Abstract
Objective
To identify reliable clinical, biological, and radiological markers predicting the failure of medical treatment in pediatric patients with acute ethmoidal rhinosinusitis (AERS) and subperiosteal abscess, facilitating informed decisions regarding the need for surgical intervention.
Design
Retrospective multicenter cohort study.
Setting
Pediatric otorhinolaryngology departments at a tertiary-care center and a public hospital in France, spanning from January 2014 to January 2024.
Participants
Children under 18 years diagnosed with Chandler stage III orbital complication of AERS, confirmed by computed tomography (CT).
Interventions
Initial treatment with antibiotics, with surgical intervention as required based on clinical evaluation.
Main Outcome Measures
Factors associated with the need for surgical intervention, including demographic data, clinical examination, C-reactive protein (CRP) levels, leukocyte count, and CT evaluation.
Results
Out of 65 patients, 31 (48%) underwent surgery and 34 (52%) were treated with antibiotics alone. In multivariate analysis adjusted for age, the significant factors associated with surgical intervention included complete eyelid closure (odds ratio (OR) = 7.6; P < .001), ophthalmoplegia (OR = 14.2; P < .001), clinical exophthalmos (OR = 25.0; P < .001), CRP level >60 mg/L (OR = 6.9; P = .006), leukocyte count >15,600 g/L (OR = 7.7; P = .002), radiological exophthalmos (OR = 6.1; P = .001), retro-septal cellulitis (OR = 3.5; P = .02), posterior ethmoid opacification (OR = 6.1; P = .03), and abscess width >4 mm (OR = 8.2; P = .01).
Conclusions and Relevance
Many patients can be managed medically. However, complete eyelid closure should prompt a CT scan. Exophthalmos, retro-septal cellulitis, or an abscess wider than 4 mm is an indication for surgical drainage. Ophthalmoplegia should be interpreted in context. CRP level >60 mg/L, a leukocyte count >15,600 g/L, and posterior ethmoid opacification should lead to close monitoring. These findings can aid in developing a clinico-bio-radiological score to guide treatment decisions, potentially improving patient outcomes by standardizing care protocols.
Keywords
Graphical Abstract
Key Message
In acute ethmoidal rhinosinusitis (AERS) in children, there is no consensus on the markers indicating the need for surgery.
This study identifies key clinical, biological, and radiological markers associated with the failure of medical treatment and the need for surgical intervention.
These findings can lead to better management and outcomes for children with severe AERS and guide the development of a clinico-bio-radiological scoring system.
Introduction
Acute ethmoidal rhinosinusitis (AERS), or acute ethmoiditis, is a predominantly-pediatric infection1 -4 occurring in about 3%-4% of children with acute rhinosinusitis. 5 It primarily affects children under 5 years old, since it is an infection of the ethmoid sinus, present at birth. 6 Ocular complications are frequent and neurological complications can be life-threatening.7 -10
The orbit is prone to contiguous spread of infection because it is separated from the ethmoidal cells only by a thin bony plate, the orbital plate of the ethmoid or “lamina papyracea,” especially in children, because the osseous septa and sinus walls are thin and porous, the suture lines are open, and the vascular foramina are permeable. 11 The early-clinical signs of AERS include periorbital edema, rhinorrhea, headache, and fever, with ophthalmological complications, such as ophthalmoplegia and exophthalmos, potentially developing progressively. 2 Diagnosis is clinical, but computed tomography (CT) scanning is necessary for complicated forms to characterize the extent of infection in the periorbital region and paranasal sinuses, to look for the presence of an abscess or venous thrombosis, and to specify the locations and dimensions of these complications. There is scientific consensus to propose medical treatment as a first approach in the absence of suppurative orbital complications (Chandler stages I and II), to propose surgical treatment in cases of severe orbital extension (Chandler stage IV) or intracranial extension (cerebral venous thrombosis Chandler stage V, and intracranial abscesses). 12
The management of orbital subperiosteal abscesses (Chandler stages III) remains controversial. This stage raises the question of surgical indication, without consensus on the prognostic factors inciting surgical intervention. 13 Recent study data favor a conservative treatment of AERS with subperiosteal abscess,4,14,15 but surgical drainage is generally recommended in locally-severe forms, when there is no improvement under medical treatment, or in case of visual impairment. 2 However, there is no consensus to propose surgical treatment as a first intention and support the choice of management. 11
The goal of this study was to identify clinical, biological, and radiological markers of medical treatment failure in AERS with Chandler stage III orbital complication, prompting the clinician to opt for surgical treatment.
