Ovarian torsion: determining the presenting features and where the delays occur

Ovarian torsion is a gynaecological emergency caused by the twisting of the ovary around its supporting ligaments, causing ischaemia and, ultimately, necrosis.^1,2 Ovarian torsion management is time-critical, as delays are more likely to result in necrosis and a subsequent decrease in fertility. Patients with ovarian torsion are most commonly of reproductive age and present with acute unilateral pelvic pain, nausea, vomiting and ovarian masses greater than 5cm.^3,4 Symptoms of fever and peritonism are less common according to the literature; however, it remains a widespread misconception that ovarian torsion must present with severe signs.^3–5 This can result in misdiagnosis and delays in treatment in those with mild pain or intermittent symptoms.⁵ Clinically diagnosing ovarian torsion is challenging, as its common features align similarly with numerous surgical diagnoses like appendicitis, pelvic inflammatory disease (PID), ectopic pregnancies and nephrolithiasis,⁶ delaying management. Treatment of ovarian torsion is surgical and includes ovary-sparing procedures, including detorsion, +/- cystectomy or oophorectomy.¹ The choice of procedure is largely influenced by the presumed viability of the ovary determined by direct visualisation.¹

Waitematā District Health Board (DHB) includes North Shore Hospital (663 beds) and Waitākere Hospital (283 beds). North Shore Hospital is a secondary hospital and the only hospital of the two that provides acute gynaecological services. An audit of complications by the North Shore gynaecology department reports 1–2 cases of delayed management of adnexal torsion monthly, potentially resulting in unnecessary oophorectomy. This study’s purpose is to assist in formulating hospital guidelines that enable clinicians to better recognise and manage ovarian torsion cases. The study shall firstly present the most common symptoms, signs and findings seen among ovarian torsion cases. Secondly, it shall examine the time frames in the clinical management of ovarian torsion to identify where delays in management occur, potentially resulting in avoidable oophorectomy.

Methods

A retrospective audit of 2-year electronic records was performed for all patients coded for ovarian torsion between 01/05/2022 and 01/05/2024 at one New Zealand hospital. Eighty-four patients met this criterion. Thirty-nine of these patients were excluded from this analysis as they were incorrectly coded as having ovarian torsion, but rather had isolated fallopian tube torsion, symptomatic ovarian cysts, no evidence of torsion intraoperatively or were incorrectly coded for ovarian torsion. One patient was initially seen in a clinic environment for an already known necrotic ovarian torsion and was treated with elective surgery. As this was not an acute case like the others, it was also excluded. Patients included in the final analyses were given a diagnosis in their discharge summary, confirmed by intraoperative visualisation, leaving a sample population of 45.

The following variables were extracted from the electronic records: age, ethnicity, menopausal status, presenting symptoms, physical examination findings, surgical procedure and the time periods between different time points of clinical management. These time points were symptom onset, initial healthcare presentation (either to their general practitioner [GP] or the emergency department [ED]), initial hospital clinician review, diagnostic imaging, gynaecology diagnosis and finally theatre. These variables were collected from admission notes, discharge summaries, operation notes, radiology reports, clinic letters and clinical notes. These data were approved for use within the scope of this research by The University of Auckland Human Participants Ethics Committee (reference number 021825). The sample population was further subdivided into two groups. The control group were those that underwent laparoscopic or open ovarian detorsion +/- cystectomy, while the case group underwent laparoscopic or open oophorectomy.

Results

Presenting symptoms and signs

During the 24 months, there were 45 cases of ovarian torsion at a New Zealand tertiary hospital. The patient characteristics and the prevalence of the main symptoms reported among the case and control groups are shown in Table 1 and 2. For patients presenting with torsion, the most common procedure performed was an oophorectomy, at 57.8%. This is similar to audits at other hospitals, with oophorectomy rates between 47–64.3%.^7–9 The most common presentation was a premenopausal woman experiencing acute unilateral lower abdominal pain with nausea, vomiting and concurrent ovarian cysts. This is consistent with the standard “textbook” presentation of ovarian torsion.

Right-sided pain was more common than left, likely due to the proximity of the right ovary to the more mobile ileum and caecum as opposed to the relatively fixed sigmoid colon on the left.⁶ Women who presented with right-sided pain were more likely to need an oophorectomy. This was likely due to the propensity to assume acute right lower quadrant pain is of surgical cause, such as appendicitis. This is supported by the fact that 44% of women with right lower quadrant pain were initially reviewed by general surgery, delaying gynaecology review.

Another pertinent finding was that 22.2% of the sample population had presented to a hospital or a clinic within a year before their torsion for symptomatic ovarian cysts. These women were much more likely to require oophorectomy. Peritonism with signs of guarding, percussion tenderness and rebound tenderness were rare findings.

View Table 1–2, Figure 1–2.

