Prior fronto-orbital advancement associated with complications from transcranial midface surgery in patients with syndromic craniosynostosis

Our center adopted posterior vault distraction osteogenesis (PVDO) as a first-line intervention for cranial expansion in syndromic craniosynostosis in 2008, and we have a growing cohort of patients undergoing transcranial midface advancement who have not had prior fronto-orbital advancement (FOA). The purpose of this study was to evaluate whether a history of FOA influences the risk profile of transcranial midface advancement in patients with syndromic craniosynostosis. Patients undergoing transcranial fronto-facial advancement from 2000 to 2022 were retrospectively divided into cohorts based on preceding history of fronto-orbital advancement (FOA− and FOA+). Perioperative outcomes including operative time, length of stay, intraoperative dural injury, and complications (Clavien-Dindo score) were compared between groups with appropriate statistics. Thirty-eight patients were included (15 in FOA− group and 23 in FOA+ group). The overall complication rate was 47% (10% minor, 37% major). Compared to the FOA− group, the FOA+ group had a higher incidence of dural tears (65% v 20%, p = 0.006) and major complications (48% v 13%, p = 0.028). These findings were recapitulated in multivariate logistic regression controlling for other predictors. Prior FOA is associated with increased rates of major complications and dural tears in patients with syndromic craniosynostosis undergoing fronto-facial surgery. Options for cranial vault expansion that avoid the frontal region, such as PVDO, may favorably alter the risk profile of fronto-facial advancement.


Introduction
A hallmark of many forms of syndromic craniosynostosis is the development of midface hypoplasia, exorbitism, and a Class III malocclusion, the phenotypic severity of which varies across different syndromes [1,2]. Such patients therefore benefit from fronto-facial corrective surgery to protect the globes and address airway obstruction and malocclusion, while also improving appearance [3][4][5]. However, frontofacial advancement carries significant morbidity [3,[6][7][8]; thus, the type and timing of surgery remains a matter of debate.
Prior studies suggest that subcranial LeFort III advancement may be associated with lower complication rates than transcranial monobloc advancement [9][10][11], and that complications of monobloc have higher severity owing to the potential creation of retrofrontal dead space which increases the risk of intracranial infection [8,11]. Mortality, although rare, has been reported for monobloc at higher rates than LeFort III, up to 4.5% of cases [7]. These findings, however, have not been uniformly replicated [12,13], and some have claimed that differences between institutional practicesincluding age at surgery, type of distraction hardware, and utilization of pericranial flaps -may influence the safety profile. Still, many institutions have limited monobloc indications to cases of midface hypoplasia with severe frontal retrusion and exorbitism [4].
Notably, comparisons of these approaches have largely been conducted in an era when fronto-orbital advancement (FOA) was the preferred initial surgical intervention for syndromic craniosynostosis [14][15][16], implying that these patients have undergone a previous FOA by the time of midface surgery. Prior FOA may complicate subsequent anterior cranial surgery due to the formation of dural adhesions and fibrosis which could increase the risk of durotomy and cerebrospinal fluid (CSF) leak upon reoperation [13].
The adoption of approaches that avoid the frontal region, such as posterior vault distraction osteogenesis (PVDO), has shifted the management paradigm for syndromic craniosynostosis [17,18]. Resulting trends include delayed anterior cranial vault surgery and decreased major craniofacial surgeries in the first 5 years of life [17]. Furthermore, the increased intracranial volume achieved with PVDO may provide the opportunity to delay the age at which a monobloc procedure is sequenced. By operating in a surgically naïve environment without fibrosis and dural adhesions, this approach has the theoretical advantages of mitigating surgical morbidity and potentially decreasing surgical time.
Our center adopted posterior vault distraction osteogenesis (PVDO) as a first-line intervention for cranial expansion in syndromic craniosynostosis in 2008, and we have a growing cohort of patients undergoing transcranial midface advancement who have not had prior fronto-orbital advancement (FOA). The purpose of this study was to evaluate whether a history of FOA influences the risk profile of transcranial midface advancement in patients with syndromic craniosynostosis utilizing a 20-year, single-institution database. A secondary objective was to investigate patient factors and comorbidities which may provide even more granular risk stratification for fronto-facial advancement.

