Antibiotic Prescribing Patterns and Stewardship Opportunities in Pediatric Facial Trauma: A Five‐Year Retrospective Analysis

ABSTRACT

Objectives

To assess antibiotic prescribing practices in pediatric facial trauma at a tertiary academic center, examine their association with infection outcomes, and identify stewardship opportunities to inform future larger‐scale evaluation of antibiotic use in pediatric facial fractures.

Methods

This retrospective chart review included pediatric patients (< 18 years) presenting via trauma activation with facial trauma to a tertiary academic medical system over 5 years. Patients were identified through the institutional data warehouse. Data collected included demographics, injury patterns, antibiotic prescribing details (route, duration, timing), surgical interventions, and clinical outcomes. The primary outcome was infection development; secondary outcomes included antibiotic‐related complications. Descriptive statistics characterized prescribing patterns, and univariate tests evaluated associations between antibiotic use and outcomes. Multivariable modeling was not performed due to only two infection events. Statistical significance was set at p < 0.05.

Results

Among 223 pediatric patients with facial fractures, the infection rate was low (< 1%). Despite this, 46.2% received antibiotics, though only 32.3% received systemic therapy. Among treated patients, 68.9% received courses exceeding 5 days (median 8 days), and over 70% received intravenous antibiotics. Surgical intervention was most strongly associated with antibiotic use, with surgical patients receiving antibiotics significantly more often than non‐surgical patients. Open fractures were associated with higher utilization and longer durations compared with closed fractures. Both documented infections occurred in antibiotic recipients, though this association was not statistically significant. Notably, 51% of antibiotic recipients were candidates for shorter courses or oral conversion.

Conclusion

These findings suggest potential antibiotic overutilization and highlight substantial opportunities to optimize stewardship in pediatric facial trauma, including limiting prophylactic duration, favoring oral therapy when appropriate, and better targeting antibiotics to high‐risk patients. Prospective, multicenter studies are needed to guide evidence‐based guideline development.

Level of Evidence

Level IV.

1. Introduction

Pediatric facial trauma is a common presentation requiring nuanced treatment approaches. The anatomical and vascular structure of the face confers some inherent protection against infection [1]; however, antibiotic utilization in these cases remains inconsistent across clinical practice. Current literature reveals significant variability in antibiotic prescribing practices with limited evidence‐based guidance specific to pediatric populations.

For facial soft tissue injuries, several studies indicate that topical antibiotics are sufficient for properly cleaned, closed, and dressed wounds, with systemic antibiotics reserved for deeper penetrating injuries or higher‐risk cases. Other studies suggest that with proper cleaning and dressing, antibiotics may not be necessary at all, and antibiotic prophylaxis may only be necessary in certain situations. Soft tissue injuries requiring prophylactic antibiotics include, but are not limited to, salivary duct, dog bites, and some high impact avulsive wounds [2, 3]. Notably, several of these foundational studies, including data on dog bite management [1, 2] derive from mixed‐age or predominantly adult populations; findings may not directly generalize to pediatric patients, who differ in wound healing capacity, immune response, and injury mechanism. Additionally, MRSA infections have become increasingly prevalent in pediatric populations, sometimes necessitating additional prophylactic measures. The type, location, and severity of the wound are major factors influencing medication selection, dosage, and duration [4].

For facial fractures, there has been an approximately 60% increase in antibiotic prescriptions for pediatric patients between 2003 and 2021, despite limited evidence supporting this practice. The Surgical Infection Society specifically advises against routine pre‐ and post‐operative antibiotics for non‐mandibular facial fractures, yet most surgeons continue to prescribe antibiotics in these scenarios [5].

While nearly half of facial traumas become colonized due to exposure to normal oral and respiratory flora, prophylactic antibiotics are not consistently proven to prevent infection. The location of fractures significantly influences antibiotic use patterns, with mandible, frontal sinus, orbital, and LeFort fractures being the most common traumas requiring antibiotic therapy [6].

