- Email: [email protected]
Novel serum inflammatory markers in patients with adnexal mass who had surgery for ovarian torsion
Novel serum inflammatory markers in patients with adnexal mass who had surgery for ovarian torsion Alexandros Daponte, M.D., F.C.O.G.,a Spyros Pournaras, M.D.,b Christos Hadjichristodoulou, M.D.,c Georgios Lialios, M.D.,a Athanasios Kallitsaris, M.D.,a Antonios N. Maniatis, M.D.,b and Ioannis E. Messinis, M.D., Ph.D.a a Department of Obstetrics and Gynaecology, b Department of Microbiology, and c Department of Hygiene and Epidemiology, University Hospital of Larissa, Larissa, Greece
Objective: To determine the diagnostic value of preoperative levels of interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-␣), interleukin-8 (IL-8), and E-selectin in patients with an adnexal mass who had surgery for ovarian torsion. Design: Prospective study. Setting: University hospital. Patient(s): Fifty consecutive women with adnexal mass who were admitted with a diagnosis of ovarian torsion. Intervention(s): All had a single preoperative measurement of IL-6, IL-8, TNF-␣, and E-selectin by ELISA. Main Outcome Measure(s): To determine if a single preoperative serum sampling of IL-6, IL-8, TNF-␣, and E-selectin by ELISA can diagnose ovarian torsion. Result(s): Thirteen patients had proven torsion of the ovary. Serum levels of IL-6 were significantly higher in the 13 patients with proven ovarian torsion compared with the 37 without ovarian torsion. Patients with IL-6 serum values ⱖ10.2 pg/mL had a 16 times higher risk of having ovarian torsion. No significant difference was found in TNF-␣, IL-8, and E-selectin between the two groups. Conclusion(s): The IL-6 levels might assist in the prompt diagnosis of ovarian torsion and allow a timely surgical intervention. (Fertil Steril威 2006;85:1469 –72. ©2006 by American Society for Reproductive Medicine.) Key Words: Ovarian torsion, IL-6, IL-8, TNF-␣, E-selectin
The clinical diagnosis of adnexal torsion usually relies only on clinical signs such as the presence of abdominal pain and ultrasonographic finding of an adnexal mass (1). A worrying fact is that many of the patients treated by an urgent diagnostic laparoscopy do not have a confirmed ovarian torsion (2). Specific laboratory markers that support the preoperative diagnosis of ovarian torsion are not available in the clinical routine today. A statistically significant association between the diagnosis of ovarian torsion and interleukin-6 (IL-6), but not tumor necrosis factor-alpha (TNF-␣), in peripheral blood was once demonstrated previously in eight ovarian torsion patients (2). Additionally, in the rat testis torsion model, TNF-␣ and interleukin-8 (IL-8) were associated with the torsion response and the neutrophiles’ aggregation, and Eselectin was associated with ischemic reperfusion injury (3, 4). A similar pathophysiology could be expected in ovarian torsion. The present study evaluated the association between a single preoperative serum IL-6, IL-8, TNF-␣, and E-selectin measurement and the operative diagnosis of ovarian torsion. Our aim was to improve the preoperative differentiation of those patients who suffer an ovarian torsion from others presenting with similar clinical signs. Received June 15, 2005; revised and accepted October 25, 2005. Reprint requests: Alexandros Daponte, M.D., F.C.O.G., Department of Obstetrics and Gynaecology, University Hospital of Larissa, 41110 Mezourlo, Larissa, Greece (FAX:⫹30 2410 611097; E-mail: dapontea@ otenet.gr).
0015-0282/06/$32.00 doi:10.1016/j.fertnstert.2005.10.056
METHODS This prospective study enrolled 50 consecutive women who were admitted to the gynecological emergency room at the University Hospital of Larissa with signs and symptoms suggestive of ovarian torsion in a 30-month period between December 2001 and June 2004. The inclusion criteria were acute abdominal pain with or without peritoneal signs, ultrasonographic finding of an ovarian cyst or mass ⬎5 cm, and the severity of the symptoms indicating a surgical intervention. We excluded women with a positive pregnancy test, clinical or ultrasonographic signs of malignancy, and clear signs and symptoms of other pathologies (e.g., acute appendicitis, urinary tract infection, or any other infection). Due to the lack of previous studies on E-selectin, IL-8, and no previous statistical significant data for TNF and ovarian torsion, it was difficult to estimate prospectively the appropriate number of patients required. As far as IL-6 was concerned, sample size calculation was based on the only previous results available, where IL-6 concentrations were reported in patients with ovarian torsion and healthy controls (2). Calculations indicated that group sample sizes of 13 patients and 13 controls were required to achieve 82% power to detect a difference between them at a significance level (alpha) of .05 using a two-sided, two-sample t-test. In our clinical setup, we recruited 37 controls until 13 patients with proven ovarian torsion were included. Therefore, 50 patients were included in our final analysis.
