarian torsion is the twisting of an ovary on its ligamentous supports; this can result in a compromised blood supply.¹ Ovarian torsion occurs most often in adolescent girls and in women of childbearing age.² Delay in diagnosis and treatment of ovarian torsion can have adverse consequences, including possible functional loss of the ovary.² Ovarian torsion is difficult to diagnose and requires immediate surgery to avoid irreversible ovarian damage.¹ The varied imaging features and nonspecific symptoms of ovarian torsion can lead to a delay in identification, and misdiagnosis often occurs.³ Ultrasonography alone is unreliable for diagnosis.⁴ The only reliable diagnosis of ovarian torsion is laparoscopy.⁵

A number of studies have reported clinical and sonographic findings associated with ovarian torsion. The clinical features of radiating pain, intermittent pain, unilateral abdominal pain, and pain lasting less than 8 hours are each associated with ovarian torsion.^6,7 The sonographic finding of an ovarian cyst ≥5 cm is associated with ovarian torsion.^5–8 Simple cysts are associated with a lower risk of ovarian torsion.⁷

Medical conditions such as vaginal bleeding, leukocytosis, nausea, vomiting, and fever; and previous histories of pelvic surgeries, ovarian cysts, and ovarian stimulation are often considered when studying ovarian torsion. There are mixed findings, with some studies reporting associations while others have not reported any associations.^5–7

Different studies have reported different clinical and sonographic features associated with ovarian torsion. In addition, not all relevant variables associated with ovarian torsion have been studied in the same sample. The purpose of the current study was to include clinical, sonographic, medical-condition, and medical-history variables potentially associated with ovarian torsion and to determine their associations with ovarian torsion, using a multivariate analytical framework.

Materials and Methods

Participants and setting

This was a retrospective cohort study of 73 consecutive patients who presented to the emergency department (ED) with acute pelvic pain and who subsequently had surgery at a suburban New York State hospital from June 1, 2010, to July 31, 2016. Of the 73 patients, 35 had ovarian torsion and 38 did not have ovarian torsion. Exclusion criteria were women who had positive pregnancy tests and who were diagnosed with ectopic pregnancy or spontaneous abortion. Institutional review board approval was obtained.

Variables

The medical records were examined for clinical and sonographic findings that consisted of four different areas: (1) demographics; (2) medical conditions and medical history; (3) pain; and (4) sonographic findings. For demographics, age (years) and race/ethnicity (white or non-white) were included. For medical conditions and history, fever, absence of leukorrhea and metrorrhagia, vaginal bleeding, nausea/vomiting, leukocytosis, history of pelvic surgery, history of ovarian cysts, and history of ovarian stimulation were included. These were all categorized as “no/yes.” For pain, intermittent pain, radiating pain, unilateral abdominal pain, and pain lasting less than 8 hours were included. These were all categorized as “no/yes.” For ovarian cysts, ovarian cyst ≥ 5 cm and simple cyst on imaging were included. These were all categorized as “no/yes.” The existing medical records that were part of the patients' clinical treatment for sonogram interpretation and ovarian-torsion diagnosis were used. A radiologist reviewed the sonograms in the medical records to measure the size of each ovarian cyst and its morphologic appearance. The diagnosis of ovarian torsion was surgically confirmed. In the surgical reports, the number of pedicle twists of the torsed ovaries were recorded.

Statistical analysis

Means and standard deviations were used to characterize the continuous age variable. Percent and frequency were used to express the categorical variables. Analysis of variance was used to compare the continuous variable to ovarian torsion. As appropriate, either Pearson's χ² test or Fisher's exact test (when any cell was <5) was used to compare the categorical variables to ovarian torsion. All variables that were statistically significant in the univariate analyses were included in a multivariate logistic regression analysis for the outcome of the presence of ovarian torsion. IBM SPSS Statistics (version 24) was used for all analyses. All p-values were two-sided.

Results

Table 1 shows the sample characteristics, compared with ovarian torsion. With regard to the pain variables, patients with ovarian torsion had a statistically significant greater percentage of intermittent pain and pain lasting less than 8 hours than patients without ovarian torsion. With regard to the ovarian torsion and cyst variables, patients with ovarian torsion had a statistically significant greater percentage of ovarian cysts ≥ 5 cm than patients without ovarian torsion. In addition, patients with ovarian torsion had a statistically significant lower percentage of simple cysts noted on imaging than patients without ovarian torsion. Both radiating pain and unilateral abdominal pain did not differ statistically. None of the demographic variables differed statistically. None of the medical conditions and history variables differed statistically.