Materials and Methods
This is a multicentric retrospective study including all patients under 18 who were hospitalized for AERS with CT-confirmed Chandler stage III orbital complication, from January 2014 to January 2024, spanning 10 years at the Pediatric Otolaryngology—Head & Neck Surgery of a tertiary-care center and the Otolaryngology—Head & Neck Surgery of a Public Hospital. This study was approved by the ethics committee of our institution (PADS24-295). Exclusion criteria were a different diagnosis, or a Chandler stage different from III.
Children were examined by a pediatrician in the emergency department, an ENT specialist, and an ophthalmologist to assess orbital complications. The clinical data collected included the side, length of stay, fever, consultation delay since the onset of eyelid edema, prior antibiotic intake in the month before the episode, clinical exophthalmos, decreased visual acuity, photomotor reflex, ophthalmoplegia, eyelid closure, and nonsteroidal anti-inflammatory drug intake. The collected paraclinical biological data included leukocyte count and C-reactive protein (CRP) levels. The paraclinical radiographic data included the dimensions of the subperiosteal abscess, the location of ethmoidal opacification, the opacification of other paranasal sinuses, radiological exophthalmos and its grade, periosteal breach, signs of retro-septal cellulitis, and the presence of bone lysis.
All patients underwent a contrast-enhanced CT scan, allowing for the analysis of the ethmoid, other paranasal sinuses, orbits, and the search for intracranial complications (such as abscess and thrombophlebitis). The CT classification of orbital complications of AERS was based on the Chandler score. 12 Radiological grading of exophthalmos was recorded according to the Cabanis classification, which ranks according to 3 severity groups based on the measured exophthalmos: 16 Grade I, if more than two-thirds of the globe was in front of the external bicanthal line; Grade II, if the external bicanthal line was flush with the posterior pole of the globe; and Grade III, if the entire globe was in front of the external bicanthal line.
Antibiotic treatment was performed during hospitalization and based on the intravenous combination of cefotaxime (third-generation cephalosporin) at a dose of 150-200 mg/kg, and rifampicin at 30 mg/kg, for a duration of 5 days. In the case of good clinical progress, an oral switch to amoxicillin-clavulanic acid at 80 mg/day was initiated after 5 days, followed by discharge to home. In the case of surgical drainage, the antibiotic regimen was adjusted based on bacterial culture results when available, with an oral switch initiated after 48 hours of clinical improvement, and at least 5 days after the start of intravenous antibiotic therapy.
Surgery was performed in case of severe initial ophthalmological complication, or the appearance of severe complication under antibiotic treatment. Severe ophthalmological complication was defined as the presence of exophthalmos, visual acuity decrease, or retro-septal cellulitis. Ophthalmoplegia was not always a criterion for surgery when it was linked with intern rectus muscle inflammation, that is, in the case of small abscess without volumetric effect of the subperiosteal abscess on the orbital contents. The second indication for surgery was failure of antibiotic treatment, defined as stagnation or worsening of clinical evaluation after 2 days of medical treatment. In this situation, a new blood sample and imaging (CT scanning or MRI) were performed. Patients who improved under medical treatment were clinically followed up.
Several surgical approaches were performed, each with specific advantages and indications. For a medial subperiosteal abscess, an external orbitotomy via a Lynch approach allows direct access to the abscess at the cost of a scar, whereas an endoscopic endonasal approach with opening of the lamina papyracea offers a less invasive intervention. For superior or superolateral abscesses, an orbitotomy directly over the abscess was the preferred approach. Combined surgery was sometimes necessary in complex cases requiring a multifocal approach or in cases of insufficiency of the endonasal approach. The choice of surgical technique depended on various factors such as the extent of the infection, the presence of complications, and the surgeon’s preferences.
Statistical Analysis
In univariate analysis, were used the chi-squared, Fisher, Mann-Whitney, and Welch tests (when variance homogeneity was not met). In multivariate analysis, a logistic regression accounting for age as a confounding factor was performed. Ninety-five percent confidence intervals were computed for all odds ratios (ORs) and continuous variables. We determined the optimal thresholds for the significant continuous variables using receiver operating characteristic (ROC) curve analysis, with the Youden Index to identify the cutoff values maximizing sensitivity and specificity in predicting the need for surgical intervention. The potential adjustment variables were selected from all collected variables in such a way that there were fewer than 20% of patients with missing data or variables with fewer than 5% missing values.
Results
Population
Using the coding of AERS, 180 patients in the tertiary-care center and 69 patients in the secondary center were identified. After a meticulous review of medical records and reports, 40 cases were excluded due to a differential diagnosis not corresponding to AERS, while 93 cases were excluded due to uncomplicated AERS. Additionally, 51 patients were eliminated from the analysis because the Chandler stage differed from III, ultimately leaving a total of 65 patients for this study focused on patients presenting with Chandler stage III orbital complication of AERS (Figure 1).
Flow diagram of patient inclusion. AERS, acute ethmoidal rhinosinusitis.
Among the 65 retained patients, 13 were operated on as a first-line intervention for a severe initial ophthalmological complication, and 52 initially received exclusive medical treatment. Of these 52 patients, 18 were operated upon following the failure of antibiotic therapy, and 34 were treated without surgery (Figure 2).
Distribution of treatments.
Descriptive Analysis and Comparison in the Function of the Surgical Status
Sixty-nine percent of the patients were male and 31% were female. The average age was 6.83 years (±4.17), and the median age was 7 years (3.00-10.0). The minimum age was 9 months, while the maximum was 16 years. The average CRP was 87.3 mg/L (±71.3), with a median of 68.5 mg/L (44.2-130). The average leukocyte count was 14,482 g/L (±4949), with a median of 15,800 g/L (10,825-18,000).
Table 1 summarizes the comparison between medically (n = 34) and surgically (n = 31) treated patients in univariate analysis. Surgery was significantly associated with ophthalmoplegia, clinical and radiological exophthalmos, the grading of exophthalmos, complete eyelid closure, CRP, leukocyte count, total ethmoidal opacification, and retro-septal cellulitis.
Comparison Between the Patients Treated by Medical Treatment Only and Surgical Treatment, According to Demographics, History-Taking, Clinical Examination, Biology, and CT-Scanning Evaluation.
Radiological exophthalmos grade according to the Cabanis classification. Significant P-values in bold type.
Abbreviations: CIs, confidence intervals; CRP, C-reactive protein; NSAIDs, nonsteroidal anti-inflammatory drugs; SD, standard deviation.
Using the ROC method with the Youden Index, the CRP threshold associated with an increased risk of surgery was 60 mg/L, the leukocyte count threshold was 15,600 g/L, and the abscess width threshold was 4 mm. In multivariate analysis adjusted for age (Table 2), the highest ORs (>8) associated with an increased risk of surgery were clinical exophthalmos, ophthalmoplegia, and an abscess width >4 mm. The statistical association between continuous variables (mean abscess width, CRP level, and leukocyte count) and the need for surgery was not significant after age adjustment in logistic regression. However, when applying the thresholds identified through ROC analysis, the association remained significant.
Multivariate Analysis Adjusted for Age, of the Risk of Undergoing Surgery, Comparing All Operated Patients (n = 31) Versus Non-Operated Patients (n = 34).
Abbreviations: CIs, confidence intervals; CRP, C-reactive protein.
No significant association was found between the choice of antibiotic treatment and the need for surgery (P = .3). The other antibiotic options were intravenous amoxicillin-clavulanate, ceftriaxone-metronidazole, or cefotaxime-rifampicin-metronidazole. The analysis comparing medically-treated patients (n = 34) with the subgroup of patients who required surgery after medical treatment failure (n = 18) yielded results close to those observed in the overall surgically-treated group (Table 3).
Subgroup Analysis of the Patients Not Initially Operated on, Showing Only the Significant Parameters.
Radiological exophthalmos grade according to the Cabanis classification.
Abbreviations: CIs, confidence intervals; CRP, C-reactive protein; SD, standard deviation
Discussion
This study confirmed that the presence of a subperiosteal abscess is not an absolute indication for surgery: more than half of the 65 patients with a subperiosteal abscess complicating AERS were treated exclusively with antibiotics. The clinical markers of severity included the presence of exophthalmos, ophthalmoplegia, or complete eyelid closure. The biological markers associated with the need for surgery were a CRP level exceeding 60 mg/L and a leukocyte count above 15,600 g/L. The radiological markers of severity were exophthalmos, retro-septal cellulitis, opacification of the posterior ethmoid, and an abscess width superior to 4 mm.