There exists a clear gap in institution‐specific data characterizing antibiotic use in pediatric facial trauma. This study aims to evaluate current practices at our institution, correlate them with infection outcomes, and identify measurable stewardship opportunities that could inform future, evidence‐based guideline development.

2. Materials and Methods

This study was a retrospective chart review of pediatric patients who presented with facial trauma at an academic, tertiary care, Level 1 hospital system between 2020 and 2025. This study was reviewed by the University of Oklahoma Health Sciences Center Institutional Review Board and was determined to be exempt from full review given its retrospective, de‐identified chart review design, with a waiver of the requirement for informed consent. We specifically identified patients under 18 years of age who arrived via trauma activation and were diagnosed with facial fractures. To minimize selection bias, all patients meeting these eligibility criteria during the study period were included, with no additional exclusions applied. Injury characteristics including location, severity, and mechanism; Facial Injury Severity Score; antibiotic treatment details (medication, dose, duration, route); surgical interventions; clinical outcomes (infection, complications); length of stay; and follow‐up information were abstracted from the electronic health record by a single reviewer. Follow‐up was based on encounters documented within the institutional electronic health record; a standardized minimum follow‐up period was not defined. The primary endpoint was development of infection, defined as a clinical diagnosis documented by the treating provider in the medical record, with a secondary endpoint of complications from antibiotic use.

Descriptive statistics were used to characterize the population and antibiotic prescribing patterns. For continuous variables, means and standard deviations (or medians and interquartile ranges for non‐normally distributed data) were calculated. For categorical variables, frequencies and percentages were reported. Missing data were infrequent and limited to a small number of cases. Analyses were conducted using available data for each variable (complete‐case analysis), with denominators reported for each comparison. No imputation was performed. To analyze associations between antibiotic use and infection outcomes, chi‐square tests, or Fisher's exact tests (for small sample sizes) were used for categorical variables. For continuous variables, t‐tests or non‐parametric equivalents were employed as appropriate. The small number of infection events (n = 2) precluded multivariable logistic regression of infection risk, which is prone to instability, overfitting, and non‐convergence when outcome events are rare; associations with infection were therefore assessed using univariate tests only and reported descriptively. A p‐value of < 0.05 was considered statistically significant for all analyses. All statistical analyses were conducted using SPSS.

3. Results

Over the five‐year study period, we identified 223 pediatric patients with facial trauma who met inclusion criteria. Nearly half (103 patients, 46.2%) received antibiotic therapy during their treatment. Of these, 72 patients (69.9% of antibiotic recipients, 32.3% of the overall cohort) received systemic (intravenous or oral) therapy, while 30 patients (29.1% of antibiotic recipients, 13.5% of the overall cohort) received topical, otic, or ophthalmic agents only. Several notable differences between patients who did and did not receive antibiotics were observed (Table 1). Those treated with antibiotics had more severe injuries overall, reflected in their higher Facial Injury Severity Scores (3.41 ± 4.65 versus 2.06 ± 3.36, p = 0.013).

TABLE 1.

Patient and fracture characteristics by antibiotic use.

Characteristic	No antibiotics	Antibiotics	p
	(n = 120)	(n = 103)
Demographics
FISS score, mean ± SD	2.06 ± 3.36	3.41 ± 4.65	0.013
Fracture location, n (%)
Mandible	8 (6.7)	14 (13.6)	0.084
Maxilla	20 (16.7)	32 (31.1)	0.011
Orbital	26 (21.7)	41 (39.8)	0.003
Nasal	18 (15.0)	26 (25.2)	0.083
Frontal	29 (24.2)	23 (22.3)	0.746
Ethmoid	13 (10.8)	9 (8.7)	0.601
Zygomatic	7 (5.8)	12 (11.7)	0.121
Injury characteristics, n (%)
Open fracture	14 (11.7)	23 (22.5)	0.03
Surgical intervention	21 (17.5)	59 (57.3)	< 0.001
Displaced fracture	52 (43.3)	65 (63.1)	0.003
Soft tissue injury	96 (80.7)	85 (83.3)	0.608

Note: p‐values from Student t‐test (continuous variables) or chi‐square test (categorical variables). Statistically significant values (p < 0.05) shown in bold.

Abbreviation: FISS, facial injury severity score.

Need for surgery emerged as the most powerful predictor of antibiotic prescribing. More than half (57.3%) of antibiotic recipients underwent surgical intervention, compared with 17.5% of those who did not receive antibiotics (p < 0.001). Several specific fracture patterns showed strong associations with antibiotic use. Orbital fractures comprised 39.8% of antibiotic recipients compared with 21.7% of non‐recipients (p = 0.003). Maxillary fractures were similarly overrepresented in the antibiotic group (31.1% versus 16.7%, p = 0.011). Open fractures and displaced fractures occurred more frequently among antibiotic recipients (p = 0.030 and p = 0.003, respectively). Mechanism of injury differed significantly between groups (p < 0.001), with motor vehicle collisions accounting for 35.0% of antibiotic recipients versus 10.8% of non‐recipients.

Cefazolin was the most prescribed antibiotic, accounting for 36.9% of cases (38 patients) (Table 2a). Nearly 30% of patients (30 patients, 29.1%) received only topical, otic, or ophthalmic antibiotics as their primary agent rather than systemic antimicrobial therapy.

TABLE 2a.

Antibiotic prescribing patterns among patients who received antibiotics (n = 103).

Characteristic	n	%
Primary antibiotic prescribed
Cefazolin	38	36.9
Ampicillin	15	14.6
Ear drops (ciprodex, ciproHC, ofloxacin)	11	10.7
Ophthalmic drops (erythromycin, polymyxin)	10	9.7
Bacitracin topical	8	7.8
Ceftriaxone	6	5.8
Amoxicillin/amoxicillin‐clavulanate	5	4.9
Other	10	9.7
Route of administration
Primary route—Intravenous	64	62.1
Primary route—Oral	8	7.8
Primary route—Topical/Otic/Ophthalmic	30	29.1
Received IV at any point during treatment	75	72.8
Systemic antibiotics (IV or Oral)	72	69.9
Duration of longest antibiotic course
Mean ± SD, days	9.38 ± 5.42	—
Median (range), days	8.0 (0–23)	—
Duration categories
Single dose/perioperative (≤ 1 day)	1	1
Short course (2–5 days)	31	30.1
Extended (> 5 days)	71	68.9

Abbreviations: IV, intravenous; SD, standard deviation.

Intravenous antibiotic use was strongly correlated with surgical status. Among patients who underwent surgery, 93.2% received IV antibiotics compared with 45.5% of non‐surgical patients (p < 0.001). Open fractures also drove higher IV utilization, with 91.3% of these patients receiving IV therapy versus 67.1% of those with closed fractures (p = 0.043).

The duration of antibiotic therapy was lengthy. On average, patients who received antibiotics received them for 9.38 ± 5.42 days, with a median of 8 days and a range spanning from 0 to 23 days. Only 1 patient (1.0%) received a single perioperative dose, while 31 patients (30.1%) received short courses of 2–5 days. The substantial majority (71 patients, 68.9%) received extended antibiotic courses exceeding 5 days.

Certain injury characteristics significantly influenced antibiotic duration (Table 2b). Open fractures led to the longest courses, averaging 12.74 ± 5.86 days compared with 8.43 ± 4.94 days for closed fractures (p < 0.001). Displaced fractures and soft tissue injuries also extended treatment duration (10.31 ± 5.75 versus 7.79 ± 4.44 days, p = 0.022; and 9.94 ± 5.54 versus 6.53 ± 3.69 days, p = 0.017, respectively). Interestingly, surgical intervention did not significantly affect antibiotic duration, with surgical patients averaging 9.73 ± 6.18 days versus 8.91 ± 4.22 days for non‐surgical patients (p = 0.450).

TABLE 2b.

Antibiotic prescribing patterns among patients who received antibiotics by characteristic.

Characteristic	Mean ± SD (days)	p
Open fracture
No (n = 79)	8.43 ± 4.94	< 0.001
Yes (n = 23)	12.74 ± 5.86
Displaced fracture
No (n = 38)	7.79 ± 4.44	0.022
Yes (n = 23)	10.31 ± 5.75
Soft tissue injury
No (n = 17)	6.53 ± 3.69	0.017
Yes (n = 86)	9.94 ± 5.54
Surgical intervention
No (n = 44)	8.91 ± 4.22	0.45
Yes (n = 59)	9.73 ± 6.18	72.8

Note: p‐values from independent samples t‐tests. Statistically significant values (p < 0.05) shown in bold.

Despite the high rate of IV antibiotic administration, the mean duration of therapy did not differ significantly between patients who received IV antibiotics and those who did not (9.61 ± 5.83 versus 8.75 ± 4.16 days, p = 0.475), with both groups having a median duration of 8 days. This observation pointed toward a significant stewardship opportunity. Among the 75 patients who received IV antibiotics, we identified 53 (70.7%) as potential candidates for IV‐to‐oral conversion or shorter antibiotic courses. These conversion candidates were defined as patients who received IV antibiotics but did not have open fractures and received more than 1 day of therapy. This represents 51.5% of all patients who received antibiotics in our study—a substantial target for intervention (Table 4).

TABLE 4.

Intravenous antibiotic use and associated clinical characteristics among patients who received antibiotics (n = 103).

Characteristic	No IV	Received IV	p
	(n = 28, 27.2%)	(n = 75, 72.8%)
Injury characteristics, n (%)
Open fracture	2 (7.1)	21 (28.0)	0.043
Surgical intervention	4 (14.3)	55 (73.3)	< 0.001
Displaced fracture	16 (57.1)	49 (65.3)	0.447
Soft tissue injury	23 (82.1)	63 (84.0)	0.817
Duration of therapy
Mean ± SD, days	8.75 ± 4.16	9.61 ± 5.83	0.475
Median, days	8	8	—
Range, days	0–15	3–23	—
Clinical outcomes, n (%)
Infection	0 (0.0)	2 (2.7)	—
Any complication	3 (10.7)	6 (8.0)	0.704

Note: p‐values from chi‐square test (categorical variables) or independent samples t‐test (continuous variables). Bolded values are statistically significant (p < 0.05).

Abbreviations: IV, intravenous; SD, standard deviation.

The overall infection rate was remarkably low at 0.9% (2 of 223 patients; Table 3). Notably, both infections occurred in patients who received antibiotics (2 of 103, 1.9%) rather than in those who did not receive antibiotics (0 of 120, 0%), though this difference was not statistically significant (Fisher's exact test, p = 0.212). Because only two infection events occurred, multivariable modeling of infection risk factors was not feasible, and associations with infection were assessed using univariate tests only. The two patients who developed infections had less severe injuries compared with those who may have been considered at higher risk for infection. Overall complications occurred in 25 patients (11.2%), with no significant difference between those who received antibiotics (9 of 103, 8.7%) and those who did not (16 of 120, 13.3%; p = 0.278).

TABLE 3.

Clinical outcomes by antibiotic use.

Outcome	Overall	No antibiotics	Antibiotics	p
	(N = 223)	(n = 120)	(n = 103)
Infection, n (%)
Yes	2 (0.9)	0 (0.0)	2 (1.9)	0.212
No	221 (99.1)	120 (100.0)	101 (98.1)	—
Any complication, n (%)
Yes	25 (11.2)	16 (13.3)	9 (8.7)	0.278
No	198 (88.8)	104 (86.7)	94 (91.3)	—

Note: p‐values from Fisher exact test (infection) and chi‐square test (complications). Neither outcome showed statistically significant differences between groups.

When examining outcomes specifically by route of administration among antibiotic recipients, both infections occurred in patients who received IV antibiotics (2 of 75, 2.7%), with no infections in the non‐IV group (0 of 28, 0%). Complication rates were similarly low in both groups, with 6 of 75 (8.0%) IV recipients and 3 of 28 (10.7%) non‐IV recipients experiencing complications (p = 0.704).

4. Discussion

This study suggests a potential mismatch between antibiotic use and actual infection risk in pediatric facial trauma. Although the observed infection rate was below 1%, nearly half of the cohort received antibiotics. Among those treated, most received extended courses exceeding 5 days (68.9%), and more than 70% received intravenous (IV) antibiotic therapy at some point during their care. Notably, the two documented infections occurred in patients who had received antibiotics rather than in untreated patients; however, this finding was not statistically significant. Nevertheless, these patterns raise important questions regarding the appropriateness of antimicrobial stewardship practices in pediatric facial trauma, particularly given that more than half of antibiotic recipients were identified as potential candidates for a shortened or less invasive regimen.

The low infection rate observed in this study aligns with existing literature, which reports infection rates ranging from 0% to 3% in similar pediatric facial trauma populations. These findings reinforce that infection remains an uncommon complication in this setting. Despite this, antibiotic prescribing remains prevalent. Prior studies have reported antibiotic use rates between 30% and 90%, and the 46.2% rate observed in this cohort falls squarely within that range.

The duration of antibiotic therapy was among the most striking findings. A median treatment duration of 8 days exceeds standard surgical prophylaxis recommendations, which usually call for discontinuation of antibiotics within 24 h postoperatively. This suggests prophylactic antibiotic regimens may evolve into extended empiric therapy even without clinical infection.

Surgical intervention was most strongly associated with antibiotic use, with 73.8% of surgical patients receiving antibiotics compared with 14.7% of non‐surgical patients. While this association is intuitive given concerns for perioperative infection, it does not fully explain the prolonged treatment durations observed among surgical patients (median of 8 days) that was not significantly different from those not undergoing surgery. This interpretation is further supported by the lack of a meaningful difference in antibiotic duration between surgical and non‐surgical patients (9.73 vs. 8.91 days), reinforcing the concern that perioperative prophylaxis may be extended beyond its intended time frame.

Fracture severity and pattern were also associated with increased antibiotic use. Specifically, patients with open fractures received antibiotics more frequently and for longer durations compared with those with closed fractures (13 versus 8.5 days). Although this approach may seem logical, it is not clearly supported by infection outcomes in this cohort, as infections were rare and occurred exclusively in patients who received antibiotics.

Mechanism of injury deserves comment. With only two documented infections in this cohort, it is difficult to draw any meaningful conclusion about which mechanisms carry higher infection risk. Of the two infections, one occurred following a gunshot wound, a penetrating mechanism, and the other followed a fall. Penetrating mechanisms, such as gunshot wounds and animal bites, are biologically plausible candidates for elevated infection risk given direct tissue contamination at the time of injury, but this cohort cannot meaningfully distinguish that possibility from chance given the small number of events. Future research adequately powered to examine mechanism as an independent risk factor, ideally through multicenter collaboration generating a larger pool of infection events, would be needed to determine whether specific mechanisms, particularly penetrating or contaminated wounds, warrant a different antibiotic stewardship approach than blunt or closed injuries.

IV antibiotic use was notably high, with nearly 75% of antibiotic‐treated patients receiving IV therapy (Table 4). Although several injury characteristics overlap with those examined in Table 1, the comparison here addresses a distinct clinical question, route of administration among patients already receiving antibiotics, rather than the decision to prescribe antibiotics at all. In routine practice, most facial fractures are amenable to oral antibiotic management when antibiotics are indicated. Despite this, there was an overwhelming preference for IV therapy, particularly among surgical patients. This finding further supports the notion that perioperative antibiotics may have been continued beyond the immediate perioperative period. Conceptually, IV antibiotics should be reserved for cases requiring prolonged therapy or severe infections; however, in this study, surgical status, not infection severity, was the primary driver of IV use. This represents a clear opportunity for improved stewardship interventions, including earlier IV‐to‐oral conversion in clinically stable patients.

