Characteristics of Superior Orbital Subperiosteal Abscesses in Children

Transcription

1 The Laryngoscope VC 2016 The American Laryngological, Rhinological and Otological Society, Inc. Characteristics of Superior Orbital Subperiosteal Abscesses in Children Lourdes Quintanilla-Dieck, MD; Sivakumar Chinnadurai, MD, MPH; Steven L. Goudy, MD; Frank W. Virgin, MD Objectives/Hypothesis: Superior pediatric orbital subperiosteal abscesses (SPAs) are less common than medial ones, and clinical features specific to patients with superior SPAs have not been well defined. Clinical characteristics between patients with superior and medial SPAs are compared to determine whether superior location is a risk factor for surgical intervention. Study Design: Retrospective cohort study. Methods: The target population consisted of patients diagnosed with an SPA and seen by the pediatric otolaryngology service at a tertiary children s hospital between January 2010 and October Imaging characteristics including proptosis, hypoglobus, intraorbital air, and abscess volume as well as treatment interventions were reviewed. Results: Forty patients between 5 and 17 years of age treated for an orbital SPA were identified. Thirteen patients were identified as having superior SPAs; 27 had medial SPAs. The average ages in the two groups were and 9.26 years, respectively. The odds ratio for surgical treatment per each increasing year of age was 1.5 (P 5.004). The proportion of patients requiring surgery was significantly different between the groups (12/13 superior vs. 13/27 medial, P 5.01). The predominant organism group cultured in surgical patients was Streptococcus anginosus (8/24, 29.17%). Superior SPA patients had significantly more proptosis, hypoglobus, and abscess volume on computed tomography scan. Conclusions: Patients with superior SPAs may present with more advanced disease, leading to a higher rate of characteristics such as proptosis, hypoglobus, and intraorbital air, factors that would predispose to surgical drainage. We found that abscess volume was the most predictive of surgery. Key Words: Pediatric, subperiosteal abscess, orbital abscess, superior, surgical drainage. Level of Evidence: 4 Laryngoscope, 127: , 2017 INTRODUCTION Subperiosteal abscess (SPA) and orbital cellulitis are serious infections that frequently involve a team approach to management including pediatric ophthalmology and otolaryngology head and neck surgery teams. They occur most commonly in children and can result in dangerous complications such as meningitis, cavernous sinus thrombosis, subdural empyema, brain abscess, blindness, and death. 1 Multiple published articles have reported comparisons between medically and surgically treated patients, attempting to identify From the Department of Pediatric Otolaryngology Head & Neck Surgery (L.Q.-D.), Oregon Health & Science University, Portland, Oregon; Department of Pediatric Otolaryngology Head & Neck Surgery (S.C., F.W.V.), Vanderbilt University, Nashville, Tennessee; Department of Pediatric Otolaryngology Head & Neck Surgery (S.L.G.), Emory University, Atlanta, Georgia, U.S.A. Editor s Note: This Manuscript was accepted for publication April 18, Presented as a poster at the American Society of Pediatric Otolaryngology Meeting during the Combined Otolaryngology Spring Meeting, Boston, Massachusetts, U.S.A., April 22 26, The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Lourdes Quintanilla-Dieck, MD, 3181 SW Sam Jackson Park Road, PV01, Portland, OR quintani@ ohsu.edu DOI: /lary risk factors for surgical intervention in pediatric orbital cellulitis. Taking these risk factors into consideration for decision making in the course of these patients treatment yields many advantages, including earlier surgical treatment, early culture results with ability to switch from broad-spectrum to directed antibiotic use, and prevention of repeat imaging and exposure to radiation. Several characteristics that increase the likelihood of surgical management have been identified, including age above 9 years, presence of proptosis, extraocular muscle restriction, elevated intraocular pressure, 1 and volume of subperiosteal abscess (> ml). 2,3 Abscess location (e.g., superior, medial, inferior) has been used as a descriptive term. 4 However, this variable has not been thoroughly explored as a predictor of the need for surgical intervention. The purpose of this study is to compare characteristics of superior SPAs compared to those in a medial location, and to evaluate abscess location as a risk factor for surgical intervention. MATERIALS AND METHODS A retrospective chart review was performed on 65 pediatric patients who were treated by the pediatric otolaryngology service at Monroe Carrell Jr. Children s Hospital at Vanderbilt for orbital cellulitis between January 2006 and October Inclusion criteria included age 5 years or greater and availability of an orbital computed tomography (CT) scan showing either 735

