- Email: [email protected]
Key Clinical Predictors in the Early Diagnosis of Adnexal Torsion in Children
Original Study Key Clinical Predictors in the Early Diagnosis of Adnexal Torsion in Children Heather Appelbaum MD 1,*, Cynthia Abraham MD 1, Jeanne Choi-Rosen MD 2, Meredith Ackerman MS 3 1
Division of Pediatric and Adolescent Gynecology, Department of Obstetrics and Gynecology, Hofstra NSLIJ School of Medicine/Steven and Alexandra Cohen Children’s Medical Center of New York, New Hyde Park, NY Department of Radiology, Hofstra NSLIJ School of Medicine/Steven and Alexandra Cohen Children’s Medical Center of New York, New Hyde Park, NY 3 Feinstein Institute for Medical Research, Northshore LIJ Health Systems, Manhasset, NY 2
a b s t r a c t Objective: Adnexal torsion (AT) accounts for 2.7% of cases of acute abdominal pain in children. When AT is undiagnosed, ovarian blood supply is compromised, eventually leading to tissue necrosis. Because the clinical presentation of AT is nonspecific, preoperative diagnosis is challenging. The purpose of this study was to identify predictors that differentiate AT from other sources of acute abdominal pain. Methods: This study was an IRB-approved retrospective chart review of girls age 4-18 y/o with acute abdominal pain who then underwent surgical evaluation. Data collected included age, menarchal status, symptoms, physical exam findings, laboratory tests, imaging studies, operative procedures and postoperative diagnosis. Factors associated with AT were included in a logistic regression model. A receiver operator characteristic (ROC) curve based on this model was then constructed in order to determine its ability to predict AT. Results: 94 patients presented with acute abdominal pain; 45 were diagnosed with AT and 49 with other causes of abdominal pain. Presence of intermittent pain (P ! .0217), non-radiating pain (P ! .0229) and increased adnexal size (P ! .0032) were significantly associated with AT in the final model. The area under the ROC curve was equal to 0.8601, suggesting excellent discrimination between AT and other causes of acute abdominal pain by using these 3 parameters. Conclusion: Key clinical and imaging findings can aid in the early diagnosis of AT in children. Future prospective studies will focus on development of a clinical predictive model for the diagnosis of AT in the pediatric population. Key Words: Adnexal torsion, Ovarian torsion, Abdominal pain
Introduction
Adnexal torsion (AT) accounts for 2.7% of all cases of children with acute abdominal pain.1 The diagnosis of AT is often missed or delayed due to its nonspecific presentation. Signs and symptoms of AT (nausea, vomiting, fever, abdominal pain, leukocytosis) may be indistinguishable from those of other conditions, leading to an inaccurate or delayed preoperative diagnosis of acute appendicitis, mesenteric adenitis, constipation, or functional ovarian cysts. 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.2 Radiologic imaging is also important when making a diagnosis of AT. However, the most common radiologic finding associated with AT of an enlarged adnexal mass (AM) is nonspecific. One case series by Warner suggested that ovarian cysts greater than 5 cm rarely caused ovarian torsion.3 Conversely, other studies report that larger cysts have a greater risk of torsion. Others suggest the presence of peripheral follicles may be more sensitive than that of an AM.4e8 Considered the most sensitive sign for AT, the whirlpool sign (the visualization of the twisted vascular The authors indicate no conflicts of interest. * Address correspondence to: Heather Appelbaum, MD, Long Island Jewish Medical Center, Department of Obstetrics & Gynecology, 270-05 76th Avenue, New Hyde Park, NY 11040; Phone: (718) 470-7660 E-mail address: [email protected] (H. Appelbaum).
pedicle) is not always appreciated on sonography.9,10 Additional nonspecific imaging signs described in surgically proven torsion include free fluid, abnormal position of the adnexa in relation to the uterus, and tenderness of the adnexa during imaging. Color Doppler sonography can be confusing because there are many series where Color Doppler flow was present in surgically proven cases of AT.11 Furthermore, there is evidence that torsion may occur with normal adnexa in response to congenitally long supportive ligaments, tubal spasm, Mǖllerian anomalies, or abrupt changes in intra-abdominal pressure. Historically, the diagnosis of AT has been presumptive with definitive diagnosis dependent on laparoscopy. When AT remains undiagnosed, blood supply is compromised, which may lead eventually to tissue necrosis and adversely affect future fertility.12 Thus, prompt management is critical in order to reduce morbidity and increase the chances of ovarian salvage. The reported ovarian salvage rate of children with AT who have then undergone untwisting ranges from 27% to 99%.13,14 Numerous case reports and reviews have commented on the incidence of individual symptoms and ultrasound findings in patients with AT.15e18 However, characteristics of the clinical presentation of prepubertal and adolescent girls with AT remain limited. The purpose of this study was to identify key predictors that would aid in differentiating AT from other sources of acute abdominal pain in children and adolescents in order to facilitate prompt management of this condition.