Fertility and Sterility姞 Vol. 85, No. 5, May 2006 Copyright ©2006 American Society for Reproductive Medicine, Published by Elsevier Inc.
1469
The patients enrolled in the study signed an informed consent and Institutional Review Board approval was obtained. All patients had surgery within 24 hours after their admission. The consultant responsible for surgery decided to use endoscopy or laparotomy. Surgeons were blinded to laboratory results before surgery. Blood samples were drawn at admission, centrifuged at 3000 ⫻ g for 10 minutes, and the sera were stored at ⫺70°C until tested. Serum concentrations of IL-6, IL-8, E-selectin, and TNF-␣ were quantitatively estimated in duplicate using an EIA method (BioSource Europe SA, Nivelles, Belgium). The EIA methods applied in this study were performed using the fully automated Triturus analyzer (Grupo Grifols S.A., Parets del Valles, Spain). The IL-6 measurement had a sensitivity of 2 pg/mL and a spectrum of 0 –1,540 pg/mL, after a total incubation time of 3 hours, using 100 L per sample with an interassay coefficient of variation (CV) of 2.2% and an intraassay coefficient of variation of 4.7%. The TNF-␣ measurement had a sensitivity of 3 pg/mL and a spectrum of 15–1,500 pg/mL, after a total incubation time of 4 hours and 30 minutes, using 200 L per sample with an interassay CV of 8% and an intraassay CV of 5.2%. The IL-8 measurement had a sensitivity 0.7 pg/mL and a spectrum of 0 –750 pg/mL, after total incubation time of 2 hours and 30 minutes, using 100 L per sample with an interassay CV of 3% and an intraassay CV of 3.5%. The E-selectin sensitivity was 0.5 pg/mL, with a spectrum of 0.8 –50 pg/mL, after a total incubation time of 2 hours and 15 minutes, using 20 L per sample with an interassay CV of 6% and an intraassay CV of 5.4%. Statistics The statistical analysis was performed using the Mann-Whitney test to assess the difference of IL-6, IL-8, TNF-␣, and E-selectin between the two groups and the Fisher’s exact test to assess if the histological diagnosis influenced the torsion. Statistical significance was defined as P⬍.05. A receiver operating characteristic (ROC) curve was constructed to help select a cutoff point of IL-6 value for the diagnosis of ovarian torsion. A two-by-two table was constructed to determine sensitivity, specificity, as well as positive and negative predictive value and likelihood ratios with confidence interval (CI) of 95% at the selected cutoff point (10.2 pg/mL). RESULTS Fifty consecutive patients were enrolled in this study. The preoperative diagnosis of ovarian torsion was confirmed during an urgent operation (41 diagnostic laparoscopies and 9 laparotomies) in 13 (26%) patients. The consultant on call decided the type of operation, and laparotomy was only performed in some patients over 40 years old. In the 13 torsion cases, surgery revealed torsion of the adnexa with a gross macroscopic appearance of ischemia and engorgement; the mean number of twists was 2 with a range of 1– 4. 1470
Daponte et al.
IL-6 as a marker for ovarian torsion
In 8 of the 13 torsion patients, detorsion of the ovary and cyst excision was the surgical procedure performed during the urgent laparoscopy; in the remaining 5 patients, an oophorectomy was performed. The surgical diagnosis among the remaining 37 patients was an ovarian cyst not in torsion, which was excised and sent for histology, and they were classified accordingly. No significant differences were found between the patients with torsion and the others in regard to mean age (patients with torsion had a mean age of 31.9 years, SD 4.212 vs. patients without torsion who had a mean age of 29.89 years, SD 3.923; P⫽.121) or body mass index (BMI) (patients with torsion had a mean age of 24.45 years, SD 1.854 vs. patients with out torsion who had a mean age of 23.8 years, SD 1.697; P⫽.351), as well as chronic diseases (e.g., hypertension, diabetes, ischemic heart diseases, asthma) or smoking. In all 13 patients with ovarian torsion, serum IL-6 concentrations were elevated with a median value of 30.2 pg/mL vs. a median value of 3.1 pg/mL in the patients without torsion (Table 1). This difference was statistically significant (P⬍.001). No significant difference was found in IL-8, TNF-␣, and E-selectin values between the two groups (Table 1). The IL-6 ROC curve is depicted in Figure 1. The area under the curve (AUC) was 0.897 (SE of 0.048 and a 95% CI of 0.802– 0.991). Using a cutoff point of 10.2 pg/mL, sensitivity was 92.3% (CI 64%–100%), specificity 78.1% (CI 60%–90.7%), positive predictive value 63.2% (CI 38.4%– 83.7%), negative predictive value 96.2% (CI 80.4%–99.9%), positive likelihood ratio 4.21 (CI 2.1 ndash;8.2), and nega-