2 Fig. 1. Patient with a superior orbital subperiosteal abscess (SPA) (arrow). a superior or medial location of SPA. To be qualified as a superior SPA, the abscess had to have 75% to 100% of its length in contact with the superior bony orbital wall, as determined in the coronal plane of the CT scan (Fig. 1). In contrast, those labeled as medial SPAs had 75% to 100% of their length adjacent to the medial wall, as visualized in either the coronal or axial plane of the CT scan (or at times both planes were used for confirmation). Prior to undertaking the study, the Vanderbilt Human Research Protection Program Institutional Review Board granted approval for our protocol. Fig. 2. Degree of proptosis determined by the difference in globe protrusion between both eyes on computed tomography scan. Fig. 3. Abscess volume calculation using the ellipsoid formula 4/ 3 3 p 3 abc, where a, b, and c correspond to the radius of each dimension. (A) Letter a represents the height, and letter b represents the width of the fluid collection, with the radius of each represented in red. (B) Letter c corresponds to the anteroposterior dimension of the abscess, with the radius represented in red. [Color figure can be viewed in the online issue, which is available at Demographic characteristics included age and gender. Each patient s clinical course was recorded from the electronic medical records, including length of stay and complications that occurred during or within 2 weeks after the hospitalization. When surgery was performed, hospital day at the time of surgery, surgical approach, and culture data were recorded. Orbital CT scans were reviewed, and the following information was gathered: laterality of the abscess, size and location of the SPA, presence and condition of the sinuses (opacified or clear), presence of intraorbital air, hypoglobus, and degree of proptosis. Proptosis was calculated on axial images by measuring the perpendicular distance from the posterior globe margin to the interzygomatic line, at the midglobe level (Fig. 2). 5 The difference between the eyes was recorded for each patient, corresponding to the difference in globe protrusion. Abscess volume was calculated using the ellipsoid formula 4/3 3 p 3 abc, where a, b, and c correspond to the radius of each dimension (Fig. 3). Hypoglobus was assessed on the coronal image that showed the largest globe size for each eye. A straight line along the orbital floor was made along with a parallel line at the edge of the globe. The space between the lines was compared between the two eyes. Hypoglobus was established when this distance was decreased on the side of the SPA when compared to the opposite eye (Fig. 4). 736

3 using the two-tailed Fisher exact test for determining whether there were differences in demographic and clinical characteristics between patients with a medial and those with a superior SPA. For continuous variables such as proptosis, an unpaired t test was performed. Fig. 4. Method of determination of presence or absence of hypoglobus. Two parallel lines were drawn for each eye, one along the orbital floor and a second at the edge of the globe. The space between the lines was compared between the two eyes. (A) A patient in whom the distance was equal between the two eyes, therefore implying no hypoglobus. (B) Decreased distance on the side of the subperiosteal abscess (right) when compared to the opposite eye, consistent with hypoglobus on the right. [Color figure can be viewed in the online issue, which is available at www. laryngoscope.com.] Data were analyzed using SPSS 11.0 software (Statistical Package for the Social Sciences; SPSS Inc., Chicago, IL) and GraphPad QuickCalcs Software (GraphPad Software, Inc., La Jolla, CA) accessed online. Statistical analysis was performed RESULTS A total of 65 pediatric patients were identified that had been admitted to Monroe Carell Jr. Children s Hospital at Vanderbilt for treatment of orbital cellulitis, between January 2006 and October 2014, and evaluated by the pediatric otolaryngology service. To keep a homogenous patient population with regard to sinus development, 6 only patients age 5 years or older were included. Additionally, only patients with medial and superior abscesses were included. All patients had either an orbital or a maxillofacial CT scan with contrast. Either the emergency department physician or the pediatrics team obtained a CT scan upon admission. Some patients had two CT scans ordered during their stay; the criteria for performing this included worsening ophthalmologic examination or lack of significant clinical improvement after multiple days of antibiotics (e.g., persistent fevers, no improvement of proptosis or periorbital cellulitis). If there were two scans done during the same admission, the second scan was included for the study, as long as it was prior to any surgical intervention. Twenty-four patients were less than 5 years old, and one patient had an inferior intraconal location to his orbital abscess. After exclusion criteria were applied, 40 patients remained in the analysis group. Of these, 27 (67.5%) had a medial SPA and 13 (32.5%) had a superior SPA. Demographic characteristics including gender and age are displayed in Table I. There were a