1083-3188/$ - see front matter Ó 2013 North American Society for Pediatric and Adolescent Gynecology. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jpag.2012.12.005
168
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
Methods
This study was an IRB approved retrospective chart review of 94 patients between the ages of 4 and 18 who presented to the Steven and Alexandra Cohen Children’s Medical Center of New York between January 2008 and February 2012 with acute abdominal pain and who subsequently underwent surgical evaluation. Charts were examined for the following: age, menarchal status, symptoms, physical exam findings (presence of abdominal tenderness, peritoneal signs, palpable mass), laboratory tests (white blood cell count), imaging studies (ultrasound or computed tomography), operative procedures, and postoperative diagnosis. AT was defined as laparoscopic visualization of the adnexa twisting on its axis. Patients were evaluated for nausea or vomiting and pain quality, including duration, severity, and pain radiation. Pelvic ultrasounds or CT scans were assessed to determine size of ovary, presence of adnexal masses, size of adnexal mass if present, heterogeneity of the ovarian parenchyma, presence of free fluid, presence of Doppler arterial and venous flow, presence of peripheral follicles, evidence of a twisted pedicle, presence of tenderness during imaging study, position of adnexae relative to uterus, and size of contralateral ovary. All radiologic imaging was interpreted by the same single blinded board-certified radiologist. Descriptive statistics (frequencies for categorical data, mean SD for continuous data) and univariate analyses using the chi-square test or Fisher exact test, as deemed appropriate, for categorical variables and the Mann-Whitney test for continuous data were used to compare patients with and without AT. Those factors that appeared to be associated with AT in the univariate analysis (P ! .25) and were of clinical significance were then included in a logistic regression model. Backwards selection was used to remove variables that did not significantly contribute information to the model, given other factors included in the model. Once the final model was established, a receiver operating characteristic (ROC) curve was constructed and the area under this curve was calculated to determine its ability to predict AT.
Table 1 Descriptive Statistics and Univariate Screen for Predictors of Adnexal Torsion
Age Menarchal status: Pre Post Regular periods Pain radiation Pain scale Duration of pain (days) Fever Nausea/Vomiting Characteristic of pain: Intermittent Constant Abdominal tenderness Peritoneal signs Palpable mass Image modality: Ultrasound CT Free fluid Twisted pedicle Doppler arterial flow Doppler venous flow Peripheral follicles Presence of mass Ovary size (cm) Mass size (cm) Heterogeneity of adnexal mass: Simple Complex Presence of ovarian tenderness at time of imaging Position of adnexa relative to uterus: Abnormal Normal Size of contralateral ovary (cm) White blood cell count O10 Postoperative diagnoses Appendicitis Hemorrhagic ovarian cyst Non-hemorrhagic ovarian cyst Ruptured cyst Hydrosalpinx/ Salpingoophoritis No abnormal findings
Adnexal Torsion (n 5 45)
No Adnexal Torsion (n 5 49)
13 2.5
19 2.7
.04
10 (22%) 35 (78%) 9 (20%) 9 6.6 3.0 16.2 57.6 7 (16%) 28 (62%)
7 (14%) 42 (86%) 7 (14%) 24 7.2 2.3 3.6 8.9 12 (25%) 27 (55%)
.32
29 11 37 11 5
(64%) (36%) (82%) (24%) (11%)
21 26 47 17 2
(43%) (57%) (96%) (35%) (4%)
39 (87%) 6 (13%) 21 (47%) 1 (2%) 8 (18%) 8 (18%) 8 (18%) 39 (87%) 6.5 3.1 6.5 3.6
49 (100%) 0 27 (55%) 0 (0%) 15 (31%) 15 (31%) 5 (10%) 24 (49%) 4.0 1.7 3.6 1.6
8 (18%) 16 (36%) 6 (13%)
7 (20%) 10 (20%) 3 (6%)
10 (22%) 13 (29%) 3.3 1.2 26
3 (6%) 25 (51%) 3.0 0.8 26 24 12 7 2 2
P value
.68* !.01 .49 .07 .28 .48 .01 .24* .38 .24* .01* .41 .44* .03 .03 .23 !.01 !.01 !.01 .61 1.00*
!.01 .35 .40
(49%) (24%) (14%) (4%) (4%)
2 (4%)
Continuous variables expressed as mean standard deviation, results for categorical variables expressed as frequencies * Fischer’s exact test use.
