Abstract
Objectives:
To evaluate the trend and factors associated with surgical management of orbital cellulitis.
Methods:
Retrospective study using the State Inpatient Databases (SID) from 2008–2015. Patients with an ICD-9-CM diagnosis code for orbital cellulitis were identified in the SID for the following states: Arkansas, Florida, Iowa, Maryland, Nebraska, New York, and Wisconsin. Surgery was defined as an ICD-9-CM procedure code for orbitotomy and/or functional endoscopic sinus surgery. The trend of surgery over time was evaluated using the Cochran-Armitage Test. Multivariable logistic regression models were used to identify patient-and hospital-level factors associated with surgery.
Results:
From 2008 to 2013, the number of hospitalizations for orbital cellulitis ranged from 1349 to 1574, but declined to 865 in 2014. From 2008 to 2015, the number of surgeries ranged from 103 to 154. For children (n=3,041), age, ophthalmologic comorbidity, and conjunctival edema were significantly associated with surgery, whereas for adults (n=7,961), male gender, private insurance, optic neuritis, and CN III/VI/VI palsy were associated with surgery.
Conclusion:
While the number of inpatient hospitalizations for orbital cellulitis has markedly declined, the number of surgeries for orbital cellulitis has remained fairly stable, leading to an observed higher proportion of hospitalized patients undergoing surgery. Future directions include extending the time frame to the present day to assess current rates of hospitalization and surgery. Knowledge of surgical trends and patient-and hospital-level characteristics associated with surgery may help improve management guidelines for and understanding of this vision-threatening disease.
Keywords: Orbital cellulitis, surgical trends, state inpatient database
INTRODUCTION
Orbital cellulitis is a potentially vision and life-threatening disease that occurs when the post-septal orbital contents become infected. Early diagnosis and treatment with intravenous antibiotics with or without surgical drainage are warranted to prevent serious complications, such as blindness, meningitis, and brain abscess.1 Repeat imaging is indicated if there is worsening vision or no clinical improvement within 24–48 hours of initiation of antibiotics. To date, however, there is no uniform consensus in the otolaryngology or ophthalmology literature regarding the timing and indications of surgical intervention for orbital cellulitis.
While there are several clinical studies that have outlined criteria for surgical management of orbital cellulitis 2–7, the majority are single-institution with small sample sizes, thus limiting the generalizability of study findings. Administrative databases collected and maintained by both public and private organizations allow patient characteristics and treatment outcomes to be studied at a population level. The State Inpatient Databases (SID) encompass all inpatient discharges from community hospitals for participating states, providing a unique and comprehensive dataset of inpatient care for a specific state that can be tracked longitudinally.8 Furthermore, the SID allows for linkage of individuals in the database to the treating hospital to provide information regarding hospital characteristics. To our knowledge, there is no study that has examined the utilization trend of surgery in the treatment of orbital cellulitis. Using the SID, the objectives of this study are to evaluate the utilization trend of surgery for orbital cellulitis as well as patient- and hospital-level factors associated with surgical management.
MATERIALS AND METHODS
Study Design
A retrospective study was performed using the State Inpatient Databases (SID) for Arkansas (AR), Florida (FL), Iowa (IA), Maryland (MD), Nebraska (NE), New York (NY), and Wisconsin (WI). These seven states were chosen to provide a representative and comprehensive population of pediatric and adult patients. The overall time frame for the study was between January 1, 2008 to September 30, 2015. Because longitudinal personal identifiers were not available for all states during the overall study period, the following time frames were used for each state: AR (2009–2015), FL (2008–2015), IA (2014–2015), MD (2014–2015), NE (2008–2015), NY (2008–2015), WI (2014–2015). The study end-date was chosen as September 30, 2015 as October 1, 2015 marked the implementation of International Classification of Diseases, Tenth Revision, Clinical Manifestation (ICD-10-CM). The study was deemed exempt from review by the Washington University School of Medicine Institutional Review Board (St. Louis, MO).
Data Source
The SID are part of a family of administrative databases developed by the Healthcare Cost and Utilization Project (HCUP) and maintained by the Agency for Healthcare Research and Quality (AHRQ). The SID contains comprehensive information on all inpatient discharge records from community hospitals for participating states, regardless of payer. 8 Hospital-specific characteristics were obtained from the American Hospital Association Annual Survey of Hospitals (Health Forum, LLC, Chicago, IL) and linked to SID data by hospital identifier.
Study Population
Patients aged one year and above identified in the SID as having an ICD-9-CM diagnosis code of orbital cellulitis (376.01) were included. Children were defined as those <18 years of age at the time of index hospitalization. Patients with out-of-state residence were excluded due to potential for loss to follow-up, as were those with missing gender, race, and personal identifiers.
Outcome Measures
The primary outcome was surgery, defined using the ICD-9-CM procedure codes for orbitotomy and/or functional endoscopic sinus surgery (Table 1). Principal exposures included patient-level and hospital-level factors associated with surgery.
Table 1.