Table 1.

Characteristics of the Sample

Variable	No ovarian torsion M (SD) or % (#)	Yes ovarian torsion M (SD) or % (#)	p-Value
(n = 38)	(n = 35)
Demographics
Age (years) [mean]	33.6 (11.25)	31.2 (12.49)	0.40
Race/ethnicity (non-white)	65.8% (25)	74.3% (26)	0.43
Medical conditions & history
Fever	0.0% (0)	5.7% (2)	0.23
Absence of leukorrhea & metrorrhagia	100.0% (38)	94.3% (33)	0.23
Vaginal bleeding	10.5% (4)	8.6% (3)	0.78
Nausea/vomiting	34.2% (13)	48.6% (17)	0.21
Leukocytosis	28.9% (11)	42.9% (15)	0.22
History of pelvic surgery	50.0% (19)	42.9% (15)	0.54
History of ovarian cyst	2.6% (1)	0.0% (0)	1.00
History ovarian stimulation	0.0% (0)	0.0% (0)	N/A
Pain
Intermittent pain	31.6% (12)	65.7% (23)	0.004
Radiating pain	23.7% (9)	20.0% (7)	0.70
Unilateral abdominal pain	94.7% (36)	91.4% (32)	0.67
Pain lasting less than 8 hours	44.7% (17)	80.0% (28)	0.002
Ovarian cyst
Ovarian cyst ≥ 5 cm	76.3% (29)	94.3% (33)	0.048
Simple cyst noted on imaging	31.6% (12)	11.4% (4)	0.049

M, mean; SD, standard deviation; N/A, not applicable.

Table 2 shows the multivariate logistic regression analysis for the outcome variable of the presence of ovarian torsion. The pain variables of intermittent pain and pain lasting less than 8 hours each were statistically significantly associated with increased odds for ovarian torsion. Ovarian cysts ≥ 5 cm were statistically significantly associated with increased odds for ovarian torsion. Simple cysts noted on imaging were statistically significantly associated with decreased odds for ovarian torsion.

Table 2.

Multivariate Logistic Regression Analysis for Presence of Ovarian Torsion

Variable	OR (95% CI)	p-Value
Intermittent pain	3.90 (1.23–12.35)	0.02
Pain lasting less than 8 hours	6.74 (1.94–23.37)	0.003
Ovarian cyst ≥ 5 cm	8.11 (1.21–54.34)	0.03
Simple cyst noted on imaging	0.17 (0.04–0.72)	0.02

OR, odds ratio; CI, confidence interval.

Discussion

Clinical pain and sonographic-cyst variables were associated with ovarian torsion in the current study. Variables associated with increased odds for ovarian torsion were: intermittent pain; pain lasting less than 8 hours; and ovarian cysts ≥ 5 cm. Simple cysts on noted on imaging were associated with decreased odds for ovarian torsion.

Intermittent pain was associated with increased odds for ovarian torsion. This was consistent with multiple studies.^6–8 Intermittent pain is a common symptom of ovarian torsion, as the episodes of pain can be separated by asymptomatic intervals when the ovary spontaneously torses and detorses.^6–8

Pain lasting less than 8 hours was associated with increased odds for ovarian torsion. This was consistent with the results of another study.⁶ Earlier studies found that the time from pain onset to a visit to a physician was short.^9,10 Age older than 40 and a delay to surgery longer than 10 hours were significantly associated with a diagnosis of adnexal necrosis.¹¹ The current authors suggest that pain usually lasts less than 8 hours when ovarian torsion occurs because patients seek medical attention sooner rather than later.

Ovarian cysts ≥ 5 cm were associated with increased odds for ovarian torsion. This is consistent with multiple studies showing that ovarian cysts ≥ 5 cm were associated with ovarian torsion.^4,6,12 The current study's findings support the theory that a large cyst can weigh an ovary down and predispose it to torsion.

Simple cysts on imaging were associated with decreased odds for ovarian torsion.⁷ A large cystic neoplasm, such as an hemorrhagic cyst, commonly predispose the ovary to swing on its vascular pedicle.⁷ The reasoning is that a hemorrhagic cyst can act as a lead point, weighing the ovary down and predisposing it to torsion.⁷ In up to 73% of cases, there is a complex adnexal or pelvoabdominal mass that can be predominantly cystic, solid, or both.¹³ The current authors suggest that a simple cyst has a decreased risk of ovarian torsion, because there is no solid component in such a cyst to weigh the ovary down and predispose it to torsion.