TABLE 1 Serum values of IL-6, TNF-␣, IL-8, and E-selectin in women with and without ovarian torsion. Torsion
IL-6a
IL-8
No (n ⫽ 37) Median 3.1 12.5 Minimum 2.1 0.81 Maximum 165b 750 Yes (n ⫽ 13) Median 30.2 8.9 Minimum 10.9 0.77 Maximum 38.8 217.3
E-Selectin TNF-␣ 66 25 200 51.5 51 63
1.09 0.01 14.2 2.6 0.01 55.9
Note: IL-6 ⫽ interleukin-6; IL-8 ⫽ interleukin-8; TNF-␣ ⫽ tumor necrosis factor-alpha. a IL-6 was significantly higher in patients with torsion than in patients without torsion (P⬍.001). b Only one control patient exhibited an excessive IL-6 value (165 pg/mL). Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
Vol. 85, No. 5, May 2006
FIGURE 1 The ROC curve to select a clinically relevant cutoff point of IL-6. The AUC is 0.897 (SE of 0.048 and a 95% CI of 0.802– 0.991). At IL-6 values of 10.2 pg/mL, the sensitivity is 1.0 and the specificity is 75%.
Recently, in testicular ischemia reperfusion injury, the first evidence of a cytokine/stress-related, kinase-signaling pathway to E-selectin expression in vivo was reported (3). Color Doppler diagnostic contribution remains controversial, and this is the reason it was not part of the inclusion criteria (6). The mouse model of myocardial ischemia reperfusion indicated that IL-6 was significantly elevated during the reperfusion phase (7). A similar mechanism might exist in ovarian torsion, and it was previously reported, in eight patients, that an elevated IL-6 is associated with ovarian torsion (2). Among situations investigated and treated surgically, IL-6 was also associated with intestinal ischemia (8) or endometrioma (9), and appendicitis (10). Additional markers might help us differentiate more accurately between these clinical situations. Therefore, we attempted to test additional parameters by extrapolating data from similar pathologies in the rat testis torsion model. Determining serum factors required for neutrophil recruitment might not only identify the patients who need emergency surgery but also lead to the design of new medical therapies to protect the germ cells. Such a treatment option would be a blocking antibody for E-selectin, as proposed in rat kidney studies (11).
Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
tive likelihood ratio 10.15 (CI 2.5–300) (Table 2). A 2 test at the level of IL-6 serum values ⱖ10.2 pg/mL indicated that above that level, patients had 16 times the risk of having an ovarian torsion (risk ratio 16.42, 95% CI 2.3– 82.2; P⬍.001). Histological findings were available in all 50 patients and, as expected, benign pathology was largely predominant (98%). In only one case (macroscopically suspicious for malignancy), the frozen section revealed an occult papillary cystadenocarcinoma (2%), and the patient had a staging laparotomy. Of the remaining 49 patients, 31 had functional cysts, 7 had endometriomas, 8 had dermoid cysts, and 3 had mucinous cystadenomas. No statistical relationship was found between the histological diagnosis of the cysts and the incidence of ovarian torsion (P⫽.192). DISCUSSION We designed this prospective study with the intention to enhance the accuracy of diagnosis in cases with a possible ovarian torsion presenting in the emergency room. The inclusion criteria correspond to the daily clinical practice. The diagnosis of ovarian torsion to date is largely unreliable (2). The serum parameters were selected according to literature findings for ovarian torsion in humans (IL-6 and TNF-␣) (2). The E-selectin and IL-8 were not measured previously in humans with possible ovarian torsion and were selected after extrapolating data from the testis torsion model in rats (4, 5). Fertility and Sterility姞
Despite the reported association in the rat testis model, the results of our serum measurements of IL-8, TNF-␣, and E-selectin were not able to predict ovarian torsion. It is possible that single serum measurements are not representative of their local increase after torsion, as was also observed in inflammatory bowel disease patients where intraperitoneal cytokine levels indicate no correlations with systemic inflammatory markers (12). It is not known whether such alterations would possibly be reflected in serial measurements of those parameters.