ICD-9-CM Diagnosis and Procedure Codes for Included Data
| ICD-9 CM Codes | Description |
|---|---|
| Inclusion Criteria | |
| 376.01 | Orbital cellulitis |
| Surgical Procedures | |
| 16.01, 16.02, 16.09 | Orbitotomy |
| 22.41, 22.42 | Frontal sinusotomy/sinusectomy |
| 22.51, 22.63 | Ethmoidotomy/ethmoidectomy |
| 22.52, 22.64 | Spheoidotomy/sphenoidectomy |
| 22.50, 22.60 | Other nasal sinusotomy/sinusectomy |
| Concomitant Diagnoses | |
| 461.X, 473.X | Acute sinusitis |
| 522.5, 522.7 | Chronic sinusitis |
| 682.0 | Periapical abscess |
| 373.13 | Facial cellulitis |
| 372.73 | Eyelid abscess |
| 376.30, 376.32, 376.34, 376.35, | Conjunctival edema |
| 376.36 | Exophthalmos |
| 368.2 | Diplopia |
| 369.X | Blindness |
| 377.30, 377.39, 377.31 | Optic neuritis |
| 378.51, 378.52, 378.53, 378.54 | CN III, IV, VI Palsy |
| 376.03, 730.08, 730.28 | Osteomyelitis |
| 377.00, 377.01 | Papilledema |
| 378.56 | Ophthalmoplegia |
| 368.4X | Visual field defect |
| 368.11 | Sudden vision loss |
| 360.00, 360.01, 360.02, 360.04 | Endophthalmitis |
Surgery utilization was defined as the overall change in the rate of surgery during the defined study period. To evaluate the trend of surgery utilization for orbital cellulitis, sub-analysis of data from the FL, NE, and NY SID was performed. These 3 states were chosen from the original 7 states because they had complete data for all years from 2008 to 2015, thus providing comprehensive longitudinal data. In addition, because we used only the first three quarters of 2015 due to the change from ICD-9-CM to ICD-10-CM, the rate of surgery in 2015 was calculated based on just the first three quarters of 2015. For evaluation of patient-level and hospital-level factors associated with surgery, data from all 7 states were used.
Patient-Level Exposure Variables
Patient demographics, primary payer status, median household income (a proxy for socioeconomic status), and associated rhinologic and ophthalmologic conditions were evaluated. Because of small numbers, the categories for race/ethnicity were collapsed into White, Black, Hispanic, or Other. Pediatric ophthalmologic comorbidities were defined using the Pediatric Medical Complexity Algorithm.9 There is no established adult comorbidity algorithm for ophthalmologic comorbidities.
Hospital-Level Variables
Hospital-level variables included bed-size, urban versus rural location, and teaching status as well as admission through the emergency department and intensive care unit (ICU) admission. ICU placement was determined by ICU-specific charges or revenue codes.
Statistical Analysis
The Cochran-Armitage Test for Trend was used to evaluate the linear trend of surgery over time. Descriptive statistics were used to explore the distribution of patient- and hospital-level characteristics in patients who received surgical treatment and those who did not. Independent samples t-test or its nonparametric equivalent, the Wilcoxon rank-sum test, was used to compare differences in characteristics between surgical and nonsurgical patients. Univariable logistic regression was used to explore patient- and hospital-level characteristics associated with surgical treatment. The adjusted role of variables significantly associated with surgery or identified as clinically important was evaluated in a hierarchical logistic regression model using the SAS GLIMMIX procedure to account for nesting of patients within hospitals. Acknowledging the large sample size used for analysis, we pre-set our alpha level at 0.01. Odds ratios with 99% confidence intervals (CI) were reported for all analyses. All statistical analyses were performed using SAS Enterprise Guide 7.1 (SAS Institute, Cary, NC).
RESULTS
Trend of Utilization of Surgery
From January 1, 2008 to September 30, 2015, a total of 10,148 children and adults were admitted for orbital cellulitis in NY, NE, and FL (Figure 1). Between 2008 and 2011, the overall number of individuals admitted for orbital cellulitis (range 1472–1574), the number who received surgery (range 110–154), and the rate of surgery (range 8–10%) remained relatively steady. Beginning in 2012, there was a reduction in the number of individuals admitted to the hospital with a diagnosis of orbital cellulitis. However, the reduction in hospital admissions was not mirrored in the number of surgeries performed; rather, the absolute number of surgeries remained fairly consistent. Thus, the rate of surgery among hospitalized patients appeared to increase starting in 2012 (Cochran-Armitage Test for Trend from 2012 to 2015, p<0.001).
Figure 1.
Utilization Trend of Surgery from 2008–2015 for NE, NY, and FL SID displaying number of individuals diagnosed with orbital cellulitis (left y-axis) and the rate of surgery (right y-axis). A) Between 2008 and 2011, both the number of orbital infections and rate of surgery remain steady for the overall cohort, but, starting in 2012 there is an increase in the rate of surgery (Cochran-Armitage Test for Trend 2012–2015, p <0.001); B)/C) Likewise, for the pediatric and adult cohort, a similar and significant trend is observed in the rate of surgery (Cochran-Armitage Test for Trend; p=0.0026 and p<0.001, respectively). *significance on Cochran-Armitage Test for Trend.