Methods
This study was approved by the institutional review board at the Children's Hospital of Philadelphia. Medical records were reviewed to identify patients with syndromic craniosynostosis undergoing fronto-facial distraction procedures between 2000 and 2022. Fronto-facial procedures included monobloc, facial bipartitions, and multi-level mid-facial osteotomies. Patients were classified into two groups based on surgical history of FOA: the FOA − group contained patients with either no prior cranial vault remodeling or only prior PVDO, and the FOA+ group contained patients with a prior FOA.
Clinical data obtained for each patient included demographic variables, syndromic diagnosis, surgical history including prior craniofacial operations, ventriculoperitoneal (VP) shunt placement/revisions, Chiari decompressions, and tracheostomies. Operative data included distractor type (external halo versus semi-buried), operative time, time under general anesthesia, blood loss, transfusion requirements, fluid resuscitation volumes, length of PICU stay, length of hospital stay, and presence of intraoperative dural tears. Complications occurring in the time between initial surgical intervention and hardware removal were recorded. Complications were graded with the Clavien-Dindo system (Table 1) and classified as minor (Clavien-Dindo I and II) or major (Clavien-Dindo IIIa-V) [19].
Data analysis was conducted using JASP (Version 0.16.3; JASP Team [20]) and R 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria. URL https:// www.R-proje ct. org/). Demographic variables were assessed using descriptive statistics. Univariate analysis was conducted with chisquare and Fisher's exact tests for nominal variables and Mann Whitney U tests for continuous measures. Trends in complications over time were assessed by dividing cases into those occurring 2000-2011 and those occurring 2012-2022 and testing with the chi-squared test. Multivariate logistic regression predicting major complications and dural tears was carried out using backwards stepwise elimination of all variables predictive of the outcome variable at the p < 0.200 level [21]. Multivariate linear regression was utilized to predict operative time and length of stay using the same method.
There were no deaths in either group. Four patients (27%) in the FOA− group had complications, compared to 14 (61%) in the FOA+ group (p = 0.039, Table 4). There was no difference in the incidence of minor complications between groups (p > 0.999), but the incidence of major complications was higher in the FOA+ group (48% vs. 13%, p = 0.028). There were no cases of postoperative CSF leak in the FOA− groups and four cases in the FOA+ group (p = 0.139). Complications in the FOA+ group were more often Clavien-Dindo IIIa or higher; however, the   Fig. 1). The complications of grade IIIa or higher consisted of 4 frontal/temporal abscesses, 3 distractor pin site infections without frank purulence, 2 lumbar drain placements for persistent CSF leak, 2 distractor dislodgements, and 2 ascending infections attributed to nasofrontal fistulas. To assess changes in complication rates over time, cases were divided into two groups (early 2000-2011, n = 14 vs. late 2012-2022, n = 24). There was no difference in major (p = 0.898) or minor (p = 0.067) complication rate across periods (Fig. 2).
There was a trend towards patients over the age of 8 at surgery having more major complications compared to those under 8 (54% vs 24%, p = 0.066). Two of 4 patients with a preoperative tracheostomy experienced deep infections necessitating reoperation, and one other patient developed a CSF leak which was managed nonoperatively.
Multivariate logistic regression found that history of FOA was the sole independent predictor associated with major complications (OR 6.0, 95% CI 1.1-32.6, p = 0.040) and intraoperative dural injury (OR 6.5, 95% CI 1.4-30.9, p = 0.018). Multivariate linear regression was used to identify variables associated with length of stay and operative time. Tracheostomy (p < 0.001), increasing anesthesia time (p = 0.031), and age at surgery (p = 0.002) were associated with longer length of hospital stay. Predictors of prolonged surgical time were early (2000-2011) time period    (p = 0.024), a diagnosis of Apert syndrome (p = 0.022), and male sex (p = 0.021, Table 5).

Discussion
There is debate over the optimal timing and type of surgical approach to address midface hypoplasia in patients with syndromic craniosynostosis. While fronto-facial advancement simultaneously treats airway obstruction, exorbitism, Class III malocclusion, intracranial pressure, and frontofacial esthetics, the procedure carries an elevated risk of serious ascending infection compared to subcranial midface advancements [9][10][11]. Our center adopted posterior vault distraction osteogenesis (PVDO) as a first-line intervention for cranial expansion in syndromic craniosynostosis in 2008, and we have a growing cohort of patients undergoing transcranial midface advancement who have not had prior fronto-orbital advancement (FOA). The data from this study suggests that prior FOA is associated with increased rates of major complications and dural tears in patients with syndromic craniosynostosis undergoing fronto-facial surgery. Secondary craniotomy has been previously implicated in infection and complications after fronto-orbital advancement [22,23] and may mediate morbidity from monobloc via multiple factors. First, dural injury is more common in secondary craniotomy, with 83% of the dural tears observed in this study occurring in patients with a prior FOA. This trend was previously noted by Munabi and colleagues who found that patients undergoing transcranial midface advancement after FOA experienced a 33% increase in dural tear rate, although the authors noted their sample size was not large enough to capture significance [13]. Esparza and Hinojosa noted a similar trend in a large sample of non-syndromic and syndromic craniosynostosis patients, with 93% of dural tears occurring in reoperations [24], some of which progressed to transcranial infection [24,25]. Second, prior frontal surgery leads to scarring and adhesions, increasing the technical challenge of surgery. In this study, prior frontal surgery was the sole independent predictor of dural injury and was also associated with a trend towards longer operations. Prior studies have linked longer operative times to infection in intracranial operations [26,27]; however, no such link was observed in the present study. Finally, as previously suggested by Fearon et al., scarring from prior operations may limit reexpansion during the early postoperative period, exacerbating retrofrontal dead space [28].
This study utilized the Clavien-Dindo classification system [19]. Complications were classified as major if they were graded IIIa or higher. While the difference in complication profile between patients with and without prior history of an FOA was significant, the heterogeneity in complications should be noted. These complications include both infections limited to distractor pin sites which resolved quickly with incision and wash out as well as deep abscesses requiring debridement and longer courses of antibiotics. Certainly, while these complications may be graded similarly owing to requiring an intervention under general anesthesia, the clinical effect on the patient is more profound in the latter. Thus, the findings from this report should be interpreted with this understanding and not overstated, as we are underpowered to make more specific comparisons of directly equivalent complications between groups.
Other patient factors which have been proposed to mediate perioperative morbidity from transcranial midface advancement are age at surgery and history of a preoperative tracheostomy. Age at surgery is confounded by so many factors-disease severity, prior surgery, prior complicationsthat it can be difficulty to study. Early reports prior to the widespread utilization of distraction osteogenesis suggested that monobloc at advanced age should be generally avoided due to complications arising from the retrofrontal dead space, which was presumed to be more significant in older patients [8]. Highly variable complication rates have been reported at various ages, and herein, while a trend was noted that patients over the age of 8 experienced more complications than those under 8, the difference was not statistically significant [8,13].
Tracheostomy as a risk factor for perioperative morbidity has been proposed due to the airway foreign body serving as a nidus for infection [27]. Still, this plausible hypothesis has not been consistently confirmed, with one large twocenter study only reporting 1 tracheostomy-related major complication following intracranial procedures over a 6.5year period [28]. Due to a low incidence of tracheostomy at the time of midface advancement in the present cohort, we are likely underpowered to assess risk in these patients, and conclusions are difficult to draw. Still, 2 of the 4 patients with tracheostomies developed major complications. Our group has broadened perioperative antibiotic prophylaxis to 72 h of ampicillin-sulbactam in this subset of patients in an attempt to counteract this problem, and this continues to be a difficult group of patients to study.
It is worth noting that dividing the cases between early and late study periods had no effect on complication rate, which decreases the likelihood that greater experience or unaccounted for changes in institutional faculty or trends influenced these findings. While this does not suggest the absence of a learning curve, it does suggest that at least in this analysis, surgeon experience was not the major driver of complications. Additionally, use of semi-buried versus external distractors was not a predictor of major complications in this study; however, our institution has transitioned away from semi-buried distractors since 2009, and thus, this study was not designed to rigorously assess this variable. The presence of a VP shunt and history of Chiari decompression was also not associated with complications on multivariate analysis, a reassuring finding.
There are several limitations to the present work. First, by virtue of the diagnostic rarity of syndromic craniosynostosis and paucity of midface distraction procedures, the sample size in this study is relatively small, even over a long study period. This no doubt limits the power of statistical inference. Similarly, due to the limited number of cases, cohorts were not perfectly balanced with respect to diagnoses and surgical history. While multivariate statistics were used to correct for these differences, conclusions drawn should be tempered with the understanding that a multitude of factors both accounted for and unaccounted for impact the development of complications in these patients. One such unbalanced factor between groups which merits mentioning is the non-uniformity in syndromic diagnoses between groups, with more patients with Crouzon syndrome in the FOA− cohort and more patients with Apert Syndrome in the FOA+ cohort. While syndrome type was not a predictor of complications at either the univariate or multivariate levels, the generalizability of these results to all syndromic classes may still be negatively impacted by this imbalance. Finally, as previously mentioned, the study period was intentionally broad so as to capture the largest sample size possible. While there was no difference in complications between early and late study periods, it is possible that underlying institutional shifts have differentially impacted the cohorts. Despite limitations, the results of this large, retrospective cohort study may provide benefit in the process of selecting a surgical approach to address midface hypoplasia in patients with syndromic craniosynostosis.

Conclusions
Prior FOA is associated with increased rates of major complications and dural tears in patients with syndromic craniosynostosis undergoing fronto-facial surgery. Options for cranial vault expansion that avoid the frontal region, such as PVDO, may favorably alter the risk profile of fronto-facial advancement.
Author contribution All the authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Connor S Wagner, Matthew E Pontell, and Michaela Hitchner. The first draft of the manuscript was written by Connor S. Wagner, and all the authors commented on previous versions of the manuscript. All the authors read and approved the final manuscript.
Availability of data and materials The data generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations
Ethics approval and consent to participate This study was approved by the Institutional Review Board at the Children's Hospital of Philadelphia.