TABLE 2 Two-by-two table of verified ovarian torsion vs. serum values of IL-6 at the selected cutoff point of 10.2 pg/mL to estimate sensitivity, specificity, and positive/negative predictive value and likelihood ratio (95% CI). IL-6 (pg/mL)a ⱖ10.2 ⬍10.2
Ovarian torsion verified
No ovarian torsion
12 1
7 25
Note: IL-6 ⫽ interleukin-6. a Sensitivity, 92.3% (CI:64%–99.8%); specificity, 78.1% (CI:60%–90.7%); positive predictive value, 63.2% (CI:38.4%– 83.7%); negative predictive value, 96.2% (CI:80.4%–99.9%); positive likelihood ratio, 4.21 (CI:2.1– 8.2); negative likelihood ratio, 10.15 (CI: 2.5–300). Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
1471
Recently, IL-6 has been associated with acute inflammation. Patients with more severe appendicitis had higher plasma levels of IL-6 and IL-10, and lower plasma levels of IL-12 and interferon-gamma than did those with uncomplicated disease (10). That study concluded that the risk for developing appendiceal perforation or gangrene may be determined, in part, by variation in the IL-6 gene expressed in serum and peritoneal concentrations (13). Our results indicated that patients with IL-6 serum values ⱖ10.2 pg/mL have a 16 times higher risk of having an ovarian torsion further strengthening the preliminary results of Cohen et al. (2). It could be noted that although the range of IL-6 in the 37 patients without torsion is 2.1–165 pg/mL, only 1 case exceeded 25 pg/mL. The observation that all patients with torsion had IL-6 levels ⱖ10.2 pg/mL indicates that the major usefulness of this test might be its negative predictive value. Nevertheless, due to the broad CI for the risk ratio, which was possibly due to the relatively low number of cases, the cutoff limit for clinical use should only be set after larger numbers of patients are studied. It appears that the estimation of serum IL-6 in the female patient with adnexal mass and acute abdominal pain with possible ovarian torsion can help in differentiating patients who will benefit from emergency surgical intervention from others who could be treated conservatively. Should the validity of this test be confirmed in larger studies, the patient might benefit from a justified urgent laparoscopy for ovarian torsion, instead of risking damage to the ovary following conservative observation. REFERENCES 1. Cohen SB, Oelsner G, Seidman DS, Admon D, Mashiach S, Goldenberg M. Laparoscopic detorsion allows sparing of the twisted ischemic adnexa. J Am Assoc Gynecol Laparosc 1999;6:139 – 43.
1472
Daponte et al.
IL-6 as a marker for ovarian torsion
2. Cohen SB, Wattiez A, Stockheim D, Seidman DS, Lidor AL, Mashiach S, et al. The accuracy of serum interleukin-6 and tumour necrosis factor as markers for ovarian torsion. Hum Reprod 2001; 16:2195–7. 3. Lysiak JJ, Nguyen QA, Kirby JL Turner TT. Ischemia-reperfusion of the murine testis stimulates the expression of proinflammatory cytokines and activation of c-jun N-terminal kinase in a pathway to Eselectin expression. Biol Reprod 2003;69:202–10. 4. Lysiak JJ, Turner SD, Nguyen QA, Singbartl K, Ley K, Turner TT. Essential role of neutrophils in germ cell-specific apoptosis following ischemia/reperfusion injury of the mouse testis. Biol Reprod 2001;65: 718 –25. 5. Lysiak JJ, Turner SD, Turner TT. Molecular pathway of germ cell apoptosis following ischemia/reperfusion of the rat testis. Biol Reprod 2000;63:1465–72. 6. Pena JE, Ufberg D, Cooney N, Denis AL. Usefulness of Doppler sonography in the diagnosis of ovarian torsion. Fertil Steril 2000;73: 1047–50. 7. Hirano T, Akira S, Tega T, Kishimoto T. Biological and clinical aspects of interleukine-6. Immunol Today 1990;11:443–9. 8. Sutherland F, Cunningham H, Pontikes L, Parsons L, Klassen J. Elevated serum interleukin 6 levels in patients with acute intestinal ischemia. Hepatogastroenterology 2003;50:419 –21. 9. Darai E, Detchev R, Hugol D, Quang NT. Serum and cyst fluid levels of interleukin (IL)-6, IL-8 and tumour necrosis factor-alpha in women with endometriomas and benign and malignant cystic ovarian tumours. Hum Reprod 2003;18:1681–5. 10. Rivera-Chavez FA, Wheeler H, Lindberg G, Munford RS, O’Keefe GE. Regional and systemic cytokine responses to acute inflammation of the vermiform appendix. Ann Surg 2003;237:408 –16. 11. Singbartl K, Ley K. Protection from ischemia-reperfusion induced severe acute renal failure by blocking E-selectin. Crit Care Med 2000; 28:2507–14. 12. Yamamoto T, Umegae S, Kitagawa T, Matsumoto K. Intraperitoneal cytokine productions and their relationship to peritoneal sepsis and systemic inflammatory markers in patients with inflammatory bowel disease. Dis Colon Rectum 2005;48:1451–9. 13. Rivera-Chavez FA, Peters-Hybki DL, Barber RC, Lindberg GM, Jialal I, Munford RS, et al. Innate immunity genes influence the severity of acute appendicitis. Ann Surg 2004;240:269 –77.
Vol. 85, No. 5, May 2006
Objective: To determine the diagnostic value of preoperative levels of interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-␣), interleukin-8 (IL-8), and E-selectin in patients with an adnexal mass who had surgery for ovarian torsion. Design: Prospective study. Setting: University hospital. Patient(s): Fifty consecutive women with adnexal mass who were admitted with a diagnosis of ovarian torsion. Intervention(s): All had a single preoperative measurement of IL-6, IL-8, TNF-␣, and E-selectin by ELISA. Main Outcome Measure(s): To determine if a single preoperative serum sampling of IL-6, IL-8, TNF-␣, and E-selectin by ELISA can diagnose ovarian torsion. Result(s): Thirteen patients had proven torsion of the ovary. Serum levels of IL-6 were significantly higher in the 13 patients with proven ovarian torsion compared with the 37 without ovarian torsion. Patients with IL-6 serum values ⱖ10.2 pg/mL had a 16 times higher risk of having ovarian torsion. No significant difference was found in TNF-␣, IL-8, and E-selectin between the two groups. Conclusion(s): The IL-6 levels might assist in the prompt diagnosis of ovarian torsion and allow a timely surgical intervention. (Fertil Steril威 2006;85:1469 –72. ©2006 by American Society for Reproductive Medicine.) Key Words: Ovarian torsion, IL-6, IL-8, TNF-␣, E-selectin
The clinical diagnosis of adnexal torsion usually relies only on clinical signs such as the presence of abdominal pain and ultrasonographic finding of an adnexal mass (1). A worrying fact is that many of the patients treated by an urgent diagnostic laparoscopy do not have a confirmed ovarian torsion (2). Specific laboratory markers that support the preoperative diagnosis of ovarian torsion are not available in the clinical routine today. A statistically significant association between the diagnosis of ovarian torsion and interleukin-6 (IL-6), but not tumor necrosis factor-alpha (TNF-␣), in peripheral blood was once demonstrated previously in eight ovarian torsion patients (2). Additionally, in the rat testis torsion model, TNF-␣ and interleukin-8 (IL-8) were associated with the torsion response and the neutrophiles’ aggregation, and Eselectin was associated with ischemic reperfusion injury (3, 4). A similar pathophysiology could be expected in ovarian torsion. The present study evaluated the association between a single preoperative serum IL-6, IL-8, TNF-␣, and E-selectin measurement and the operative diagnosis of ovarian torsion. Our aim was to improve the preoperative differentiation of those patients who suffer an ovarian torsion from others presenting with similar clinical signs. Received June 15, 2005; revised and accepted October 25, 2005. Reprint requests: Alexandros Daponte, M.D., F.C.O.G., Department of Obstetrics and Gynaecology, University Hospital of Larissa, 41110 Mezourlo, Larissa, Greece (FAX:⫹30 2410 611097; E-mail: dapontea@ otenet.gr).
0015-0282/06/$32.00 doi:10.1016/j.fertnstert.2005.10.056
METHODS This prospective study enrolled 50 consecutive women who were admitted to the gynecological emergency room at the University Hospital of Larissa with signs and symptoms suggestive of ovarian torsion in a 30-month period between December 2001 and June 2004. The inclusion criteria were acute abdominal pain with or without peritoneal signs, ultrasonographic finding of an ovarian cyst or mass ⬎5 cm, and the severity of the symptoms indicating a surgical intervention. We excluded women with a positive pregnancy test, clinical or ultrasonographic signs of malignancy, and clear signs and symptoms of other pathologies (e.g., acute appendicitis, urinary tract infection, or any other infection). Due to the lack of previous studies on E-selectin, IL-8, and no previous statistical significant data for TNF and ovarian torsion, it was difficult to estimate prospectively the appropriate number of patients required. As far as IL-6 was concerned, sample size calculation was based on the only previous results available, where IL-6 concentrations were reported in patients with ovarian torsion and healthy controls (2). Calculations indicated that group sample sizes of 13 patients and 13 controls were required to achieve 82% power to detect a difference between them at a significance level (alpha) of .05 using a two-sided, two-sample t-test. In our clinical setup, we recruited 37 controls until 13 patients with proven ovarian torsion were included. Therefore, 50 patients were included in our final analysis.
Fertility and Sterility姞 Vol. 85, No. 5, May 2006 Copyright ©2006 American Society for Reproductive Medicine, Published by Elsevier Inc.
1469
The patients enrolled in the study signed an informed consent and Institutional Review Board approval was obtained. All patients had surgery within 24 hours after their admission. The consultant responsible for surgery decided to use endoscopy or laparotomy. Surgeons were blinded to laboratory results before surgery. Blood samples were drawn at admission, centrifuged at 3000 ⫻ g for 10 minutes, and the sera were stored at ⫺70°C until tested. Serum concentrations of IL-6, IL-8, E-selectin, and TNF-␣ were quantitatively estimated in duplicate using an EIA method (BioSource Europe SA, Nivelles, Belgium). The EIA methods applied in this study were performed using the fully automated Triturus analyzer (Grupo Grifols S.A., Parets del Valles, Spain). The IL-6 measurement had a sensitivity of 2 pg/mL and a spectrum of 0 –1,540 pg/mL, after a total incubation time of 3 hours, using 100 L per sample with an interassay coefficient of variation (CV) of 2.2% and an intraassay coefficient of variation of 4.7%. The TNF-␣ measurement had a sensitivity of 3 pg/mL and a spectrum of 15–1,500 pg/mL, after a total incubation time of 4 hours and 30 minutes, using 200 L per sample with an interassay CV of 8% and an intraassay CV of 5.2%. The IL-8 measurement had a sensitivity 0.7 pg/mL and a spectrum of 0 –750 pg/mL, after total incubation time of 2 hours and 30 minutes, using 100 L per sample with an interassay CV of 3% and an intraassay CV of 3.5%. The E-selectin sensitivity was 0.5 pg/mL, with a spectrum of 0.8 –50 pg/mL, after a total incubation time of 2 hours and 15 minutes, using 20 L per sample with an interassay CV of 6% and an intraassay CV of 5.4%. Statistics The statistical analysis was performed using the Mann-Whitney test to assess the difference of IL-6, IL-8, TNF-␣, and E-selectin between the two groups and the Fisher’s exact test to assess if the histological diagnosis influenced the torsion. Statistical significance was defined as P⬍.05. A receiver operating characteristic (ROC) curve was constructed to help select a cutoff point of IL-6 value for the diagnosis of ovarian torsion. A two-by-two table was constructed to determine sensitivity, specificity, as well as positive and negative predictive value and likelihood ratios with confidence interval (CI) of 95% at the selected cutoff point (10.2 pg/mL). RESULTS Fifty consecutive patients were enrolled in this study. The preoperative diagnosis of ovarian torsion was confirmed during an urgent operation (41 diagnostic laparoscopies and 9 laparotomies) in 13 (26%) patients. The consultant on call decided the type of operation, and laparotomy was only performed in some patients over 40 years old. In the 13 torsion cases, surgery revealed torsion of the adnexa with a gross macroscopic appearance of ischemia and engorgement; the mean number of twists was 2 with a range of 1– 4. 1470
Daponte et al.
IL-6 as a marker for ovarian torsion
In 8 of the 13 torsion patients, detorsion of the ovary and cyst excision was the surgical procedure performed during the urgent laparoscopy; in the remaining 5 patients, an oophorectomy was performed. The surgical diagnosis among the remaining 37 patients was an ovarian cyst not in torsion, which was excised and sent for histology, and they were classified accordingly. No significant differences were found between the patients with torsion and the others in regard to mean age (patients with torsion had a mean age of 31.9 years, SD 4.212 vs. patients without torsion who had a mean age of 29.89 years, SD 3.923; P⫽.121) or body mass index (BMI) (patients with torsion had a mean age of 24.45 years, SD 1.854 vs. patients with out torsion who had a mean age of 23.8 years, SD 1.697; P⫽.351), as well as chronic diseases (e.g., hypertension, diabetes, ischemic heart diseases, asthma) or smoking. In all 13 patients with ovarian torsion, serum IL-6 concentrations were elevated with a median value of 30.2 pg/mL vs. a median value of 3.1 pg/mL in the patients without torsion (Table 1). This difference was statistically significant (P⬍.001). No significant difference was found in IL-8, TNF-␣, and E-selectin values between the two groups (Table 1). The IL-6 ROC curve is depicted in Figure 1. The area under the curve (AUC) was 0.897 (SE of 0.048 and a 95% CI of 0.802– 0.991). Using a cutoff point of 10.2 pg/mL, sensitivity was 92.3% (CI 64%–100%), specificity 78.1% (CI 60%–90.7%), positive predictive value 63.2% (CI 38.4%– 83.7%), negative predictive value 96.2% (CI 80.4%–99.9%), positive likelihood ratio 4.21 (CI 2.1 ndash;8.2), and nega-
TABLE 1 Serum values of IL-6, TNF-␣, IL-8, and E-selectin in women with and without ovarian torsion. Torsion
IL-6a
IL-8
No (n ⫽ 37) Median 3.1 12.5 Minimum 2.1 0.81 Maximum 165b 750 Yes (n ⫽ 13) Median 30.2 8.9 Minimum 10.9 0.77 Maximum 38.8 217.3
E-Selectin TNF-␣ 66 25 200 51.5 51 63
1.09 0.01 14.2 2.6 0.01 55.9
Note: IL-6 ⫽ interleukin-6; IL-8 ⫽ interleukin-8; TNF-␣ ⫽ tumor necrosis factor-alpha. a IL-6 was significantly higher in patients with torsion than in patients without torsion (P⬍.001). b Only one control patient exhibited an excessive IL-6 value (165 pg/mL). Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
Vol. 85, No. 5, May 2006
FIGURE 1 The ROC curve to select a clinically relevant cutoff point of IL-6. The AUC is 0.897 (SE of 0.048 and a 95% CI of 0.802– 0.991). At IL-6 values of 10.2 pg/mL, the sensitivity is 1.0 and the specificity is 75%.