+ Total number of patients and surgery rates in 2015 are calculated based on the first three-quarters of 2015.
When stratified by age, differences in the trend of hospitalizations and surgeries for orbital cellulitis between the pediatric and adult population are observed. Between 2008 to 2015, the absolute number of pediatric hospitalizations declined over four distinct time intervals, while a similar but more gradual reduction in absolute number of surgeries occurred. When grouped into four time periods (2008 to 2009, 2010 to 2011, 2012 to 2013, and 2014 to 2015), the average number of hospitalizations markedly declined from 500 and 408 in the first and second periods, respectively, to 334 and 192 in the third and fourth periods, respectively. Thus, when comparing hospitalizations between these four time intervals, the following is observed: 22% decline in hospitalizations between the first and second time periods; 18% decline between the second and third time periods; and marked 42% decline in total hospitalizations between the third and fourth time periods. Compared to the first time period (2008–2009), there was a 62% decrease in mean hospitalizations by the fourth time period (2014–2015). From 2008 to 2009, the average number of surgeries was 54 and progressively declined to 38 by the fourth period, representing an overall 30% decrease in mean surgeries between the first and last time periods. As a result of this marked reduction in the number of children admitted for orbital cellulitis but modest reduction in the number of surgeries performed, there appears to be a consequent increase in the rate of surgery among hospitalized children beginning in 2012 (Cochran-Armitage Test for Trend from 2012 to 2015, p= 0.0026).
In contrast to that of the pediatric population, the decline in absolute number of hospitalizations and surgeries for adults was observed over 2 as opposed to 4 distinct time periods. From 2008 to 2013, the average number of hospitalizations was 1,032 and declined by 38% to 642 between 2014 to 2015. During these two time intervals, the average number of surgeries remained relatively constant (n=78 and n=76, respectively). As a result, when comparing the average rate of surgery during these two time periods, a 4.3% absolute and 57% relative increase from the rate of 7.5% in the first period to 11.8% in the second period are observed. Similar to that observed in the pediatric population, there is a trend in the increase in the rate of surgery among hospitalized adults starting in 2012 that becomes most pronounced in 2014 (Cochran-Armitage Test for Trend from 2012 to 2015, p<0.001). The trend remained consistent among both the pediatric and adult populations for the individual states (Supplemental Figure). The National Inpatient Sample (NIS) was also queried, and a similar decline in the total number of individuals admitted with orbital cellulitis from 2013 to 2014 was noted (10,410 versus 6,850 individuals, respectively).
Patient Demographics and Characteristics
A total of 3,041 pediatric patients were admitted with a diagnosis of orbital cellulitis in all 7 states from 2008–2015 (Table 2). As described in the methods, not all states contributed equally to each year due to incomplete longitudinal patient identifiers. Of those admitted for orbital cellulitis, 397 (13%) underwent surgery, of whom 277 (70%) were male. Children who received surgery were older compared to those who received medical management, with a median age of 10 (IQR 7–13) years compared to 5 (IQR 2–10) years, respectively (p<0.001). Medicaid was the most common payer for both the surgical (n= 190, 48%) and nonsurgical cohort (n= 1434, 54%).
Table 2.
Patient Characteristics and their Association with Surgery on Univariable Analysis
| Pediatric Population (N = 3,041) | Adult Population (N = 7,961) | |||||
|---|---|---|---|---|---|---|
| Patient Demographics | Surgical, N (%) 397 (13) |
Nonsurgical, N (%) 2644 (87) |
OR (99%CI) | Surgical, N (%) 681 (9) |