Results
Descriptive statistics and the univariate screen for predictors of AT are outlined in Table 1. Ninety-four patients who presented with acute abdominal pain underwent surgical evaluation. 45 of these patients had AT. Of the 49 patients who did not have AT, 24 patients were diagnosed with appendicitis, 12 with hemorrhagic ovarian cyst, 7 with non-hemorrhagic ovarian cyst, 2 with ruptured ovarian cyst, 1 with hydrosalpinx, 1 with salpingooophoritis, and 2 with no pathology. Presence of abdominal tenderness, palpable mass, intermittent pain, non-radiating pain, and AM were included in the logistic regression model as these were potential predictors associated with AT. Size of the ovary and size of the AM were also considered to be likely predictors of AT and thus were also included in the model. Of these parameters, presence of intermittent pain (P ! .0217), non-radiating pain (P ! .0229), and enlarged adnexa
(P ! .0032) were the only variables that were significantly associated with AT in the final model (Table 2). Number of patients with intermittent abdominal pain in the AT and non-AT groups were 29 and 21, respectively. Number of patients with non-radiating abdominal pain in the AT and non-AT groups were 9 and 24, respectively. Mean AM size in patients with AT and without AT were 6.5 cm and 3.6 cm, respectively. An ROC was created based on the final model Table 2 Logistic Regression Model for Ovarian Torsion Variable Intercept Characteristic of pain (Constant vs Intermittent) Pain radiation (No vs Yes) Mass size
b coefficient Odds Ratio
95% CI
P value
4.3562 1.8669
6.468
.0011 (1.314, 31.834) .0217
1.7075
5.515
(1.267, 24.001) .0229
0.5542
1.740
(1.204, 2.516)
Hosmer-Lemeshow goodness-of-fit test: P ! .8512
.0032
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
Fig. 1. Receiver Operating Curve for Logistic Regression Model including Pain radiation, Characteristic of Pain, and Mass Size as Predictors of Adnexal Torsion.
and the area under the curve equaled 0.8601, a value that suggests excellent discrimination between AT and other causes of acute abdominal pain (Fig. 1). Discussion
In this study, the relationship between potential prognostic factors and diagnoses of acute abdominal pain was evaluated. Of the 22 variables that were investigated, 5 were found to be potential predictors for a logistic regression model, but only 3, when combined, were found to be highly effective in discriminating between AT and other sources of acute abdominal pain, as determined by the area under the ROC curve. These 3 were presence of intermittent pain, presence of non-radiating pain, and increased adnexal size. The strong association between these variables and AT is analogous to findings from other studies. In a retrospective chart review performed by Chang et al19 of forty-nine children with AT, 32% had severe abdominal pain, 14.2% had a history of recurrent pain for more than 2 weeks, and 84% had abdominal ultrasound findings that were suggestive of ovarian pathology. In a retrospective chart review performed by Oltmann et al20 of 328 operative ovarian cases, 97% presented with abdominal pain and 29.6% of patients ultimately were diagnosed with ovarian torsion. Oltmann et al also found that the presence of a pelvic mass 5 cm or larger had 83% sensitivity for torsion and, thus, these investigators concluded that a strategy of earlier use of diagnostic laparoscopy, especially with a pelvic mass larger than 5 cm, would improve rates of ovarian preservation. Standard management of AT for many years involved oophorectomy (and possible laparotomy) secondary to delays in diagnosis that limited preservation of ovarian viability.21 Now, laparoscopic untwisting has become a fundamental tenet of AT management and has been found
169
to be efficacious in children with AT.22e24 In 2010, however, Guthrie et al25 noted that the incidence of ATrelated oophorectomy had not changed between 2000 and 2006. These findings further emphasize the need for a tool that can assist in making an earlier diagnosis of AT in order to improve ovarian viability and ovarian salvage rates. Based on a PubMed search using the keywords, “adnexal torsion,” “ovarian torsion,” and “children,” no study to date has compared the occurrence of specific clinical findings between patients with AT and those with other causes of acute abdominal pain in the pediatric population. Despite the robust association between the triad of intermittent pain, non-radiating pain and increased adnexal size and AT, the absence of these signs does not necessarily reduce the urgency of surgical evaluation