Clinical Ophthalmological Complications
This study emphasizes the importance of ophthalmological examination—exophthalmos, ophthalmoplegia, eyelid closure or decreased visual acuity—in surgical decision-making. In particular, the presence of clinical exophthalmos (OR = 25.0; P < .001) is the most significant marker of the risk of poor outcomes under antibiotic treatment. This can be explained by the fact that it reflects a mechanical mass effect on the ocular globe due to a combination of abscess volume, fat inflammation, and soft tissue swelling. Complete eyelid closure also suggests a marked increase in intra-orbital pressure and complicates ophthalmologic monitoring, preventing proper assessment of ocular motility, visual acuity, and signs of optic nerve compromise. An increase in intra-orbital pressure may lead to ischemic optic neuropathy and irreversible vision loss. Cohen et al 17 identified ophthalmological complications as important indicators necessitating immediate surgical intervention. For Quintanilla-Dieck et al, 18 the presence of exophthalmos and intra-orbital gas bubbles was highly predictive of the need for surgical drainage. These results highlight the importance of thorough ophthalmological evaluation in patients presenting with AERS: A specialized ophthalmological examination should be systematic, particularly in cases of doubt.
Biological Inflammatory Markers
High CRP levels seem to be associated with an increased risk of complications, notably orbital abscess, as already reported in studies by Shifman et al 19 and Bülbül et al. 20 In this study, biological inflammatory markers were higher in the patients who underwent surgery. The loss of statistical significance for CRP and leukocyte count as continuous variables after age adjustment may be due to age-related immune response variations, nonlinear risk relationships, and the influence of extreme values. Threshold-based analysis enhances discrimination, explaining why specific CRP and leukocyte cutoffs remain significant predictors of surgery despite age adjustment. New studies are required to determine how to weight biological results in the function of age in the decision-making process.
Abscess Dimensions
Unlike previous studies that focused on the maximum dimension of abscesses, this study highlights the importance of the abscess’ width rather than its length in the decision for surgical intervention: An abscess wider than 4 mm is associated with the need for surgical intervention. Gavriel et al 21 noted that larger abscesses with a length >17 mm and a width >4.5 mm were associated with a higher likelihood of requiring surgical intervention, while McKerlie et al 22 identified an abscess volume threshold of 1.18 mL beyond which surgical intervention is strongly recommended. This observation is logical: The width of the abscess corresponds to its bulging within the orbit, thus its mass effect on the orbital content, unlike the length, which corresponds to the extension along the orbital plate of the ethmoid or the frontal bone. An increased width thus has a significant effect on exophthalmos and the inflammation of the internal rectus muscle. This suggests that the focal-compressive threat posed by a wide abscess is more dangerous than the diffuse pressures exerted by a longer abscess. These findings emphasize the importance of considering both the width and volume of abscesses in radiological evaluation and therapeutic decision-making and encourage radiologists to systematically measure these values. 2
The loss of statistical significance for mean abscess width after age adjustment may be attributed to a correlation between orbital size and abscess width when treated as a continuous variable, a nonlinear risk relationship, and the superior discriminative power of a threshold. Considering anatomical variations may be important in pediatric populations: Ratios between the abscess volume and the orbital volume are likely superior to absolute dimensions when defining the need for surgical drainage.
Other Radiological Signs
CT scan analysis of the orbit is crucial, particularly in young children, for whom ophthalmologic examination may be challenging, and in cases of complete eyelid closure. This study found a significant relationship between the presence of radiological exophthalmos, its grading, and the need for surgical intervention. Radiological exophthalmos may detect subtler cases that are not yet clinically apparent, resulting in a lower OR for the need for surgery compared with clinical exophthalmos. Retro-septal cellulitis is characterized by thickening of the orbital soft tissues, retro-orbital fatty opacification, blurring of the orbital fat contours, and sometimes even a distortion of the extrinsic muscles. Sansa-Perna et al demonstrated that early identification of radiological signs of retro-septal cellulitis is essential for guiding treatment and reducing the risk of severe complications such as orbital abscess and vision loss. 23 Children with posterior or total ethmoidal opacification were more frequently directed toward surgical intervention than those with only anterior opacification. Bülbül et al 20 also observed a relationship between extensive ethmoidal opacification and the necessity of surgical intervention. This link can be explained by the fact that a complete ethmoidal opacification is likely a reflection of more diffuse inflammation and a larger purulent opacification, making medical treatment alone less effective.