greater number of male patients in both groups: 72.4% were males in the medial SPA group and 53.8% in the superior SPA groups. There was no significant difference in the mean age of superior and medial SPA patients (10.92 vs. 9.26, respectively; P <.15). There was also no significant difference in the age of superior SPA patients who underwent surgery and medial SPA patients who underwent surgery (P 5.641). However, within the medial SPA group, surgically treated patients were significantly older than medically treated patients (P 5.021). Taking all patients into account, patients who required surgery TABLE I. Demographic Characteristics in Pediatric Patients With Medial and Superior Subperiosteal Abscesses. Characteristic Medial Abscess Group Superior Abscess Group P Value Total 27 (67.5%) 13 (32.5%) Male 21 (72.4%) 7 (53.8%) Female 8 (27.6%) 6 (46.2%) Age range, yr 5 to 17 5 to 14 Age, yr, mean 6 standard deviation Median age, yr Comorbidities Idiopathic thrombocytopenic purpura; gastroesophageal reflux disease; rheumatoid arthritis; asthma in two patients Craniosynostosis; positivity for flu virus 737

4 TABLE II. Clinical and Imaging Characteristics of the Two Comparison Groups. Characteristic Medial Abscess Group, n 5 27 Superior Abscess Group, n 5 13 P Value Clinical Fever upon presentation 11 (40.7%) 8 (61.5%).310* Leukocytosis >12 on presentation 15 (55.6%) 5 (38.5%).501* Antibiotic therapy prior to admission 14 (51.9%) 8 (61.5%).737* Required surgery 13 (48%) 12 (92.3%).013* Mean age, yr FESS 7 0 External orbitotomy 4 2 FESS 1 external orbitotomy 2 10 Average hospital day of surgery Average length of stay, days Imaging Presence of at least one frontal sinus 26 (96.2%) 13 (100%) 1 Proptosis, mm, mean Hypoglobus 15 (55.6%) 13 (100%).004* Intraorbital air 1 (3.7%) 10 (76.9%) <.001* Abscess volume, ml, mean *Fisher exact test. Unpaired t test. Only 25 patients in the medial abscess group had computed tomography scan images that included all three planes, and therefore the ability to calculate abscess volume. FESS 5 functional endoscopic sinus surgery. were, on average, 3.52 years older than those who did not (95% confidence interval [CI]: ). The odds ratio of needing surgical treatment increases by 1.5 with each year of age above 5 (P 5.004, 95% CI: ). Comparison of clinical characteristics between the two groups was carried out (Table II). Surgical intervention was made in the great majority of cases due to concerning eye findings (decreased vision, diplopia, and/or concerning eye exam). If present on presentation, the patient was taken to surgery on the first hospital day (15/25 surgical patients). Otherwise, a worsening or unchanged eye exam over the next 1 to 2 days was the reason for delayed surgical treatment. In only one case was the surgical indication unclear. A total of 13/27 (48%) patients in the medial SPA group required surgical treatment as compared to 12/13 (92.3%) patients in the superior SPA group (P 5.013). The surgical treatment of 7/27 (26%) medial SPA patients consisted of only functional endoscopic sinus surgery (FESS), whereas none were treated solely with FESS in the superior abscess group. Of the nine medial SPA patients who underwent FESS, all had maxillary antrostomy and either anterior or total ethmoidectomy, and only one had a frontal sinusotomy. Of the 10 superior SPA patients who had FESS, all had maxillary antrostomy and anterior or total ethmoidectomy, and six also had a frontal sinusotomy. One patient in the medial SPA group required a revision FESS for recurrence of the abscess collection six days following initial drainage, and this revision included frontal sinusotomy. There were no other complications or recurrences in the immediate postoperative period or within 2 weeks after surgery. Time to surgery was shorter for the superior abscess group (average day of hospitalization vs ), but did not reach statistical significance (P 5.067). Twenty-four of the patients who underwent surgery had intraoperative cultures performed, and eight of these grew a species within the TABLE III. Results of Intraoperative Cultures in 24 Patients From Our Cohort. Result Streptococcus anginosus group (including Staphylococcus intermedius and Streptococcus constellatus) Medial Abscess Patients Superior Abscess Patients 7 1 No growth 2 3 No predominant organism 1 2 Haemophilus influenza 2 Methicillin-sensitive Staphylococcus aureus 1 1 Streptococcus pyogenes 1 1 Staphylococcus epidermidis 1 Enterobacter agglomerans 1 Staphylococcus haemolyticus 1 Streptococcus pneumoniae 1 Lactobacillus 1 Fungus (Bipolaris sp.) 1 738