Radiating pain and unilateral abdominal pain were not associated with ovarian torsion. There have been conflicting results in studies using these variables.^6,7 The current study's results were consistent with the studies that had reported no associations.

The following clinical medical conditions and history variables were not associated with ovarian torsion: fever; absence of leukorrhea and metrorrhagia; vaginal bleeding; nausea/vomiting; leukocytosis; history of ovarian cyst; and history of ovarian stimulation. There have been reports of conflicting associations of ovarian torsion with these variables. The current study's results were consistent with the studies that reported no associations.^5–7

There were no significant associations with history of pelvic surgery with ovarian torsion. Patients with histories of pelvic surgery, especially tubal ligation, are potentially at increased risk for ovarian torsion. A previous tubal ligation can cause hypermobile adnexa, inducing torsion at the mesosalpinx.⁴ Another possible reason for increased risk of torsion is that adhesions provide a site around which the ovarian pedicle can twist. In the current study, history of pelvic surgery occurred in 43% of cases (15/35 patients), which was comparable to the rates found in other studies.^4,6 However, the current study did not find any significant difference for history of pelvic surgery between patients with ovarian torsion and without ovarian torsion. These results were consistent with studies that have reported no associations between history of pelvic surgery and ovarian torsion.^4,6

This study was limited by a small sample size. In addition, only patients who underwent surgery were included. Thus, the current authors did not have knowledge of the diagnoses of patients who presented to the ED with abdominal pain that was suspicious for ovarian torsion but were discharged without interventions. However, all patients with undiagnosed ovarian torsion experienced severe and unrelenting pain due to ovarian necrosis and, therefore, would have returned at a later time to the ED for assessment.

Conclusions

Intermittent pain, pain lasting less than 8 hours, and ovarian cysts ≥ 5 cm are frequently shown clinical and sonographic indicators that are associated with increased odds for ovarian torsion. A simple cyst noted on imaging is associated with decreased odds for ovarian torsion. Clinicians seeing patients presenting with intermittent pain, pain lasting less than 8 hours, and ovarian cysts ≥ 5 cm and whose work-ups strongly indicates the presence of ovarian torsion should bring these patients to the operating room for further treatment.

Footnotes

Author Disclosure Statement

No financial conflicts of interest exist.

References

Ashwal

, Hiersch

, Krissi

, et al. Characteristics and management of ovarian torsion in premenarchal compared with postmenarchal patients. Obstet Gynecol, 2015; 126:514.

Bar-On

, Mashiach

, Stockheim

, et al. Emergency laparoscopy for suspected ovarian torsion: Are we too hasty to operate?. Fertil Steril, 2010; 93:2012.

Shadinger

, Andreotti

, Kurian

. Preoperative sonographic and clinical characteristics as predictors of ovarian torsion. J Ultrasound Med, 2008; 27:7.

Pena

, Ufberg

, Cooner

, Denis

. Usefulness of Doppler sonography in the diagnosis of ovarian torsion. Fertil Steril, 2000; 73:1047.

Asfour

, Varma

, Menon

. Clinical risk factors for ovarian torsion. J Obstet Gynaecol, 2015; 35:721.

Huchon

, Staraci

, Fauconnier

. Adnexal torsion: A predictive score for pre-operative diagnosis. Hum Reprod, 2010; 25:2276.

Appelbaum

, Abraham

, Choi-Rosen

, Ackerman

. Key clinical predictors in the early diagnosis of adnexal torsion in children. J Pediatr Adolesc Gynecol, 2013; 26:167.

Damigos

, Johns

, Ross

. An update on the diagnosis and management of ovarian torsion. Obstet Gynaecol, 2012; 14:229.

Nichols

, Julian

. Torsion of the adnexa. Clin Obstet Gynecol, 1985; 28:375.

10.

Goldstein

. Acute and chronic pelvic pain. Pediatr Clin North Am, 1989; 36:573.

11.

Mazouni

, Bretelle

, Menard

, Blanc

, Gamerre

. Diagnosis of adnexal torsion and predictive factors of adnexal necrosis. Gynecol Obstet Fertil, 2005; 33:102.

12.

Bouguizane

, Bibi

, Farhat

, et al. Adnexal torsion: A report of 135 cases. J Gynecol Obstet Biol Reprod (Paris), 2003; 32:535.

13.

Albayram

, Hamper

. Ovarian and adnexal torsion: Spectrum of sonographic findings with pathologic correlation. J Ultrasound Med, 2001; 20:1083.