Medical Treatment and Surgery
The antibiotic combination of cefotaxime and rifampicin was established with the pediatric infectious disease team to cover germs frequently isolated in the rhinopharyngeal sphere and staphylococcus aureus. This combination is justified by the local bacterial ecology, and the superior clinical efficacy of this combination when compared with that of data reported in the literature. 24 Previous studies, notably that by Brook 25 on AERS, have also highlighted the importance of considering local particularities in the choice of antibiotic therapy. Furthermore, the results of the study by Bory et al 26 on the use of the cefotaxime-rifampicin combination in cervical abscesses in our area reinforce the argument in favor of this therapeutic strategy in the treatment of severe infections, with a superior efficacy of rifampicin over Staphylococcus aureus compared with clindamycin. For patients transferred from other institutions, the treatment was switched to our protocol unless high-dose (150 mg/kg/day) intravenous amoxicillin-clavulanic acid or another third-generation cephalosporin (eg, ceftriaxone) had been initiated. In such cases, the same third-generation cephalosporin was maintained and combined with rifampicin. The 5 day duration of intravenous antibiotic therapy corresponds to the time usually required to achieve near-complete regression of the orbital edema—an indicator of good infection control—before switching to oral therapy. In our practice, we combine systemic treatment with nasal irrigation using saline solution, but we do not use topical corticosteroids or vasoconstrictors. No effect of NSAIDs intake could be demonstrated in this study, despite their suspected role in promoting acute complications of pediatric rhinosinusitis. 27
The analysis did not reveal a clear difference in efficacy between the different surgical techniques. Although some teams suggested that the combined surgical approach might offer superior benefits, particularly in complex cases, it is important to note that no definitive conclusion has been established regarding the superiority of a specific surgical approach. 28 The expertise and experience of the surgeon play a crucial role in the choice of surgical technique.
Methodology
The cohort size constitutes one of the strengths of this study, associated with a bicentric inclusion, from a university hospital and a peripheral hospital, which helps limit recruitment bias. The main limitation of this study is its retrospective nature. First, it is subject to potential biases such as missing data or coding errors. However, these errors are increasingly rare due to the standardization of coding practices and the digitization of medical records. Second, some patients may not have been included if no orbital complications were detected during the clinical examination, as they would not have undergone a CT scan. This absence of imaging might have led to missed diagnoses of well-tolerated Chandler stage III orbital complications. Nonetheless, we believe it is not justified to deviate from current management recommendations by systematically performing CT scans in clinically well-tolerated cases. Third, some rare data could not be retrieved during the review of patient records. CRP and leukocyte data were missing for 13 patients. Abscess dimensions could not be obtained for some patients (11 lengths and 31 widths) because they were not mentioned in the reports, and the images were not available in numerical format, preventing reevaluation of the abscess dimensions. The comparison between the subgroup of 18 patients initially treated medically and later requiring surgery and the patients treated medically only appears to confirm that the predefined surgical criteria were appropriate. However, given the small sample size, these results should be interpreted with caution.
Clinical Implications
Considering the results of this study, several clinical management recommendations can be established for the management of children with Chandler stage III orbital complication of AERS. Many patients can be managed medically; however, certain severity criteria should warrant careful attention. (1) Complete eyelid closure should prompt an immediate CT scan, as clinical assessment is difficult in these cases. (2) Exophthalmos, retro-septal cellulitis, or an abscess wider than 4 mm is an indication for immediate surgical drainage. (3) Ophthalmoplegia should be interpreted in context: If it results from simple inflammation of the medial rectus muscle without the aforementioned radiological criteria, medical treatment may be considered. However, if associated with the radiological signs listed above, surgical drainage is strongly recommended. (4) CRP level above 60 mg/L, a leukocyte count exceeding 15,600 g/L, and posterior ethmoid opacification are not absolute indications for surgery but serve as risk factors for poor outcomes under medical treatment, warranting closer monitoring.
If an initial decision is made for medical treatment, the absence of clinical improvement after 48 hours of IV antibiotics, worsening clinical signs or deterioration of biological markers under medical treatment should prompt consideration of surgical drainage.
We encourage clinicians to perform a precise search for these markers as they can identify patients requiring surgical intervention or a close follow-up. To further advance this approach, the development of a clinico-bio-radiological score based on these markers and accounting for age remains to be carried out and should be the subject of future studies.
Footnotes
Authors Contributions
Eric Moreddu: conceptualization, acquisition, resources, analysis, interpretation of data, manuscript draft, manuscript revision, and final approval. Marie-Eva Rossi: acquisition, manuscript revision, and final approval. Dalia Bellal: acquisition, manuscript revision, and final approval. Fadi Alshawareb: acquisition, resources, manuscript revision, and final approval. Richard Nicollas: acquisition, resources, manuscript revision, and final approval. Jawad Qarbal: acquisition, analysis, interpretation of data, manuscript draft, and final approval.
Data Availability Statement
The datasets generated during the current study are available from the corresponding author on reasonable request.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Considerations
This study was approved by the ethics committee of our institution (PADS24-295).
Consent to Participate
The requirement for informed consent to participate has been waived by the relevant ethics committee, after ensuring that no patient has exercised their right to object to the use of their data with our institution.