in a patient who continues to have a high suspicion of ovarian torsion. But this study does demonstrate key clinical and imaging parameters that can aid in the early diagnosis of AT in children. Limitations of this study are its retrospective nature and small number of subjects. Future prospective studies will focus on development of a clinical predictive model for the assessment of AT in children in order to facilitate preoperative diagnosis and more accurately identify the patients who will benefit from surgical intervention. References 1. Breech LL, Hillard PJ: Adnexal torsion in pediatric and adolescent girls. Curr Opin Obstet Gynecol 2005; 17:483 2. Oelsner G, Shashar D: Adnexal torsion. Clin Obstet Gynecol 2006; 49:459 3. Warner BW, Kuhn JC, Barr LL: Conservative management of large ovarian cysts in children: the value of serial pelvic ultrasonography. Surgery 1992; 112:749 4. Albayram F, Hamper UM: Ovarian and adnexal torsion: spectrum of sonographic findings withpathologic correlation. J Ultrasound Med 2001; 20:1083 5. Stark JE, Siegel MJ: Ovarian torsion in prepubertal and pubertal girls: sonographic findings. AJR Am J Roentgenol 1994; 163:1479 6. Warner MA, Fleischer AC, Edell SL, et al: Uterine adnexal torsion: sonographic findings. Radiology 1985; 154:773 7. Graif M, Shalev J, Strauss S, et al: Torsion of the ovary: sonographic features. AJR Am J Roentgenol 1984; 143:1331 8. Shadinger L, Andreotti R, Kurian R: Preoperative sonographic and clinical characteristics as predictors of ovarian torsion. J Ultrasound Med 2008; 27:7 9. Lee EJ, Kwon HC, Joo HJ, et al: Diagnosis of ovarian torsion with color doppler sonography: depiction of twisted vascular pedicle. J Ultrasound Med 1998; 17:83 10. Mashiach R, Melamed N, Gilad N, et al: Sonographic diagnosis of ovarian torsion: accuracy and predictive factors. J Ultrasound Med 2011; 30:1205 11. Servaes S, Zurakowski D, Laufer MR, et al: Sonographic findings of ovarian torsion in children. Pediatr Radiol 2007; 37:446 12. Wilkinson C, Sanderson A: Adnexal torsion e a multimodality imaging review. Clin Radiol 2012; 67:476 13. Oelsner G, Bider D, Goldenberg M, et al: Long-term follow-up of the twisted ischemic adnexa managed by detorsion. Fertil Steril 1993; 60:976 € n O, Aldemir H, et al: Long-term results of conservative 14. Celik A, Ergu management of adnexal torsion in children. J Pediatr Surg 2005; 40:704 15. Kao JK, Chiu CC, Wang PY, et al: Pediatric ovarian torsion in a medical center in Taiwan: a case analysis. Pediatr Neonatol 2012; 53:55 16. Meyer JS, Harmon CM, Harty P, et al: Ovarian torsion: clinical and imaging presentation in children. J Pediatr Surg 1995; 30:1433 17. Rossi BV, Ference EH, Zurakowski D, et al: The clinical presentation and surgical managment of adnexal torsion in the pediatric and adolescent population. J Pediatr Adolesc Gynecol 2012; 25:109 18. Tsafrir Z, Azem F, Hasson J, et al: Risk factors, symptoms, and treatment of ovarian torsion in children: the twelve-year experience of one center. J Minim Invasive Gynecol 2012; 19:29 19. Chang YJ, Yan DC, Kong MS, et al: Adnexal torsion in children. Pediatr Emerg Care 2008; 24:534 20. Oltmann SC, Fischer A, Barber R, et al: Cannot exclude torsionea 15-year review. J Pediatr Surg 2009; 44:1212 21. Ogburn T, Wurzel J, Espey E, et al: Adnexal torsion: experience at a single university center. J Reprod Med 2005; 50:591
170
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
22. Aziz D, Davis V, Allen L, et al: Ovarian torsion in children: is oophorectomy necessary? J Pediatr Surg 2004; 39:750 23. Rousseau V, Massicot R, Darwish AA, et al: Emergency management and conservative surgery of ovarian torsion in children: a report of 40 cases. J Pediatr Adolesc Gynecol 2008; 21:201
24. Mayer JP, Bettolli M, Kolberg-Schwerdt A, et al: Laparoscopic approach to ovarian mass in children and adolescents: already a standard in therapy. J Laparoendosc Adv Surg Tech A 2009; 19(Suppl 1):S111 25. Guthrie BD, Adler MD, Powell EC: Incidence and trends of pediatric ovarian torsion hospitalizations in the United States, 2000-2006. Pediatrics 2010; 125:532
Division of Pediatric and Adolescent Gynecology, Department of Obstetrics and Gynecology, Hofstra NSLIJ School of Medicine/Steven and Alexandra Cohen Children’s Medical Center of New York, New Hyde Park, NY Department of Radiology, Hofstra NSLIJ School of Medicine/Steven and Alexandra Cohen Children’s Medical Center of New York, New Hyde Park, NY 3 Feinstein Institute for Medical Research, Northshore LIJ Health Systems, Manhasset, NY 2
a b s t r a c t Objective: Adnexal torsion (AT) accounts for 2.7% of cases of acute abdominal pain in children. When AT is undiagnosed, ovarian blood supply is compromised, eventually leading to tissue necrosis. Because the clinical presentation of AT is nonspecific, preoperative diagnosis is challenging. The purpose of this study was to identify predictors that differentiate AT from other sources of acute abdominal pain. Methods: This study was an IRB-approved retrospective chart review of girls age 4-18 y/o with acute abdominal pain who then underwent surgical evaluation. Data collected included age, menarchal status, symptoms, physical exam findings, laboratory tests, imaging studies, operative procedures and postoperative diagnosis. Factors associated with AT were included in a logistic regression model. A receiver operator characteristic (ROC) curve based on this model was then constructed in order to determine its ability to predict AT. Results: 94 patients presented with acute abdominal pain; 45 were diagnosed with AT and 49 with other causes of abdominal pain. Presence of intermittent pain (P ! .0217), non-radiating pain (P ! .0229) and increased adnexal size (P ! .0032) were significantly associated with AT in the final model. The area under the ROC curve was equal to 0.8601, suggesting excellent discrimination between AT and other causes of acute abdominal pain by using these 3 parameters. Conclusion: Key clinical and imaging findings can aid in the early diagnosis of AT in children. Future prospective studies will focus on development of a clinical predictive model for the diagnosis of AT in the pediatric population. Key Words: Adnexal torsion, Ovarian torsion, Abdominal pain
Introduction
Adnexal torsion (AT) accounts for 2.7% of all cases of children with acute abdominal pain.1 The diagnosis of AT is often missed or delayed due to its nonspecific presentation. Signs and symptoms of AT (nausea, vomiting, fever, abdominal pain, leukocytosis) may be indistinguishable from those of other conditions, leading to an inaccurate or delayed preoperative diagnosis of acute appendicitis, mesenteric adenitis, constipation, or functional ovarian cysts. 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.2 Radiologic imaging is also important when making a diagnosis of AT. However, the most common radiologic finding associated with AT of an enlarged adnexal mass (AM) is nonspecific. One case series by Warner suggested that ovarian cysts greater than 5 cm rarely caused ovarian torsion.3 Conversely, other studies report that larger cysts have a greater risk of torsion. Others suggest the presence of peripheral follicles may be more sensitive than that of an AM.4e8 Considered the most sensitive sign for AT, the whirlpool sign (the visualization of the twisted vascular The authors indicate no conflicts of interest. * Address correspondence to: Heather Appelbaum, MD, Long Island Jewish Medical Center, Department of Obstetrics & Gynecology, 270-05 76th Avenue, New Hyde Park, NY 11040; Phone: (718) 470-7660 E-mail address: [email protected] (H. Appelbaum).
pedicle) is not always appreciated on sonography.9,10 Additional nonspecific imaging signs described in surgically proven torsion include free fluid, abnormal position of the adnexa in relation to the uterus, and tenderness of the adnexa during imaging. Color Doppler sonography can be confusing because there are many series where Color Doppler flow was present in surgically proven cases of AT.11 Furthermore, there is evidence that torsion may occur with normal adnexa in response to congenitally long supportive ligaments, tubal spasm, Mǖllerian anomalies, or abrupt changes in intra-abdominal pressure. Historically, the diagnosis of AT has been presumptive with definitive diagnosis dependent on laparoscopy. When AT remains undiagnosed, blood supply is compromised, which may lead eventually to tissue necrosis and adversely affect future fertility.12 Thus, prompt management is critical in order to reduce morbidity and increase the chances of ovarian salvage. The reported ovarian salvage rate of children with AT who have then undergone untwisting ranges from 27% to 99%.13,14 Numerous case reports and reviews have commented on the incidence of individual symptoms and ultrasound findings in patients with AT.15e18 However, characteristics of the clinical presentation of prepubertal and adolescent girls with AT remain limited. The purpose of this study was to identify key predictors that would aid in differentiating AT from other sources of acute abdominal pain in children and adolescents in order to facilitate prompt management of this condition.