5 A receiver operating characteristic curve for multiple variables was generated and is displayed in Figure 5. Evaluations of area under the curve for abscess size and for proptosis were 0.88 and 0.78, with P values of.000 and.005, respectively. For abscess size, a cutoff of 0.67 ml yielded a sensitivity of 71% and specificity of 93% for predicting surgery (odds ratio of 42; P 5.00). For proptosis, a cutoff of 3.85 mm showed 71% sensitivity and 79% specificity (odds ratio of 8.5; P 5.06). When these variables were combined into a multivariate logistic model, it was clear that abscess size was driving the relationship and not proptosis. When controlling for abscess size, proptosis is no longer significant. Using abscess size cutoff as the only predictor for surgery yields 86.8% accuracy for this dataset. Fig. 5. Receiver operating characteristic (ROC) curve showing analysis for abscess size and proptosis. Streptococcus anginosus group (Streptococcus anginosus, Staphylococcus intermedius, or Streptococcus constellatus). The majority of these patients (7/8) were in the medial SPA group. Table III shows other culture results in our group of patients. Analysis of the orbital CT scans revealed that all patients in the superior SPA group had bilateral frontal sinuses present, and all had either partial or complete opacification of both the frontal sinus and the ethmoid sinuses ipsilateral to the abscess. In the medial SPA group, 20/27 patients had bilateral frontal sinus development, 6/27 had only one frontal sinus formed, and 1/27 had neither frontal sinus present. Of the 24 patients who had a well-formed frontal sinus ipsilateral to the SPA, only 15 had opacification of this sinus. Meanwhile, all 27 patients with a medial SPA had opacification of the ipsilateral ethmoid sinuses. Other imaging characteristics, including abscess volume, proptosis, intraorbital air and hypoglobus, were explored in detail and compared between the two groups (Table II). Within the surgical medial SPA group, there was no significant difference in abscess volume between patients treated with only FESS and those who required external orbitotomy (P 5.186). The superior SPA group had a mean proptosis of mm, whereas the medial SPA group s mean proptosis was mm (P 5.031). There was no statistically significant difference in the number of patients with proptosis 4 mmin the superior and medial SPA groups (P 5.12). The rate of requiring surgery was 84% (16/19) in patients with 4 mm or more of proptosis, and this number increased to 93.33% (14/15) in patients with 5 mm or more of proptosis. All of the patients in the superior SPA group had at least a mild level of hypoglobus, whereas only 15/27 (55.56%) medial SPA patients had any hypoglobus (P 5.004). Only 1/27 medial SPA patients had intraorbital air on imaging, whereas 10/13 superior SPA patients had this finding (P <.001). DISCUSSION Several studies have used the location of the SPA to describe the cohort of patients (e.g., the number of patients with each abscess location who eventually require surgery and what approach is used for each). 4,7,8 However the majority of these studies have a small number of patients with superiorly located SPAs, and they do not analyze this characteristic in such a way to use it as an independent risk factor for surgery. One study by Oxford and McClay analyzed their patients with superior versus medial SPAs and found that 5/43 (11.6%) had a superior location, and that these patients were significantly older and had a longer length of stay. 4 This study made treatment recommendations for patients with medial abscesses, but excluded the superior SPA group, likely due to the small number of patients with an abscess in this location. However, all five patients required surgical treatment, supporting our position that superior SPAs have a higher likelihood of requiring surgery given that they present with greater severity of clinical characteristics such as proptosis. We identified a significant difference in hypoglobus between the groups. In addition, superior SPA patients had a significantly higher rate of intraorbital air on radiographic imaging. The exact etiology of this finding is unclear; some have postulated it could imply presence of anaerobic infection or gas-forming bacteria, but we found no evidence of the latter in our group of patients (Table III). Consistent with prior literature, the most common organism group cultured was S anginosus (positive in 8/24 cultures). 4 Interestingly, the majority of these patients (7/8) were in the medial SPA group. It is also possible that the adjacent sinus has decompressed and transmitted air through a bony dehiscence into the orbit. Regardless, the finding of intraorbital air in the context of a relatively large fluid collection would prompt many physicians to advocate for surgical treatment early, likely contributing to the higher rate of surgery in superior abscess patients. Our current analysis includes a larger number of superiorly located SPAs than has been previously reported. This may be due to our inclusion criteria of age 5 or older, because an older group of patients would have a higher rate of fully or partially developed frontal 739

