This study analyzes different clinical, radiological and laboratory factors as predictors of ovarian torsion in children. Although clinical symptoms and ultrasound data provide the basis for a diagnosis of suspicion, the NLR has the strongest predictive ability for ovarian torsion in infants. Diagnosis complexity in these patients can cause delays in optimal treatment, and therefore seeking potential predictive biomarkers is important for the prognosis of this disease.
The incidence of ovarian torsion is estimated to be between 0.5 and 2 cases per 10,000 patients, representing approximately 2–3% of all visits for abdominal pain in EDs (1, 14). Symptoms can also be very vague and may range from mild to severe pelvic or abdominal pain, nausea, vomiting or fever, and can even mimic other etiologies of abdominal pain, including acute appendicitis, mesenteric adenitis, constipation, functional ovarian cysts, renal colic, pyelonephritis, ectopic pregnancy, and even colitis (15, 16). Furthermore, in the pediatric population, reproductive organs lie high in the abdomen and may be difficult to evaluate on physical examination, which may additionally complicate reaching an appropriate diagnosis (17, 18). Abdominal pain is the most common presentation, which was present in more than 70% of patients in both groups. Vomiting and/or nausea are the most common accompanying complaints (19). In our study, we found that the presence of vomiting was more frequently associated with OT, which may be secondary to the parasympathetic reaction induced by ischemia. In contrast, we observed that hyperthermia was more frequent in patients without OT, more related to abdominal or pelvic infectious causes.
Prolonged interval between the onset of pain and the diagnosis of torsion correlates with a decreased rate of ovarian salvage (14). However, it is difficult to influence the duration between the first symptoms and consultation in the ED. In our study, patients in the OT-group had a shorter time from symptom onset to ED consultation than girls in the non-OT group. The ischemia maintained during the time in the OT may explain that the pain is less bearable in these cases and that is why the time to go to the ED is shorter. Although ovarian torsion is more common in the postmenarcheal girls due to the increased prevalence of ovarian cysts in these patients, it can also be found in premenarcheal children (20). Other common acute adnexal pathologies such as simple ovarian cysts with or without rupture are more frequent during menstruation, and may be confused with an OT. This may justify the findings found in our study, where the majority of girls in the non-OT group were postmenarcheal, unlike those in the OT group, and consequently, ovarian torsion may still remain a potential diagnosis in both pre- and postmenarcheal girls. The association of vomiting, shorter time from symptom onset and premenarcheal age have been identified by other authors as clinical predictors of OT in girls (21). However, these studies do not include ultrasound data or inflammatory laboratory parameters.
Ultrasonography is the preferred imaging modality when ovarian pathology is clinically suspected. Signs such as increased ovarian size, the presence of a complex mass, and free fluid can indicate adnexal torsion (4). However, the ultrasound appearance of ovarian torsion varies depending on factors like the duration and extent of torsion, whether it's complete or incomplete, and the presence or absence of an ovarian mass. Color Doppler sonography has emerged as a potential tool for identifying interruptions in ovarian blood flow in recent years (22, 23). Yet, it's essential to recognize that the presence of vascular flow on Doppler studies does not conclusively exclude torsion, nor does the flow absence confirm OT. Shadinger et al. found arterial flow in 54% and venous flow in 33% of patients with pathologically proven OT (4). In our study we found significantly higher ovarian volume and ovarian volume ratio in patients with OT, with no differences in the rates of absence of Doppler flow or pelvic free fluid. We also observed a predominant occurrence of torsion on the right side in both groups, similar to previous studies (24). This might be attributed to the presence of the sigmoid colon in the left iliac fossa, which reduces the mobility of the tubal structure and consequently lowers the risk of left adnexal torsion.
Some laboratory data have been tested to predict OT, although several authors found them not helpful in the diagnostic process (25, 26). CRP levels rise in response to inflammation and tissue necrosis several hours after torsion, making it of little use for early diagnosis (27). In addition, alternating CRP concentrations during the menstrual cycle have been reported, which makes interpretation difficult (28). In OT, tissue ischemia initiates systemic inflammation that can be quantified in peripheral blood. In adults, markers such as interleukin-6, interleukin-8, tumor necrosis factor-α, and E-selectin have been proposed, but low availability and high cost make their use in clinical practice complicated (29). In contrast, white blood cell data reflect systemic inflammation, being universally available, quickly analyzed, and cost effective. In this context, the usefulness of NLR as a diagnostic marker of OT has been described, mainly in adult women (10, 30). However, there is scarce experience in pediatric patients. Nissen et al. analyzed laboratory data from 18 girls with OT and 58 controls with ovarian pathology other than torsion, and observed that NLR and PLR allowed differentiation between the two cohorts (11). Nonetheless, they did not include ultrasound data in the analysis, so our present study represents a novelty in this aspect. Furthermore, our study is the first to analyze different inflammatory indices (NLR, PLR, SIRI, SII), which are easily calculated from blood count data, and also compares them with clinical and ultrasound data, similar to what occurs in clinical practice. The results obtained demonstrate that the NLR is the most sensitive and specific predictor for the diagnosis of OT in girls. On the one hand, it is a more objective parameter than clinical data such as time from symptom onset, which is sometimes not very accurate since the presentation sometimes begins with progressive pain. In addition, it avoids the explorer-dependent variability of radiological tests such as ultrasound.
Inflammatory response secondary to ovarian ischemia results in neutrophilia due to chemotaxis and increased release of these cells from the bone marrow to the peripheral blood, which is combined with lymphopenia induced by elevated levels of endogenous cortisol due to ischemia (31). All this leads to an elevation of the NLR by two combined pathways. This may explain the higher AUC with respect to other inflammatory indices such as PLR, SIRI or SII that involves monocytes or platelets, as NLR translates the combined cellular response of neutrophilia and lymphopenia. These observations may extend beyond female reproductive organs, as the role of NLR as a predictor of testicular torsion in adolescents has also been recently described (6). Other authors have described the usefulness of NLR for the differentiation of ruptured ovarian cysts and adnexal torsion, although no analysis of additional inflammatory indices has been performed (32). Tayyar et al. reported a marked reduction of platelets in patients with OT with comparably unaltered lymphocyte counts, which conditioned a low predictive capacity of the PLR, consistent with the results obtained in our study (33). In summary, the main advantage of NLR is that it incorporates the informative aspects of two variables representing contrasting immune pathways through the leukocyte subtype ratio, which provides a more accurate depiction of the overall impact of alterations in ovarian torsion. NLR demonstrated superior discriminatory power compared to leukocytes, neutrophils, PLR, SIRI and SII, as evidenced by higher AUC values in ROC curves analysis.
The main strength of this study is the inclusion of clinical, ultrasound and laboratory data, which allows ovarian torsion to be analyzed from a holistic perspective, which constitutes a novelty in this aspect. This translates into high applicability, as it includes the same parameters that are used in routine clinical practice in this type of patient. However, our study has limitations that should be taken into account. The retrospective design is the main limitation, since it only allows us to analyze the data previously collected in the medical record. In addition, the absence of similar studies in pediatric patients makes it difficult to compare the results obtained. The sample size is also limited, despite the multicenter participation, due to the relatively low incidence of this pathology. Therefore, caution should be adopted when extrapolating or generalizing these results. Prospective studies are required to validate these findings.