1083-3188/$ - see front matter Ó 2013 North American Society for Pediatric and Adolescent Gynecology. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jpag.2012.12.005
168
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
Methods
This study was an IRB approved retrospective chart review of 94 patients between the ages of 4 and 18 who presented to the Steven and Alexandra Cohen Children’s Medical Center of New York between January 2008 and February 2012 with acute abdominal pain and who subsequently underwent surgical evaluation. Charts were examined for the following: age, menarchal status, symptoms, physical exam findings (presence of abdominal tenderness, peritoneal signs, palpable mass), laboratory tests (white blood cell count), imaging studies (ultrasound or computed tomography), operative procedures, and postoperative diagnosis. AT was defined as laparoscopic visualization of the adnexa twisting on its axis. Patients were evaluated for nausea or vomiting and pain quality, including duration, severity, and pain radiation. Pelvic ultrasounds or CT scans were assessed to determine size of ovary, presence of adnexal masses, size of adnexal mass if present, heterogeneity of the ovarian parenchyma, presence of free fluid, presence of Doppler arterial and venous flow, presence of peripheral follicles, evidence of a twisted pedicle, presence of tenderness during imaging study, position of adnexae relative to uterus, and size of contralateral ovary. All radiologic imaging was interpreted by the same single blinded board-certified radiologist. Descriptive statistics (frequencies for categorical data, mean SD for continuous data) and univariate analyses using the chi-square test or Fisher exact test, as deemed appropriate, for categorical variables and the Mann-Whitney test for continuous data were used to compare patients with and without AT. Those factors that appeared to be associated with AT in the univariate analysis (P ! .25) and were of clinical significance were then included in a logistic regression model. Backwards selection was used to remove variables that did not significantly contribute information to the model, given other factors included in the model. Once the final model was established, a receiver operating characteristic (ROC) curve was constructed and the area under this curve was calculated to determine its ability to predict AT.
Table 1 Descriptive Statistics and Univariate Screen for Predictors of Adnexal Torsion
Age Menarchal status: Pre Post Regular periods Pain radiation Pain scale Duration of pain (days) Fever Nausea/Vomiting Characteristic of pain: Intermittent Constant Abdominal tenderness Peritoneal signs Palpable mass Image modality: Ultrasound CT Free fluid Twisted pedicle Doppler arterial flow Doppler venous flow Peripheral follicles Presence of mass Ovary size (cm) Mass size (cm) Heterogeneity of adnexal mass: Simple Complex Presence of ovarian tenderness at time of imaging Position of adnexa relative to uterus: Abnormal Normal Size of contralateral ovary (cm) White blood cell count O10 Postoperative diagnoses Appendicitis Hemorrhagic ovarian cyst Non-hemorrhagic ovarian cyst Ruptured cyst Hydrosalpinx/ Salpingoophoritis No abnormal findings
Adnexal Torsion (n 5 45)
No Adnexal Torsion (n 5 49)
13 2.5
19 2.7
.04
10 (22%) 35 (78%) 9 (20%) 9 6.6 3.0 16.2 57.6 7 (16%) 28 (62%)
7 (14%) 42 (86%) 7 (14%) 24 7.2 2.3 3.6 8.9 12 (25%) 27 (55%)
.32
29 11 37 11 5
(64%) (36%) (82%) (24%) (11%)
21 26 47 17 2
(43%) (57%) (96%) (35%) (4%)
39 (87%) 6 (13%) 21 (47%) 1 (2%) 8 (18%) 8 (18%) 8 (18%) 39 (87%) 6.5 3.1 6.5 3.6
49 (100%) 0 27 (55%) 0 (0%) 15 (31%) 15 (31%) 5 (10%) 24 (49%) 4.0 1.7 3.6 1.6
8 (18%) 16 (36%) 6 (13%)
7 (20%) 10 (20%) 3 (6%)
10 (22%) 13 (29%) 3.3 1.2 26
3 (6%) 25 (51%) 3.0 0.8 26 24 12 7 2 2
P value
.68* !.01 .49 .07 .28 .48 .01 .24* .38 .24* .01* .41 .44* .03 .03 .23 !.01 !.01 !.01 .61 1.00*
!.01 .35 .40
(49%) (24%) (14%) (4%) (4%)
2 (4%)
Continuous variables expressed as mean standard deviation, results for categorical variables expressed as frequencies * Fischer’s exact test use.