Nonsurgical, N (%) 7280 (91) |
OR (99%CI) |
| Sex | ||||||
| Male | 277 (70) | 1550 (59) | 1.63 (1.21 – 2.20) | 398 (58) | 3186 (44) | 1.81 (1.47 – 2.23) |
| Age (years)- | ||||||
| Median (IQR) | 10 (7–13) | 5 (2–10) | 51 (35 – 64) | 53 (38 – 70) | ||
| < 5 | 42 (11) | 1127 (43) | Ref | |||
| 5 – 8 | 97 (24) | 640 (24) | 4.07 (2.49 – 6.66) | |||
| 9 – 13 | 161 (41) | 537 (20) | 8.05 (5.05 – 12.82) | |||
| 14 – 17 | 97 (24) | 340 (13) | 7.66 (4.64 – 12.64) | |||
| 18 – 24 | 82 (12) | 617 (9) | Ref | |||
| 25 – 44 | 176 (26) | 1916 (26) | 0.69 (0.48 – 1.00) | |||
| 45 – 64 | 260 (38) | 2399 (33) | 0.82 (0.58 – 1.15) | |||
| 65 – 84 | 141 (21) | 1764 (24) | 0.60 (0.41 – 0.88) | |||
| >= 85 | 22 (3) | 584 (8) | 0.28 (0.15 – 0.54) | |||
| Race | ||||||
| White | 161 (43) | 1005 (40) | Ref | 397 (61) | 4214 (60) | Ref |
| Black | 135 (36) | 665 (26) | 1.27 (0.91 – 1.76) | 140 (21) | 1350 (19) | 1.10 (0.84 – 1.44) |
| Hispanic | 43 (12) | 530 (21) | 0.51 (0.32 – 0.81) | 80 (12) | 956 (13) | 0.89 (0.64 – 1.23) |
| Other | 35 (9) | 314 (13) | 0.70 (0.42 – 1.16) | 40 (6) | 551 (8) | 0.77 (0.50 – 1.20) |
| Median Household Income of Patient Zip Code (by quartile) | ||||||
| 4 (lowest) | 126 (33) | 865 (34) | Ref | 200 (31) | 2204 (31) | Ref |
| 3 | 90 (24) | 609 (24) | 1.02 (0.69 – 1.49) | 152 (23) | 1724 (24) | 0.97 (0.73 – 1.30) |
| 2 | 96 (25) | 593 (24) | 1.11 (0.76 – 1.62) | 160 (24) | 1660 (24) | 1.06 (0.80 – 1.41) |
| 1 (highest) | 67 (18) | 444 (18) | 1.04 (0.68 – 1.57) | 144 (22) | 1446 (21) | 1.10 (0.82 – 1.47) |
| Insurance | ||||||
| Medicare | * | * | - | 182 (27) | 2802 (38) | Ref |
| Medicaid | 190 (48) | 1434 (54) | Ref | 119 (17) | 1493 (21) | 1.23 (0.89 – 1.68) |
| Private | 169 (43) | 999 (38) | 1.28 (0.95 – 1.71) | 264 (39) | 1890 (26) | 2.15 (1.66 – 2.79) |
| Self-Pay | 20 (5) | 126 (5) | 1.20 (0.63 – 2.30) | 63 (9) | 723 (10) | 1.34 (0.91 – 1.98) |
| Other | 18 (4) | 79 (3) | 1.72 (0.85 – 3.47) | 53 (8) | 369 (5) | 2.21 (1.44 – 3.39) |
| Child Ophthalmology Comorbidity | 28 (7) | 58 (2) | 3.38 (1.84 – 6.23) | |||
| Rhinologic/Ophthalmologic Concomitant Diagnosis | ||||||
| Acute Sinusitis | 202 (51) | 361 (14) | 6.55 (4.87 – 8.82) | 222 (33) | 351 (5) | 9.55 (7.41 – 12.30) |
| By Site | ||||||
| Location not specified | 33 (8) | 93 (4) | 2.49 (1.45 – 4.27) | 34 (5) | 106 (1) | 3.56 (2.12 – 5.97) |
| Maxillary | 79 (20) | 116 (4) | 5.42 (3.61 – 8.13) | 101 (15) | 124 (2) | 10.05 (6.99 – 14.44) |
| Ethmoid | 98 (25) | 123 (5) | 6.72 (4.58 – 9.85) | 105 (15) | 74 (1) | 17.75 (11.82 – 26.66) |
| Frontal | 27 (7) | 33 (1) | 5.77 (2.92 – 11.44) | 70 (10) | 46 (1) | 18.02 (10.92 – 29.73) |
| Sphenoid | 15 (4) | 13 (1) | 7.95 (2.96 – 21.30) | 20 (3) | 19 (0) | 11.56 (5.03 – 26.56) |
| Pansinusitis | 61 (15) | 92 (3) | 5.04 (3.21 – 7.91) | 51 (7) | 65 (1) | 8.99 (5.49 – 14.72) |
| Chronic Sinusitis | 185 (47) | 745 (28) | 2.22 (1.68 – 2.95) | 294 (43) | 859 (12) | 5.68 (4.56 – 7.08) |
| Periapical Abscess- | * | 32 (1) | - | 20 (3) | 151 (2) | 1.43 (0.77 – 2.66) |
| Facial Cellulitis | 43 (11) | 312 (12) | 0.91 (0.58 – 1.42) | 103 (15) | 1461 (20) | 0.71 (0.53 – 0.95) |
| Eyelid Abscess | 27 (7) | 120 (5) | 1.54 (0.87 – 2.71) | 44 (6) | 370 (5) | 1.29 (0.84 – 1.97) |
| Acute Conjunctivitis | * | 133 (5) | - | 13 (2) | 636 (9) | 0.20 (0.10 – 0.42) |
| Conjunctival Edema | 20 (5) | 28 (1) | 4.96 (2.30 – 10.68) | 31 (5) | 142 (2) | 2.40 (1.42 – 4.04) |
| Exophthalmos | 68 (17) | 95 (4) | 5.55 (3.59 – 8.58) | 100 (15) | 258 (4) | 4.69 (3.39 – 6.47) |
| Diplopia | 20 (5) | 16 (1) | 8.71 (3.63 – 20.89) | 31 (5) | 98 (1) | 3.50 (2.03 – 6.01) |
| Blindness | * | 12 (1) | - | 33 (5) | 288 (4) | 1.24 (0.76 – 2.01) |
| Optic Neuritis | * | * | - | 16 (2) | 26 (1) | 6.71 (2.94 – 15.32) |
| CN III, IV, and/or VI Palsy | * | * | - | 17 (3) | 34 (1) | 5.46 (2.52 – 11.81) |
| Osteomyelitis | 4 (12) | 59 (2) | 6.17 (3.67 – 10.37) | 30 (4) | 40 (1) | 8.34 (4.44 – 15.68) |
Per HCUP, variables with 11 or fewer patients cannot be reported.