6 sinuses that could predispose to having a superiorly located abscess within the orbit. As would be expected, all patients with a superior SPA had either partial or complete opacification of the ipsilateral frontal sinus, supporting the theory that these infections are secondary to sinusitis, likely frontal, in the absence of other identifiable inciting factors. There was no significant difference between the two groups in terms of age or the length of time between admission and surgical treatment. Proptosis has been found to be predictive of surgery. An article by Rahbar et al. found an estimated probability of surgery of 92% in the presence of 2 mm or more of proptosis. 7 Interestingly, in this article, a multiple logistic regression analysis indicated that proptosis was the only significant predictor of surgical intervention, and abscess volume did not improve the predictive value of the model. Our results differ in that a proptosis cutoff of 3.85 mm had the best predictive value for surgery, with 71% sensitivity and 79% specificity. One possible explanation for the higher cutoff point in the present study is that a higher proportion of superior SPA patients may imply a higher rate of severe disease at presentation. We found that abscess size was the main predictor for surgery, canceling out the significance of proptosis in a multivariate logistic model. Reports in the literature describe a wide range of abscess volume cutoffs from 1.25 to 3.8 ml. 2,3,9 This variance may be related to the different methods of volume measurement by different institutions. In our cohort of patients, surgical abscess volumes were smaller than those previously reported in the literature (lower limit of 0.67 ml). This could be explained by a tendency to take patients to surgery sooner at our institution based mostly on ophthalmologic symptoms and less stringent criteria compared to other places. However, 76.92% of superior SPA patients had an abscess size of 0.67 ml or greater, whereas only 36% of medial SPA patients had this characteristic (P ). This smaller cutoff point may be a reflection of the higher number of superior SPA patients in our cohort compared to patients in other studies, who tend to present with greater severity of ophthalmologic symptoms and therefore a higher rate of needing surgery. There are several limitations to this study. The inherent limitations of any retrospective chart review would apply. This study group was also from a tertiary referral center, where the severity of presentation may be higher. Additionally, a selection bias may exist given that these patients were selected from the cohort of patients with orbital cellulitis who required pediatric otolaryngology consultation for assistance in management. Therefore, the percentage of patients requiring surgery may be higher because those presenting at an earlier stage who had a quick improvement would not have required surgical consultation during their hospital stay. Lastly, the decision to proceed with surgery is made by a team of clinicians and therefore can be considered a subjective outcome, especially because the criteria can vary between different providers. CONCLUSION The clinical course, physical examination, and imaging findings of patients with orbital cellulitis should all be used to determine the best treatment plan. If the patient has clinical worsening in their symptoms or examination while undergoing antibiotic therapy, few would argue against surgical therapy. However, the patient s initial presentation can help in this decision making as well. Important factors include age, the location of the SPA, and abscess volume. We found an odds ratio of requiring surgery of 29 when the abscess volume was 0.67 ml or greater. Patients with a superior SPA will require surgical therapy more frequently than medial SPAs. This may be related to characteristics commonly found in superior SPAs, such as greater proptosis and abscess volume, which in turn lead to more severe ophthalmologic findings. BIBLIOGRAPHY 1. Smith JM, Bratton EM, Dewitt P, Davies BW, Hink EM, Durairaj VD. Predicting the need for surgical intervention in pediatric orbital cellulitis. Am J Ophthalmol 2015;158: Gavriel H, Yeheskeli E, Aviram E, Yehoshua L, Eviatar E. Dimension of subperiosteal orbital abscess as an indication for surgical management in children. Otolaryngol Head Neck Surg 2011;145: Todman MS, Enzer YR. Medical management versus surgical intervention of pediatric orbital cellulitis: the importance of subperiosteal abscess volume as a new criterion. Ophthal Plast Reconstr Surg 2011;27: Oxford LE, McClay J. Medical and surgical management of subperiosteal orbital abscess secondary to acute sinusitis in children. Int J Pediatr Otorhinolaryngol 2006;70: Nugent RA, Belkin RI, Neigel JM, et al. Graves orbitopathy: correlation of CT and clinical findings. Radiology 1990;177: Shah RK, Dhingra JK, Carter BL, Rebeiz EE. Paranasal sinus development: a radiographic study. Laryngoscope 2003;113: Rahbar R, Robson CD, Petersen RA, et al. Management of orbital subperiosteal abscess in children. Arch Otolaryngol Head Neck Surg 2001;31: Tanna N, Preciado DA, Clary MS, Choi SS. Surgical treatment of subperiosteal abscess. Arch Otolaryngol Head Neck Surg 2008;134: Le TD, Liu ES, Adatia FA, Buncic JR, Blaser S. The effect of adding orbital computed tomography findings to the Chandler criteria for classifying pediatric orbital cellulitis in predicting which patients will require surgical intervention. J AAPOS 2014;18: