New trends and insights in facial fracture treatment in the United States

Patient selection

The 2012 to 2016 ACS NSQIP participant databases were accessed on April 1st, 2018 and queried to identify all patients undergoing facial bone fractures operations. Table 1 shows the Current Procedural Terminology (CPT) codes used to identify patients with facial bone fracture operations. CPT codes were grouped by anatomical location for analysis. Of the patients undergoing facial bone fracture treatment, 158 patients were excluded to avoid biased outcomes and false complication rates due to concurrent procedures in non-facial areas or to facial bone fixation procedures after osteotomy for head and neck tumor extirpation. Excluded patients were identified using the postoperative diagnosis codes 9 and 10 (ICD 9 & 10). Patients were sorted into a single-facial-fracture group and a multiple-facial-fractures group. In a different analysis, patients were sorted into a plastic surgery specialty group and a non-plastic surgery specialty group. The American Society of Anesthesiologists (ASA) classification, quarter of admission, transfer pattern, surgical specialty, body mass index, diabetes mellitus, hypertension, smoking, functional status, and demographic data were collected and analyzed for each gender group and type of facial fractures in both single and multiple-fractures groups as well as for plastic and non-plastic surgery groups (ENT, orthopedic surgeon, general surgeon, neurosurgeon, vascular surgeon, and urologist).

Surgical specialty

Patient epidemiological data was divided into two groups and compared based on the specialty of the surgeon: patients operated on by a plastic surgeon and patients operated on by a non-plastic surgeon. Furthermore, patients with postoperative complications were divided into two groups and compared based on the specialty of the surgeon as previously stated [Figure 1]. Ophthalmologists, oral & maxillofacial surgeons are not included in the ACS-NSQIP database.

Figure 1. Patient-selection process. From years 2012 to 2016, the ACS-NSQIP contained 3,832,117 cases from different area hospitals in the United States. Of those, 3,354 patients underwent facial fractures treatment.

Outcome and postoperative complications

Complications analysis was made for single/multiple facial fracture groups and plastic/non-plastic groups. Our primary outcomes of interest were operation time and length of stay. Postoperative complications included wound complications, mortality, return to the operating room, and major bleeding. Surgical complications were defined as any wound infections, acute wound dehiscence, graft failure, and need for intraoperative blood transfusion. Medical complications included any defined ACS-NSQIP nonsurgical endpoint, such as pneumonia, pulmonary embolism, postoperative renal insufficiency (creatinine > 2 mg/dL), urinary tract infection, stroke, myocardial infarction, symptomatic deep venous thrombosis, or sepsis. Complications were treated as a dichotomous variable (none vs. one or more). All complications were identified within 30 days of the index procedure.

Statistical analysis

Bivariate analysis was used for epidemiologic and postoperative variables. The t-test was used for continuous variables and the Chi-square test or Fisher’s exact test was performed to compare categorical variables. All statistical analyses were performed using R software version 3.4.4 (http://www.R-project.org) with a statistical significance set at the alpha value of 0.05.

Single-fracture vs. multiple-fracture

Among all facial fracture repair patients, 2,563 (79.6%) had single facial fractures and 657 (20.4 %) had multiple facial fractures. In the single-fracture group, the most common fracture was mandibular. In the multiple-fractures group, there was an average of 2.28 (SD = 0.62) fractures per patient [1,500 fractures in 657 people] and the most common fractures were mandibular and zygomaticomaxillary fractures [Figure 2]. Zygomaticomaxillary fractures were significantly higher in multiple fractures (P = 0.027). In both fracture groups, trends showed a higher prevalence in male patients than in female patients. The most prevalent age group was in the twenties. There was no statistical difference in age or sex between the two groups [Table 2]. Preoperative factors showed no statistically significant difference based on single-vs-multiple fractures except for hypertension (P = 0.0437). There was a statistically significant difference in surgeon specialty between the single and multiple-fracture groups (P < 0.0001).