The paradoxical finding that both infections occurred in antibiotic‐treated patients warrants discussion. This observation likely reflects several factors rather than suggesting antibiotics are harmful. First, patients who received antibiotics had inherently more severe injuries with higher baseline infection risk, as evidenced by higher FISS scores, more open fractures, and more surgical interventions. Second, with only two total infections, statistical power to detect any protective effect was severely limited. Third, selection bias likely influenced prescribing patterns, with clinicians prescribing antibiotics to patients perceived as high‐risk based on clinical features not fully captured in our data. This finding underscores that antibiotics alone do not eliminate infection risk and highlights the importance of comprehensive infection prevention strategies including meticulous wound care and appropriate surgical technique.

Several practice improvement opportunities emerge from these findings. First, targeted educational initiatives emphasizing the low baseline infection risk in pediatric facial trauma may help reduce unnecessary prolonged antibiotic courses (> 5 days). Second, establishing criteria to identify candidates for early IV‐to‐oral antibiotic conversion could improve stewardship practices, reduce costs, and enhance patient experience; whether this would affect infection risk cannot be determined from this dataset and would require prospective evaluation. Our identification of 53 patients (51.5% of all antibiotic recipients) as conversion candidates represents a concrete, measurable stewardship target. As noted previously, cefazolin was the most frequently administered antibiotic among the included patients. Preoperative cefazolin was excluded from the conversion analysis, as its administration reflects evidence‐based practice for perioperative surgical prophylaxis. Postoperative cefazolin and all other eligible antibiotics were retained in our conversion analysis. Third, the relatively high (46.2%) antibiotic utilization rate in the context of an infection rate below 1% warrants critical evaluation. Development of clearer institutional guidelines, particularly for closed, minimally displaced fractures without significant soft tissue injury, may reduce unnecessary antibiotic exposure while maintaining excellent clinical outcomes.

This study has several limitations. As a single‐institution, retrospective study, our findings may be subject to selection bias, since referral patterns, surgeon preferences, and prescribing culture at our institution may not reflect practices elsewhere, limiting the generalizability of our results. Its retrospective design may be subject to unmeasured confounding, and the small number of infections limited our ability to analyze infection risk factors or assess the protective effects of antibiotics. Given the extremely low infection rate observed (0.9%, 2 of 223 patients), statistical power to detect associations with infection was limited; multivariable modeling of infection risk was not undertaken, as logistic regression performed with so few outcome events is prone to instability and overfitting. In addition, infection outcomes were based on documented diagnoses, which may have missed subclinical or unreported cases, and follow‐up duration in this cohort was not standardized; without longitudinal surveillance beyond the immediate treatment period, delayed infections presenting outside our system may have gone undetected, potentially underestimating the true infection rate. We also lacked data on antimicrobial resistance and could not assess patient compliance with prescribed regimens or distinguish between prophylactic and therapeutic intent.

Larger prospective studies are needed to better define true infection risk and should stratify patients by injury characteristics to identify those most likely to benefit from prophylactic antibiotics. Ultimately, randomized trials comparing antibiotic strategies, including shortened courses or no antibiotics for low‐risk patients, would provide the strongest evidence‐based recommendations. Multicenter studies would further enhance generalizability and clinical applicability. Nonetheless, our identification of 53 patients (51.5% of antibiotic recipients) as potential conversion candidates represents a measurable, hypothesis‐generating target that could inform future prospective evaluation of stewardship interventions in this population.

5. Conclusion

In summary, we examined pediatric patients with facial trauma over a defined period, captured detailed antibiotic prescribing practices, including route and duration, and identified specific, measurable stewardship opportunities based on real‐world patterns. The low infection rate observed is consistent with prior reports. These findings are hypothesis‐generating and derive from a single‐institution retrospective cohort; confirmation in larger, prospective, multicenter studies is needed before they can inform practice changes.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

LaMar J., Schaefer J., Henkle G., Hasanjee A., Ramjee V., and Mims M., “Antibiotic Prescribing Patterns and Stewardship Opportunities in Pediatric Facial Trauma: A Five‐Year Retrospective Analysis,” Laryngoscope Investigative Otolaryngology 11, no. 5 (2026): e70574, 10.1002/lio2.70574.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.