Recently, in testicular ischemia reperfusion injury, the first evidence of a cytokine/stress-related, kinase-signaling pathway to E-selectin expression in vivo was reported (3). Color Doppler diagnostic contribution remains controversial, and this is the reason it was not part of the inclusion criteria (6). The mouse model of myocardial ischemia reperfusion indicated that IL-6 was significantly elevated during the reperfusion phase (7). A similar mechanism might exist in ovarian torsion, and it was previously reported, in eight patients, that an elevated IL-6 is associated with ovarian torsion (2). Among situations investigated and treated surgically, IL-6 was also associated with intestinal ischemia (8) or endometrioma (9), and appendicitis (10). Additional markers might help us differentiate more accurately between these clinical situations. Therefore, we attempted to test additional parameters by extrapolating data from similar pathologies in the rat testis torsion model. Determining serum factors required for neutrophil recruitment might not only identify the patients who need emergency surgery but also lead to the design of new medical therapies to protect the germ cells. Such a treatment option would be a blocking antibody for E-selectin, as proposed in rat kidney studies (11).
Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
tive likelihood ratio 10.15 (CI 2.5–300) (Table 2). A 2 test at the level of IL-6 serum values ⱖ10.2 pg/mL indicated that above that level, patients had 16 times the risk of having an ovarian torsion (risk ratio 16.42, 95% CI 2.3– 82.2; P⬍.001). Histological findings were available in all 50 patients and, as expected, benign pathology was largely predominant (98%). In only one case (macroscopically suspicious for malignancy), the frozen section revealed an occult papillary cystadenocarcinoma (2%), and the patient had a staging laparotomy. Of the remaining 49 patients, 31 had functional cysts, 7 had endometriomas, 8 had dermoid cysts, and 3 had mucinous cystadenomas. No statistical relationship was found between the histological diagnosis of the cysts and the incidence of ovarian torsion (P⫽.192). DISCUSSION We designed this prospective study with the intention to enhance the accuracy of diagnosis in cases with a possible ovarian torsion presenting in the emergency room. The inclusion criteria correspond to the daily clinical practice. The diagnosis of ovarian torsion to date is largely unreliable (2). The serum parameters were selected according to literature findings for ovarian torsion in humans (IL-6 and TNF-␣) (2). The E-selectin and IL-8 were not measured previously in humans with possible ovarian torsion and were selected after extrapolating data from the testis torsion model in rats (4, 5). Fertility and Sterility姞
Despite the reported association in the rat testis model, the results of our serum measurements of IL-8, TNF-␣, and E-selectin were not able to predict ovarian torsion. It is possible that single serum measurements are not representative of their local increase after torsion, as was also observed in inflammatory bowel disease patients where intraperitoneal cytokine levels indicate no correlations with systemic inflammatory markers (12). It is not known whether such alterations would possibly be reflected in serial measurements of those parameters.
TABLE 2 Two-by-two table of verified ovarian torsion vs. serum values of IL-6 at the selected cutoff point of 10.2 pg/mL to estimate sensitivity, specificity, and positive/negative predictive value and likelihood ratio (95% CI). IL-6 (pg/mL)a ⱖ10.2 ⬍10.2
Ovarian torsion verified
No ovarian torsion
12 1
7 25
Note: IL-6 ⫽ interleukin-6. a Sensitivity, 92.3% (CI:64%–99.8%); specificity, 78.1% (CI:60%–90.7%); positive predictive value, 63.2% (CI:38.4%– 83.7%); negative predictive value, 96.2% (CI:80.4%–99.9%); positive likelihood ratio, 4.21 (CI:2.1– 8.2); negative likelihood ratio, 10.15 (CI: 2.5–300). Daponte. IL-6 as a marker for ovarian torsion. Fertil Steril 2006.