Results
Descriptive statistics and the univariate screen for predictors of AT are outlined in Table 1. Ninety-four patients who presented with acute abdominal pain underwent surgical evaluation. 45 of these patients had AT. Of the 49 patients who did not have AT, 24 patients were diagnosed with appendicitis, 12 with hemorrhagic ovarian cyst, 7 with non-hemorrhagic ovarian cyst, 2 with ruptured ovarian cyst, 1 with hydrosalpinx, 1 with salpingooophoritis, and 2 with no pathology. Presence of abdominal tenderness, palpable mass, intermittent pain, non-radiating pain, and AM were included in the logistic regression model as these were potential predictors associated with AT. Size of the ovary and size of the AM were also considered to be likely predictors of AT and thus were also included in the model. Of these parameters, presence of intermittent pain (P ! .0217), non-radiating pain (P ! .0229), and enlarged adnexa
(P ! .0032) were the only variables that were significantly associated with AT in the final model (Table 2). Number of patients with intermittent abdominal pain in the AT and non-AT groups were 29 and 21, respectively. Number of patients with non-radiating abdominal pain in the AT and non-AT groups were 9 and 24, respectively. Mean AM size in patients with AT and without AT were 6.5 cm and 3.6 cm, respectively. An ROC was created based on the final model Table 2 Logistic Regression Model for Ovarian Torsion Variable Intercept Characteristic of pain (Constant vs Intermittent) Pain radiation (No vs Yes) Mass size
b coefficient Odds Ratio
95% CI
P value
4.3562 1.8669
6.468
.0011 (1.314, 31.834) .0217
1.7075
5.515
(1.267, 24.001) .0229
0.5542
1.740
(1.204, 2.516)
Hosmer-Lemeshow goodness-of-fit test: P ! .8512
.0032
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
Fig. 1. Receiver Operating Curve for Logistic Regression Model including Pain radiation, Characteristic of Pain, and Mass Size as Predictors of Adnexal Torsion.
and the area under the curve equaled 0.8601, a value that suggests excellent discrimination between AT and other causes of acute abdominal pain (Fig. 1). Discussion
In this study, the relationship between potential prognostic factors and diagnoses of acute abdominal pain was evaluated. Of the 22 variables that were investigated, 5 were found to be potential predictors for a logistic regression model, but only 3, when combined, were found to be highly effective in discriminating between AT and other sources of acute abdominal pain, as determined by the area under the ROC curve. These 3 were presence of intermittent pain, presence of non-radiating pain, and increased adnexal size. The strong association between these variables and AT is analogous to findings from other studies. In a retrospective chart review performed by Chang et al19 of forty-nine children with AT, 32% had severe abdominal pain, 14.2% had a history of recurrent pain for more than 2 weeks, and 84% had abdominal ultrasound findings that were suggestive of ovarian pathology. In a retrospective chart review performed by Oltmann et al20 of 328 operative ovarian cases, 97% presented with abdominal pain and 29.6% of patients ultimately were diagnosed with ovarian torsion. Oltmann et al also found that the presence of a pelvic mass 5 cm or larger had 83% sensitivity for torsion and, thus, these investigators concluded that a strategy of earlier use of diagnostic laparoscopy, especially with a pelvic mass larger than 5 cm, would improve rates of ovarian preservation. Standard management of AT for many years involved oophorectomy (and possible laparotomy) secondary to delays in diagnosis that limited preservation of ovarian viability.21 Now, laparoscopic untwisting has become a fundamental tenet of AT management and has been found
169
to be efficacious in children with AT.22e24 In 2010, however, Guthrie et al25 noted that the incidence of ATrelated oophorectomy had not changed between 2000 and 2006. These findings further emphasize the need for a tool that can assist in making an earlier diagnosis of AT in order to improve ovarian viability and ovarian salvage rates. Based on a PubMed search using the keywords, “adnexal torsion,” “ovarian torsion,” and “children,” no study to date has compared the occurrence of specific clinical findings between patients with AT and those with other causes of acute abdominal pain in the pediatric population. Despite the robust association between the triad of intermittent pain, non-radiating pain and increased adnexal size and AT, the absence of these signs does not necessarily reduce the urgency of surgical evaluation in a patient who continues to have a high suspicion of ovarian torsion. But this study does demonstrate key clinical and imaging parameters that can aid in the early diagnosis of AT in children. Limitations of this study are its retrospective nature and small number of subjects. Future prospective studies will focus on development of a clinical predictive model for the assessment of AT in children in order to facilitate preoperative diagnosis and more accurately identify the patients who will benefit from surgical intervention. References 1. Breech LL, Hillard PJ: Adnexal torsion in pediatric and adolescent girls. Curr Opin Obstet Gynecol 2005; 17:483 2. Oelsner G, Shashar D: Adnexal torsion. Clin Obstet Gynecol 2006; 49:459 3. Warner BW, Kuhn JC, Barr LL: Conservative management of large ovarian cysts in children: the value of serial pelvic ultrasonography. Surgery 1992; 112:749 4. Albayram F, Hamper UM: Ovarian and adnexal torsion: spectrum of sonographic findings withpathologic correlation. J Ultrasound Med 2001; 20:1083 5. Stark JE, Siegel MJ: Ovarian torsion in prepubertal and pubertal girls: sonographic findings. AJR Am J Roentgenol 1994; 163:1479 6. Warner MA, Fleischer AC, Edell SL, et al: Uterine adnexal torsion: sonographic findings. Radiology 1985; 154:773 7. Graif M, Shalev J, Strauss S, et al: Torsion of the ovary: sonographic features. AJR Am J Roentgenol 1984; 143:1331 8. Shadinger L, Andreotti R, Kurian R: Preoperative sonographic and clinical characteristics as predictors of ovarian torsion. J Ultrasound Med 2008; 27:7 9. Lee EJ, Kwon HC, Joo HJ, et al: Diagnosis of ovarian torsion with color doppler sonography: depiction of twisted vascular pedicle. J Ultrasound Med 1998; 17:83 10. Mashiach R, Melamed N, Gilad N, et al: Sonographic diagnosis of ovarian torsion: accuracy and predictive factors. J Ultrasound Med 2011; 30:1205 11. Servaes S, Zurakowski D, Laufer MR, et al: Sonographic findings of ovarian torsion in children. Pediatr Radiol 2007; 37:446 12. Wilkinson C, Sanderson A: Adnexal torsion e a multimodality imaging review. Clin Radiol 2012; 67:476 13. Oelsner G, Bider D, Goldenberg M, et al: Long-term follow-up of the twisted ischemic adnexa managed by detorsion. Fertil Steril 1993; 60:976 € n O, Aldemir H, et al: Long-term results of conservative 14. Celik A, Ergu management of adnexal torsion in children. J Pediatr Surg 2005; 40:704 15. Kao JK, Chiu CC, Wang PY, et al: Pediatric ovarian torsion in a medical center in Taiwan: a case analysis. Pediatr Neonatol 2012; 53:55 16. Meyer JS, Harmon CM, Harty P, et al: Ovarian torsion: clinical and imaging presentation in children. J Pediatr Surg 1995; 30:1433 17. Rossi BV, Ference EH, Zurakowski D, et al: The clinical presentation and surgical managment of adnexal torsion in the pediatric and adolescent population. J Pediatr Adolesc Gynecol 2012; 25:109 18. Tsafrir Z, Azem F, Hasson J, et al: Risk factors, symptoms, and treatment of ovarian torsion in children: the twelve-year experience of one center. J Minim Invasive Gynecol 2012; 19:29 19. Chang YJ, Yan DC, Kong MS, et al: Adnexal torsion in children. Pediatr Emerg Care 2008; 24:534 20. Oltmann SC, Fischer A, Barber R, et al: Cannot exclude torsionea 15-year review. J Pediatr Surg 2009; 44:1212 21. Ogburn T, Wurzel J, Espey E, et al: Adnexal torsion: experience at a single university center. J Reprod Med 2005; 50:591
170
H. Appelbaum et al. / J Pediatr Adolesc Gynecol 26 (2013) 167e170
22. Aziz D, Davis V, Allen L, et al: Ovarian torsion in children: is oophorectomy necessary? J Pediatr Surg 2004; 39:750 23. Rousseau V, Massicot R, Darwish AA, et al: Emergency management and conservative surgery of ovarian torsion in children: a report of 40 cases. J Pediatr Adolesc Gynecol 2008; 21:201
24. Mayer JP, Bettolli M, Kolberg-Schwerdt A, et al: Laparoscopic approach to ovarian mass in children and adolescents: already a standard in therapy. J Laparoendosc Adv Surg Tech A 2009; 19(Suppl 1):S111 25. Guthrie BD, Adler MD, Powell EC: Incidence and trends of pediatric ovarian torsion hospitalizations in the United States, 2000-2006. Pediatrics 2010; 125:532