+ Percentages reported are based on available data. Missing data was excluded from analysis.
++ Variables excluded from Table 1 because there were <11 surgical patients with the following diagnosis in both the pediatric & adult population: Papilledema, ophthalmoplegia, visual field defect, sudden vision loss, endophthalmitis.
A total of 7,961 adults were admitted with a diagnosis of orbital cellulitis (Table 2). Of those, 681 (8.55%) underwent surgery, of whom 398 (58%) were male. The median age (IQR) of patients was 51 (IQR 35–64) years and 53 (IQR 38–70) years for those who did and did not receive surgery, respectively. The most common type of insurance for adults requiring surgery was private (n=264, 39%) compared to Medicare for adults receiving medical management (n=2802, 38%).
Hospital-Level Factors
Of the pediatric patients who underwent surgery, 372 (99%) were evaluated at a hospital in an urban location, and 334 (89%) received surgery at a teaching hospital (Table 3). In adults who required surgery, 639 (97%) were treated at a hospital in an urban location, and 488 (74%) were managed at a teaching hospital (Table 3).
Table 3.
Hospital Characteristics and their Association with Surgery in Univariable Analysis
| Pediatric Population (N = 3,041) | Adult Population (N = 7,961) | |||||
|---|---|---|---|---|---|---|
| Surgical, N (%) 397 (13) |
Nonsurgical, N (%) 2644 (87) |
OR (99%CI) | Surgical, N (%) 681 (9) |
Nonsurgical, N (%) 7280 (91) |
OR (99%CI) | |
| Size | ||||||
| Small (<150 beds) | 25 (7) | 139 (5) | Ref | 41 (6) | 583 (8) | Ref |
| Medium (150 – 400 beds) | 67 (18) | 645 (25) | 0.58 (0.30 – 1.11) | 132 (20) | 1719 (24) | 1.09 (0.68 – 1.76) |
| Large (>400 beds) | 284 (75) | 1797 (70) | 0.88 (0.49 – 1.58) | 483 (74) | 4771 (68) | 1.44 (0.93 – 2.22) |
| Location | ||||||
| Rural | * | 122 (5) | - | 17 (3) | 447 (6) | Ref |
| Urban | 372 (99) | 2459 (95) | - | 639 (97) | 6626 (94) | 2.54 (1.33 – 4.83) |
| Teaching | 334 (89) | 1946 (75) | 2.60 (1.68 – 4.02) | 488 (74) | 4250 (60) | 1.93 (1.52 – 2.45) |
| Service | ||||||
| General Medical & Surgical | 322 (85) | 2371 (92) | Ref | 635 (97) | 6919 (98) | Ref |
| Otolaryngology / Ophthalmology | * | * | - | 14 (2) | 82 (1) | 1.86 (0.88 – 3.95) |
| Children’s General Medical & Surgical | 47 (13) | 201 (8) | 1.72 (1.10 – 2.69) | * | * | - |
| Other | * | * | - | * | 69 (1) | - |
| HCUP ED Indicator | 298 (75) | 2054 (78) | 0.87 (0.63 – 1.19) | 495 (73) | 6109 (84) | 0.51 (0.40 – 0.65) |
| ICU | 56 (14) | 73 (3) | 5.78 (3.58 – 9.36) | 116 (17) | 700 (10) | 1.93 (1.46 – 2.56) |
| Weekend Admission | 98 (25) | 691 (26) | 0.93 (0.67 – 1.28) | 139 (20) | 1780 (24) | 0.79 (0.61 – 1.02) |
Per HCUP, variables with 11 or fewer patients cannot be reported.
Multivariable Logistic Regression Model of Surgery
A hierarchical logistic regression model was used to account for variability in individual patient and hospital characteristics on the likelihood for surgery (Tables 4 and 5).
Table 4.