Surgical specialty

Among the 3,354 cases analyzed in this study, 1,418 (44.0%) patients were operated on by plastic surgeons and 1,802 (56.0%) were operated on by non-plastic surgeons. Plastic surgeons treated zygomatic and orbital fractures more frequently than non-plastic surgeons (P = 0.0004 and P = 0.011, respectively) [Figure 4]. Of single fractures, most were ENT cases, and of the multiple fractures, most were plastic cases [Figure 5]. Patients operated on by plastic surgeons were less likely to be identified as white or Hispanic compared to non-plastic surgeons (P = 0.0253 and P < 0.0001, respectively). There was an association between fracture location and surgical specialty [Figure 6] (P < 0.0001). In regards to preoperative factors, hypertension showed a statistically significant difference based on surgical specialty (P = 0.0109), while smoking, diabetes and Body Mass Index (BMI) showed no statistically significant differences (P = 0.498, P = 0.550 and P = 0.939, respectively). There was a statistically significant difference in patient ASA classification and functional status (P = 0.0352 and P = 0.0445, respectively) based on surgical specialty [Table 3].

Figure 5. Facial fracture treatment prevalence by fracture location and sex. The bar graph above shows the fracture location distribution of facial fracture patients from 2012 to 2016. Most male patients underwent mandibular fracture treatment, while most female patients underwent orbital fracture treatment.

Figure 6. Comparison of outcome. The plastic group had a shorter mean operation time and a longer average length of stay than the non-plastic group (P = 0.0007, P < 0.0001, respectively).

Outcomes and postoperative complications

Plastic surgeons’ operating time was less than that of non-plastic surgeons (P = 0.0007), with a mean duration of 110 min (plastic) and 121 min (non-plastic). The average length of stay was higher for the plastic group, with a mean of 1.65 days (plastic) and 1.01 days (non-plastic) (P < 0.00001) [Figure 7]. From 2012 to 2016, 3,220 cases were evaluated for postoperative complications within 30 days of operation. Out of the 3,220 facial fracture patients, 256 had complications (7.96%). Surgical complication, medical complication, and reoperation rate were 2.67, 0.84, and 2.70%, respectively. The mortality rate was 0.25%. All postoperative complication variables showed no statistically significant differences between the plastic and non-plastic groups [Table 4].

Figure 7. Operating surgical specialty in single vs. multiple-fractures groups. The Y-axis represents the percentage of patients undergoing facial fracture treatment for each surgical subspecialty, which is shown on the x-axis. Of single fractures, most operating surgeons are ENT surgeons, and of the multiple, most are plastic surgeons.

Facial fracture epidemiology

Data analysis showed that male subjects received facial trauma treatment more frequently (76.6%) than female subjects, consistent with previously reported studies^[6-16]. Mandibular fractures were the most common in male patients, whereas orbital fractures were most common in female patients [Figure 5]. Previous studies have found a high prevalence of orbital fractures in women, which may be related to domestic violence^[23]. Mean age was 39.05 years, indicating an older population. Patients in their twenties were the most frequently reported for fracture treatment, possibly due to their increased physical ability and engagement in risky environments [Figure 3]. Our prevalence data for fracture treatment based on race, functional status and ASA classification confirm previous trends and validate their reliability.

Single vs. multiple fracture groups

There were no statistically significant differences in demographic variables between the single and multiple fracture groups. In regards to preoperative conditions, there were statistically significant differences in hypertension between the single and multiple fracture groups; 17.1% of patients with multiple fractures had preoperative hypertension compared to 14.4% in the single fracture group (P = 0.0437). This finding corroborates studies reporting hypertension as a risk factor for decreased bone mineral density (BMD) and bone fractures^[24]. There was a statistically significant association between surgical specialty and the single and multiple fracture groups (P < 0.0001); of single fractures, most were operated on by ENT (57.67%, 1,473 cases) and more than half of multiple fractures were operated on by plastic surgeons (51.29%, 327 cases) [Figure 7].