Management of ovarian torsion

The clinical journey of ovarian torsion management, as defined in the methods, was consistent across patients. Tracking each patient’s progress through a systematic set of time points allowed for easier comparison and identification of delays. Obtaining electronic clinical records ensured accuracy in these time frames. The major source of inaccuracies came from admission notes that subjectively document the time from symptom onset to initial presentation, particularly when symptoms started more than 12 hours before presentation. Generally, if symptoms started fewer than 12 hours before presentation, the exact time of symptom onset was well documented. Beyond 12 hours, the time of symptom onset was rarely accurately documented, and generalisations were used, such as “for 1 day”, or “yesterday evening”. Thus, if the exact duration was not clearly defined, then this was rounded to the nearest 12 hours, e.g., if the patient had symptoms for 1 day, then the time used was 24 hours.

The stage titled “initial presentation” is significant, as it shows when patients first sought medical attention from symptom onset. For 76.6% of patients, their first presentation was to the ED, while the remainder presented to their GP. Exceptions included cases referred directly to the surgical acute diagnostic unit (ADU) from the GP and thus were reviewed by a general surgical clinician.

Pelvic ultrasound is the preferred diagnostic imaging modality due to its lack of radiation, non-invasiveness and cost-effectiveness compared with computed tomography (CT) or magnetic resonance imaging (MRI).¹⁰ However, ultrasound findings can be variable, and though doppler increases sensitivity, the absence of blood flow is a relatively late sign.¹¹ Therefore, the presence of arterial perfusion should not rule out torsion.¹² Due to the non-specific presentation, CT scans are often the initial imaging performed. CT findings can be utilised to make a diagnosis, with retrospective studies showing twisted pedicles and ovarian masses are strong positive predictive findings for ovarian torsion.¹³ However, 66.6% of patients who initially received a CT scan underwent a further pelvic ultrasound to necessitate diagnosis anyway. Recent radiographic literature identifies a need for better familiarisation of the CT signs of ovarian torsion, such that a diagnosis may be made upon the first scan.¹⁴

Time frames

The average time from symptom onset to theatre was 72.8 hours. Patients who received an oophorectomy had an average time of 83.6 hours compared with 57.9 hours for those who underwent ovarian detorsion +/- cystectomy, a difference of 25.7 hours. This reflects the time-critical nature of the condition and is consistent with literature showing detorsion and revascularisation is possible up to 72 hours. Interestingly, however, there was no significant difference in time from initial healthcare presentation to theatre between the two groups, with an average time of 28.7 hours for detorsion and 28.3 hours for oophorectomy. Similar audits in Australia, New Zealand and the United States of America (USA) showed average times from initial presentation to theatre were much lower than the studied hospital, ranging from 6 to 22 hours,^7–9,15 suggesting more delays are occurring irrespective of the procedure. If these global delays were to be decreased, even to a comparable level to these other tertiary centres, perhaps the overall oophorectomy rate would also decrease.

To observe the effect of symptom acuteness on outcomes, the sample population was divided into three groups based on the time from symptom onset to initial presentation: fewer than 12 hours, 12–24 hours and greater than 24 hours. Oophorectomy rates increased as the time from symptom onset to first presentation increased (Figure 1). Interestingly, however, those with symptoms for greater than 24 hours experienced the shortest time from initial presentation to theatre, at 24.4 hours. For those with symptoms for 12–24 hours and fewer than 12 hours, this time was 34 and 28.7 hours respectively. This raises the question of why management was more efficient for those with prolonged symptoms. A total of 22.2% of the sample population had a healthcare presentation for symptomatic ovarian cysts in the year before their torsion. For those presenting beyond 24 hours, the prevalence of recent symptomatic cysts was 58.3%. This suggests that a recent presentation for ovarian cysts is more likely to delay a future presentation for ovarian torsion, when it should in fact prompt the diagnosis of torsion. This delay explains the greater prevalence of oophorectomy among this group.

The final analysis aimed to identify where in the clinical management the greatest delays occurred (Figure 2). The goal was to find areas where improvements would yield substantial reductions in delay to theatre. Figure 2 shows that the greatest delays were from the initial hospital clinician review to pelvic ultrasound and from gynaecology diagnosis to theatre. A similar USA study showed a reduced time from initial triage to initial clinician assessment (1.9 hours), initial review to ultrasound (5.5 hours) and overall time from initial presentation to operating theatre (OT) (22.4 hours).¹⁵

Discussion

The ideal outcome for women experiencing ovarian torsion is to have a detorsion +/- cystectomy, especially in the case of premenopausal women who have not completed their family, as losing an ovary can significantly affect future fertility. While the results of this audit show that patients who present more acutely (<24 hours of symptoms) are more likely to avoid an oophorectomy, their management in hospital is still delayed, resulting in average oophorectomy rates above 50%. This delay is evident when compared to data from other tertiary centres. The solution to reduce delays and rates of oophorectomy is threefold: better identification of ovarian torsion, more efficient treatment and a greater emphasis on trialling ovarian conservation.