Association of Risk Factors with Surgery in the Pediatric Population in Multivariable Analysis
| Variables | OR (99% CI) |
|---|---|
| Teaching Hospital | 2.13 (1.11 – 4.12) |
| Male | 1.14 (0.78 – 1.68) |
| Age (years) | |
| < 5 | Ref |
| 5 – 8 | 2.38 (1.34 – 4.24) |
| 9 – 13 | 4.43 (2.55 – 7.69) |
| 14 – 17 | 4.55 (2.48 – 8.36) |
| Race | |
| White | Ref |
| Black | 1.34 (0.86 – 2.07) |
| Hispanic | 0.93 (0.51 – 1.71) |
| Other | 0.65 (0.33 – 1.27) |
| Ophthalmology Comorbidity | 2.88 (1.25 – 6.65) |
| Acute Sinusitis By Location*– | |
| Location not specified | 5.06 (2.43 – 10.53) |
| Maxillary | 1.71 (0.79 – 3.72) |
| Ethmoid | 8.51 (4.05 – 17.87) |
| Frontal | 1.04 (0.39 – 2.79) |
| Sphenoid | 2.56 (0.71 – 9.28) |
| Pansinusitis | 10.37 (5.36 – 20.05) |
| Chronic Sinusitis | 4.29 (2.76 – 6.67) |
| Conjunctival Edema | 3.17 (1.09 – 9.25) |
| Exophthalmos | 2.08 (1.15 – 3.75) |
| Diplopia | 2.48 (0.81 – 7.55) |
| Osteomyelitis | 3.13 (1.57 – 6.27) |
The odds for having surgery when a specific sinus site is involved compared to when there is no involvement of the specific sinus of interest
Table 5.
Association of Risk Factors with Surgery in the Adult Population in Multivariable Analysis
| Variables | OR (99% CI) |
|---|---|
| Hospital Location | 2.09 (0.92 – 4.69) |
| Teaching Hospital | 1.67 (1.18 – 2.34) |
| Male | 1.44 (1.12 – 1.86) |
| Age – | |
| 18 – 25 | Ref |
| 25 – 45 | 1.07 (0.67 – 1.71) |
| 45 – 65 | 1.13 (0.72 – 1.77) |
| 65 – 85 | 1.18 (0.65 – 2.12) |
| >= 85 | 0.73 (0.32 – 1.68) |
| Primary Payer - | |
| Medicare | Ref |
| Medicaid | 1.12 (0.69 – 1.81) |
| Private | 1.54 (1.01 – 2.36) |
| Self-Pay | 1.06 (0.61 – 1.87) |
| Other | 1.42 (0.76 – 2.66) |
| Acute Sinusitis By Site*- | |
| Location not specified | 3.88 (2.06 – 7.29) |
| Maxillary | 2.75 (1.50 – 5.04) |
| Ethmoid | 6.49 (3.39 – 12.45) |
| Frontal | 4.47 (2.27 – 8.81) |
| Sphenoid | 3.73 (1.14 – 12.21) |
| Pansinusitis | 8.17 (4.35 – 15.33) |
| Chronic Sinusitis | 5.97 (4.58 – 7.78) |
| Facial Cellulitis | 0.66 (0.47 – 0.93) |
| Acute Conjunctivitis | 0.28 (0.13 – 0.62) |
| Conjunctival Edema | 1.03 (0.52 – 2.04) |
| Exophthalmos | 2.52 (1.65 – 3.83) |
| Diplopia | 1.72 (0.87 – 3.41) |
| Optic Neuritis | 3.21 (1.12 – 9.14) |
| Cranial Nerve III/IV/VI Palsy | 2.65 (1.00 – 7.04) |
| Osteomyelitis | 4.36 (1.93 – 9.85) |
The odds for having surgery when a specific sinus site is involved compared to when there is no involvement of the specific sinus of interest
For pediatric patients, age was a significant predictor of surgery with patients aged 9 years and above significantly more likely to have surgery compared to patients less than 5 years of age (Table 4). When stratified by age, patients aged 14–17 years of age were most likely to receive surgery compared to those less than 5 years of age (OR 4.55, 99% CI 2.48 – 8.36). Of the rhinologic and ophthalmologic signs of orbital cellulitis, acute ethmoid sinusitis and acute pansinusitis were associated with the highest likelihood of surgery (OR 8.51, 99% CI 4.05–17.87 and OR 10.37, 99% CI 5.36–20.05, respectively). Of the hospital-level variables, treatment at a teaching hospital was associated with a 2.1-fold increase in surgery.
For adult patients, males had a higher likelihood of surgery (OR 1.44, 99% CI 1.12–1.86) compared to females (Table 5). Involvement of any of the sinuses was associated with surgery. Of the patients with acute sinusitis who underwent surgery, patients with acute pansinusitis had the highest OR for surgery (OR 8.17, 99% CI 4.35–15.33), followed by acute ethmoid sinusitis (OR 6.49, 99% CI 3.39–12.45), and acute frontal sinusitis (4.47, 99% CI 2.27–8.81). Patients with severe ophthalmic signs of exophthalmos (OR 2.52, 99% CI 1.65–3.83), optic neuritis (OR 3.21, 99% CI 1.12–9.14), and/or CN III, IV, or VI palsy (OR 2.65, 99% CI 1.00–7.04) were more likely to undergo surgery.