Plastic vs. non-plastic groups

Epidemiological data showed statistically significant differences in the patient’s reported race by surgical specialty; surgeons who operated on patients identified as white or Hispanic were less likely to be in the plastic group. This could be explained by the shortage and asymmetrical distribution of plastic surgeons in the United States (over 25 million people lack geographic access to the specialty)^[25]. There were no statistically significant differences in patient age, sex, BMI, diabetes, and smoking between the plastic and non-plastic group. There was an association between preoperative hypertension, functional status and ASA classification between the plastic and non-plastic group (P = 0.0109, P = 0.0445, and P = 0.0352, respectively); this further emphasizes plastic surgeons’ set of skills in treating multiple fractures, for which hypertension is a risk factor. ASA classification and functional status findings indicate that plastic surgeons operated more frequently on healthier patients (lower ASA classification number) and less frequently on patients with a dependent-labeled functional status compared to non-plastic surgeons.

Fracture location

Overall, the most common fractures were mandibular then orbital [Figure 2], consistent with previous studies. Traumatic mandibular fractures are common due to the mobility, protuberance and large surface area of the mandible. Interpersonal violence is the leading cause of mandible fractures in men^[26,27]. Figure 4 shows fracture treatment location by surgical specialty. Statistical analysis showed that specialty is dependent on the fracture location, with plastic surgeons operating on orbital, nasal, zygomatic and maxillary areas more frequently whereas non-plastic surgeons treating mandibular, zygomaxillary and frontal fractures more frequently (P < 0.0001). This finding shows again that fracture location treatment seems to be dependent on each specialty’s unique expertise and set of skills.

Postoperative complications and outcomes

Out of the 3,220 facial fracture patients, 256 had complications (7.96%). Surgical complications and unplanned reoperation were the most common (2.67 and 2.70%, respectively). Similar postoperative complication rates have been reported previously^[4,28-30]. Our data analysis showed no statistically significant differences in postoperative complications between the plastic and non-plastic groups (55.5% and 44.5%, respectively). Kim et al. 2013 found similar results for 1,147 patients from 2005 to 2011 in the US^[4].

Our primary outcomes of interest were the average length of stay and mean operation time. A large database is necessary for the length of stay and operation time evaluation since potential variables (such as age and sex) could confound these outcomes^[31]. There was an association between operation time and surgical specialty (P = 0.0007) as well as between average length of stay between surgical specialty groups (P < 0.00001); plastic surgeons’ mean operating time was lower and patients’ average length of stay was longer than the non-plastic group. Prolonged operation time has been correlated with an increased risk of SSI, as well as with increased use of healthcare resources and higher costs^[32,33]. No previous studies have investigated the length of stay in association with postoperative complications of facial fractures or surgical specialty.

Limitations

This study presents important limitations for consideration. Variables in the ACS-NSQIP database are standardized yet incomplete. Outcomes and risk factors that are not continuously collected in the database could be of high significance to the study. These variables include but are not limited to diplopia, malocclusion, hypoesthesia, asymmetry, and malunion. Another important variable is the cause of fracture; this would allow the evaluation of concurrent procedures and co-existing conditions. Moreover, the lack of data captured beyond 30 days postoperatively limits the collection of subacute or delayed complications as well as the overall aesthetic results of treatment. Finally, only patients over the age of 16 were available for assessment in the ACS-NSQIP database. Hence, the data age distribution is truncated and is not representative of the whole patient population. In addition, the sample in this may be considered heterogeneous: it includes patients with different types of facial fractures that cannot be categorized into smaller groups with similar features based on the severity and type of fractures. Further studies may look into these different groups by classifying the fractures by site and severity. Ophthalmologists and Oral & Maxillofacial surgeons are not included in the ACS NSQIP database. As a result, the data analyzed may be an incomplete representation of managed facial trauma injuries in the US. The Comprehensive Face Injury (CFI) score, which assigns a severity mean to each face injury in a reliable, repeatable manner, is one of the significant severity rating methods available in the literature today^[34]. Despite the fact that the ACS-NSQIP database offers a sample of exceptional value for its size, it is likely impossible to accurately characterize the severity of the patients who make up the sample.