Identification of torsion is challenging due to its non-specific presentation. Future guidelines should increase awareness and serve to educate clinicians about the severity of ovarian torsion. Twenty-eight hours from triage to theatre is far too long of a delay to treat a condition where the loss of an organ is involved and thus should take priority over other acute cases. Furthermore, the risk of torsion in women with pre-existing ovarian cysts must be emphasised to all clinicians, as they are at greater risk of developing torsion while more likely to experience delays. Perhaps new clinical guidelines targeted at gynaecology and emergency clinicians can be of benefit. Clinical guidelines should reflect the variability in clinical features to prevent tunnel vision in diagnosis and to dismantle stringent expectations of how torsion should present. The prime example is peritonism, which is a relatively rare finding, and pain of acute onset, which may not always be the case due to intermittent torsion. Another suggestion could be to upskill junior gynaecology trainees to use bedside ultrasound in the diagnostic process, which could identify large ovarian cysts sooner than a formal pelvic ultrasound or CT. A proven example of successful education was from the American Pediatric Surgical Association (APSA), which implemented one such programme that over a 10-year period more than halved the rate of salpingo-oophorectomy and encouraged ovarian conservation procedures.¹⁶

Alongside improving recognition of torsion, reducing delays to OT and detorsion is critical. Key areas to target are reducing time to pelvic ultrasound and time from gynaecology diagnosis to theatre. Reducing the time to imaging requires multiple considerations. As previously mentioned, most patients receive CT initially, which can add delays to ultrasound. Perhaps one option to nullify this is to develop clear CT diagnostic criteria that can be used to diagnose ovarian torsion, bypassing the need for an additional pelvic ultrasound if there is high clinical suspicion of ovarian torsion. Key features include large ovarian cysts, signs of peri-ovarian inflammation and torted pedicles.¹³ The inability of CT to demonstrate blood flow does put it at a disadvantage to ultrasound. However, the absence of arterial flow is usually only a late sign,¹¹ and there is a lack of consistency among studies on how common this is, with some showing it is absent in a majority of patients while in others only a minority show an absence of ovarian perfusion.^17,18 Further research that looks at patients who were transferred to theatre solely after a CT scan could be insightful in showing the sensitivities of CT findings and to see if this expedites management and improves outcomes. Another suggestion is that the general surgery/ED team could consider ordering a pre-emptive ultrasound alongside referring to gynaecology. This would remove the need for gynaecology to re-see the patient before booking the ultrasound.

Research is now showing returning function of necrotic appearing ovaries, suggesting oophorectomy may not always be necessary. Within the operation notes examined in this study, there were instances where the ovaries disintegrated on contact due to necrosis; however, frequently a decision for oophorectomy was made due to a failure to show signs of revascularisation upon detorsion of necrotic appearing ovaries. However, evidence shows necrotic appearing ovaries that do not appear to immediately revascularise still have the potential to return to function. A study among 12 children who all received detorsion found improved vascularity and follicular development in follow-up sonography, despite 76% of the ovaries showing moderate to severe signs of necrosis on intraoperative visualisation, with some having no return of colour on detorsion.¹⁹ There were also no major complications in these children.¹⁹ There are relatively few studies addressing the preservation of necrotic-appearing ovaries and monitoring for return of function in adult women. Perhaps our ability to judge necrosis on simple visualisation is also flawed, as a 2021 retrospective study found that of 31 ovaries that were visually judged as necrotic, only five of them (16%) had histopathologically confirmed necrosis.²⁰ Twenty (64.5%) had haemorrhage or venous congestion and six (19%) had normal ovarian tissue.²⁰ Evidence supports conserving necrotic-appearing ovaries as opposed to removing them.

Conclusion

For patients presenting with ovarian torsion, the most common symptoms were unilateral abdominal pain, nausea and vomiting, as well as the presence of ovarian cysts on imaging, either CT or ultrasound scan (USS). The majority of the patients presenting to a tertiary New Zealand hospital in this 2-year period underwent an oophorectomy, with delays in management largely arising from initial presentation to diagnostic pelvic ultrasound and diagnosis to theatre. These delays resulted in an average time from initial presentation to theatre that was comparably longer than other tertiary centres in New Zealand, Australia and the USA.^7–9,15

CT imaging is often the first-line scan for acute abdomen presentations. Thus, more study is needed to determine if relying solely on CT findings, alongside a clinical picture suggestive of ovarian torsion, can reduce time to diagnosis and improve outcomes, compared to the commonly used CT followed by ultrasound.

Although the study is limited in its sample size and would likely benefit from further longitudinal analysis that captures more patients, we recommend the development and implementation of guidelines that improve torsion recognition and expedite both imaging and surgery to reduce oophorectomy rates. Furthermore, women presenting with symptomatic ovarian cysts should be counselled on the risk of future torsion and advised to seek urgent acute medical attention for changes in symptoms to reduce the risk of oophorectomy.