DISCUSSION
Orbital cellulitis is a serious infection of the post-septum periorbital structures that can result in vision- and life-threatening sequelae. The treatment of orbital and periorbital infections is an area of debate in the ophthalmology and otolaryngology literature with no consensus over if and when surgical management should be initiated. In this study, we found a dramatic reduction in hospital admissions for orbital cellulitis for both children and adults with a similar albeit smaller reduction in the number of surgeries performed for both populations. As a result, when examining the rate of surgery performed among hospitalized individuals, there was a large increase in the utilization of surgery in both groups during our study period. We also identified patient-level and hospital-level characteristics associated with surgery.
In both the adult and pediatric population, there appears to be an overall increase in the rate of surgery among hospitalized patients starting in 2012 that becomes most apparent in 2014. Interestingly, in 2014, there was also a significant decline in the total number of individuals admitted for orbital cellulitis in our study population as well as in the NIS database, suggesting that while fewer patients are being admitted for orbital cellulitis, those that are hospitalized are more likely to undergo surgery, possibly because either the hospitalized incidence of orbital cellulitis and/or disease severity are changing.
Overall, 9.8% of all patients admitted for orbital cellulitis underwent surgery. In the pediatric cohort, 13% of patients admitted for orbital cellulitis underwent surgery, which is consistent with that reported by Marchiano et al. 10 using the Nationwide Inpatient Database (NIS) from 2002–2010 and Mahalingam-Dhingra et al. 11 using the Kid’s Inpatient Database (KID) from 2006. Consistent with prior studies, children undergoing surgery tended to be male and older than those managed medically. When grouped by age, children 9 years or older were most likely to undergo surgery in multivariable analysis. This finding is consistent with the study done by Garcia et al.2 which found that conservative management was successful in treating subperiosteal abscesses secondary to orbital cellulitis in children less than 9 years of age. Microbial analysis of cultures from children less than 9 years of age with orbital cellulitis by Harris12 demonstrated that children in this age group often have simple, monomicrobial infections compared to more complex, polymicrobial infections seen in older patients. In addition, the growth of the sinus cavities with age results in larger cavities that can serve as a nidus for polymicrobial infections refractive to antibiotic therapy alone.12
Among the paranasal sinuses, involvement of the ethmoid and pansinusitis are most commonly associated with orbital cellulitis.4 Similarly, in our study, among individuals with acute sinusitis, those with acute ethmoid sinusitis or acute pansinusitis were most likely to receive surgery, irrespective of age. In both the pediatric and adult populations, those with more severe ophthalmologic signs were more likely to undergo surgical intervention, likely reflective of the severity and degree of orbital tissue involvement. The median length of stay for both children and adults was also longer for those who underwent surgery, most likely due to the complexity of disease for surgical patients.
Although regional differences in surgical trends have been studied with patients in the western geographic region of the United States more likely to undergo surgical management11, hospital characteristics have not been described in terms of its influence on surgical versus medical management. In our study, the majority of surgical patients were treated at large hospitals (>400 beds) in urban locations. In addition, admission to a hospital with a teaching affiliation was associated with a higher likelihood for surgery. These findings may be reflective of the severity of disease, requiring a higher acuity of care provided by tertiary hospitals with academic affiliations, lower threshold for surgical intervention at tertiary hospitals due to surgeon experience and/or volume, and/or community referral patterns.
A principle limitation of this study is the reliance on the accuracy of ICD-9-CM diagnosis codes. Although the ICD-9-CM diagnosis code for orbital cellulitis was used as the inclusion criteria, this code does not distinguish between orbital abscess and subperiosteal abscess, or other ophthalmologic conditions that could have been coded as such. However, by limiting our study to those patients with the ICD-9-CM diagnosis code for orbital cellulitis, we hope to have captured the spectrum of the disease, including those with the most severe disease. In addition, although imaging is often used to confirm orbital and subperiosteal abscesses, the ICD-9-CM codes for imaging do not provide specific details regarding radiologic findings. Despite these limitations, this is the first study to use the SID database from 2008 to 2015 to study longitudinal hospitalization and surgical trends as well as hospital characteristics and clinical features of pediatric and adult patients with orbital cellulitis who received surgical management.
CONCLUSION
This study describes the longitudinal trend of surgery for orbital cellulitis from 2008–2015 as well as factors associated with surgery in both the pediatric and adult population. In 2014, there was a sharp reduction in the number of patients admitted with orbital cellulitis, suggesting an overall decline in the severity of disease incidence, changes in hospital-level or provider practices, access to care, or other explanations for which are beyond the scope of this paper. Future studies to extend the time frame to the present day would be invaluable in assessing if the rate of surgery will continue to increase or return to historic averages. In addition, examining the indications for surgery from patient chart review would allow for extraction of more granular detail than is possible with a large administrative database. By exploring the utilization trend of surgery for orbital cellulitis as well as factors associated with surgery, we believe the findings herein may help to improve management guidelines and further understanding of this potentially vision- and life-threatening disease.