1471
Recently, IL-6 has been associated with acute inflammation. Patients with more severe appendicitis had higher plasma levels of IL-6 and IL-10, and lower plasma levels of IL-12 and interferon-gamma than did those with uncomplicated disease (10). That study concluded that the risk for developing appendiceal perforation or gangrene may be determined, in part, by variation in the IL-6 gene expressed in serum and peritoneal concentrations (13). Our results indicated that patients with IL-6 serum values ⱖ10.2 pg/mL have a 16 times higher risk of having an ovarian torsion further strengthening the preliminary results of Cohen et al. (2). It could be noted that although the range of IL-6 in the 37 patients without torsion is 2.1–165 pg/mL, only 1 case exceeded 25 pg/mL. The observation that all patients with torsion had IL-6 levels ⱖ10.2 pg/mL indicates that the major usefulness of this test might be its negative predictive value. Nevertheless, due to the broad CI for the risk ratio, which was possibly due to the relatively low number of cases, the cutoff limit for clinical use should only be set after larger numbers of patients are studied. It appears that the estimation of serum IL-6 in the female patient with adnexal mass and acute abdominal pain with possible ovarian torsion can help in differentiating patients who will benefit from emergency surgical intervention from others who could be treated conservatively. Should the validity of this test be confirmed in larger studies, the patient might benefit from a justified urgent laparoscopy for ovarian torsion, instead of risking damage to the ovary following conservative observation. REFERENCES 1. Cohen SB, Oelsner G, Seidman DS, Admon D, Mashiach S, Goldenberg M. Laparoscopic detorsion allows sparing of the twisted ischemic adnexa. J Am Assoc Gynecol Laparosc 1999;6:139 – 43.
1472
Daponte et al.
IL-6 as a marker for ovarian torsion
2. Cohen SB, Wattiez A, Stockheim D, Seidman DS, Lidor AL, Mashiach S, et al. The accuracy of serum interleukin-6 and tumour necrosis factor as markers for ovarian torsion. Hum Reprod 2001; 16:2195–7. 3. Lysiak JJ, Nguyen QA, Kirby JL Turner TT. Ischemia-reperfusion of the murine testis stimulates the expression of proinflammatory cytokines and activation of c-jun N-terminal kinase in a pathway to Eselectin expression. Biol Reprod 2003;69:202–10. 4. Lysiak JJ, Turner SD, Nguyen QA, Singbartl K, Ley K, Turner TT. Essential role of neutrophils in germ cell-specific apoptosis following ischemia/reperfusion injury of the mouse testis. Biol Reprod 2001;65: 718 –25. 5. Lysiak JJ, Turner SD, Turner TT. Molecular pathway of germ cell apoptosis following ischemia/reperfusion of the rat testis. Biol Reprod 2000;63:1465–72. 6. Pena JE, Ufberg D, Cooney N, Denis AL. Usefulness of Doppler sonography in the diagnosis of ovarian torsion. Fertil Steril 2000;73: 1047–50. 7. Hirano T, Akira S, Tega T, Kishimoto T. Biological and clinical aspects of interleukine-6. Immunol Today 1990;11:443–9. 8. Sutherland F, Cunningham H, Pontikes L, Parsons L, Klassen J. Elevated serum interleukin 6 levels in patients with acute intestinal ischemia. Hepatogastroenterology 2003;50:419 –21. 9. Darai E, Detchev R, Hugol D, Quang NT. Serum and cyst fluid levels of interleukin (IL)-6, IL-8 and tumour necrosis factor-alpha in women with endometriomas and benign and malignant cystic ovarian tumours. Hum Reprod 2003;18:1681–5. 10. Rivera-Chavez FA, Wheeler H, Lindberg G, Munford RS, O’Keefe GE. Regional and systemic cytokine responses to acute inflammation of the vermiform appendix. Ann Surg 2003;237:408 –16. 11. Singbartl K, Ley K. Protection from ischemia-reperfusion induced severe acute renal failure by blocking E-selectin. Crit Care Med 2000; 28:2507–14. 12. Yamamoto T, Umegae S, Kitagawa T, Matsumoto K. Intraperitoneal cytokine productions and their relationship to peritoneal sepsis and systemic inflammatory markers in patients with inflammatory bowel disease. Dis Colon Rectum 2005;48:1451–9. 13. Rivera-Chavez FA, Peters-Hybki DL, Barber RC, Lindberg GM, Jialal I, Munford RS, et al. Innate immunity genes influence the severity of acute appendicitis. Ann Surg 2004;240:269 –77.
Vol. 85, No. 5, May 2006