Supplementary Material
Supplemental Figure. Utilization Trend of Surgery from 2008–2015 for FL and NY demonstrating that while the number of orbital infections and rate of surgery remain steady for the overall cohort in each state between 2008 and 2011, there is an increase in the rate of surgery that starts in 2012 and is most prominent by 2014. Moreover, in 2014, there is a significant decline in the overall number of patients admitted for orbital cellulitis in both FL and NY. Although the trend remains consistent for NE, that data cannot be shown as HCUP prevents the report of data for less than 11 patients.
+ Total number of patients and surgery rates in 2015 are calculated based on the first three-quarters of 2015.
Acknowledgments
This study was supported by the “Development of Clinician/Researchers in Academic ENT” T32 DC00022 from the National Institutes of Deafness and Other Communication Disorders and the Center for Administrative Data Research (CADR), part of the Washington University Institute of Clinical and Translational Sciences which is supported by the NIH/National Center for Advancing Translational Sciences (NCATS) CTSA grant UL1TR002345. Support for CADR was also provided in part by grant R24HS19455 from the Agency for Healthcare Research and Quality (AHRQ). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Footnotes
The authors have no other funding, financial relationships, or conflicts of interest to disclose.
Presented at the Combined Sections Meeting in Del Coronado, California, USA, January 22–24, 2019.
LEVEL OF EVIDENCE: NA
REFERENCES
- 1.Jain A, Rubin PA. Orbital cellulitis in children. International ophthalmology clinics 2001; 41:71–86. [DOI] [PubMed] [Google Scholar]
- 2.Garcia GH, Harris GJ. Criteria for nonsurgical management of subperiosteal abscess of the orbit: analysis of outcomes 1988–1998. Ophthalmology 2000; 107:1454–1456; discussion 1457–1458. [DOI] [PubMed] [Google Scholar]
- 3.Rahbar R, Robson CD, Petersen RA, et al. Management of orbital subperiosteal abscess in children. Archives of Otolaryngology–Head & Neck Surgery 2001; 127:281–286. [DOI] [PubMed] [Google Scholar]
- 4.Yang M, Quah BL, Seah LL, Looi A. Orbital cellulitis in children-medical treatment versus surgical management. Orbit (Amsterdam, Netherlands) 2009; 28:124–136. [DOI] [PubMed] [Google Scholar]
- 5.Taubenslag KJ, Chelnis JG, Mawn LA. Management of frontal sinusitis-associated subperiosteal abscess in children less than 9 years of age. Journal of AAPOS : the official publication of the American Association for Pediatric Ophthalmology and Strabismus 2016; 20:527–531.e521. [DOI] [PubMed] [Google Scholar]
- 6.Nation J, Lopez A, Grover N, Carvalho D, Vinocur D, Jiang W. Management of Large-Volume Subperiosteal Abscesses of the Orbit: Medical vs Surgical Outcomes. Otolaryngology--head and neck surgery : official journal of American Academy of Otolaryngology-Head and Neck Surgery 2017; 157:891–897. [DOI] [PubMed] [Google Scholar]
- 7.Botting AM, McIntosh D, Mahadevan M. Paediatric pre- and post-septal peri-orbital infections are different diseases. A retrospective review of 262 cases. International journal of pediatric otorhinolaryngology 2008; 72:377–383. [DOI] [PubMed] [Google Scholar]
- 8.Healthcare Cost and Utilization Project. Overview of the State Inpatient Databases (SID). Available at: https://www.hcup-us.ahrq.gov/sidoverview.jsp.
- 9.Simon TD, Cawthon ML, Stanford S et al. Pediatric Medical Complexity Algorithm: A New Method to Stratify Children by Medical Complexity 2014; 133:e1647–e1654. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Marchiano E, Raikundalia MD, Carniol ET et al. Characteristics of patients treated for orbital cellulitis: An analysis of inpatient data. The Laryngoscope 2016; 126:554–559. [DOI] [PubMed] [Google Scholar]
- 11.Mahalingam-Dhingra A, Lander L, Preciado DA, Taylormoore J, Shah RK. Orbital and periorbital infections: a national perspective. Archives of otolaryngology--head & neck surgery 2011; 137:769–773. [DOI] [PubMed] [Google Scholar]
- 12.Harris GJ. Age as a factor in the bacteriology and response to treatment of subperiosteal abscess of the orbit. Transactions of the American Ophthalmological Society 1993; 91:441–516. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure. Utilization Trend of Surgery from 2008–2015 for FL and NY demonstrating that while the number of orbital infections and rate of surgery remain steady for the overall cohort in each state between 2008 and 2011, there is an increase in the rate of surgery that starts in 2012 and is most prominent by 2014. Moreover, in 2014, there is a significant decline in the overall number of patients admitted for orbital cellulitis in both FL and NY. Although the trend remains consistent for NE, that data cannot be shown as HCUP prevents the report of data for less than 11 patients.
+ Total number of patients and surgery rates in 2015 are calculated based on the first three-quarters of 2015.