26 yr experience treating frontal sinus fractures_jh paper

$: 26 yr experience treating frontal sinus fractures_jh paper$
PEDIATRIC/CRANIOFACIAL

Twenty-Six–Year Experience Treating FrontalSinus Fractures: A Novel Algorithm Based onAnatomical Fracture Pattern and Failure ofConventional Techniques

Eduardo D. Rodriguez, M.D.,D.D.S.

Matthew G. Stanwix, M.D.Arthur J. Nam, M.D.

Hugo St. Hilaire, M.D.Oliver P. Simmons, M.D.Michael R. Christy, M.D.

Michael P. Grant, M.D.Paul N. Manson, M.D.

Baltimore, Md.

Background: Frontal sinus fracture treatment strategies lack statistical power.The authors propose statistically valid treatment protocols for frontal sinusfracture based on injury pattern, nasofrontal outflow tract injury, andcomplication(s).Methods: An institutional review board–approved retrospective review wasconducted on frontal sinus fracture patients from 1979 to 2005. Fractureswere categorized by location, displacement, comminution, and nasofrontaloutflow tract injury. Demographic data, treatment, and complications werecompiled.Results: One thousand ninety-seven frontal sinus fracture patients wereidentified; 87 died and 153 were excluded because of insufficient data,leaving a cohort of 857 patients. The most common injury was simultaneousdisplaced anteroposterior walls (38.4 percent). Nasofrontal outflow tractinjury constituted the majority (70.7 percent), with 67 percent having adiagnosis of obstruction. Of the 857 patients, 504 (58.8 percent) underwentsurgery, with a 10.4 percent complication rate; and 353 were observed, witha 3.1 percent complication rate. All complications except one involvednasofrontal outflow tract injury (98.5 percent). Nasofrontal outflow tractinjuries with obstruction were best managed by obliteration or cranialization(complication rates: 9 and 10 percent, respectively). Fat obliteration andosteoneogenesis had the highest complication rates (22 and 42.9 percent,respectively). The authors’ treatment algorithm provides a receiver operat-ing characteristic area under the curve of 0.8621.Conclusions: A frontal sinus fracture treatment algorithm is proposed andstatistically validated. Nasofrontal outflow tract involvement with obstructionis best managed by obliteration or cranialization. Osteoneogenesis and fatobliteration are associated with unacceptable complication rates. Observa-tion is safe when the nasofrontal outflow tract is intact. (Plast. Reconstr.Surg. 122: 1850, 2008.)

Frontal sinus fracture management hasevolved with greater emphasis on nasofrontaloutflow tract injury.1–20 Complications are de-

scribed; however, series have insufficient power to

allow significant conclusions to be drawn regard-ing treatment effectiveness based on fracture pat-terns with consideration for nasofrontal outflowtract injury. These fractures represent 5 to 15percent of all craniomaxillofacial fractures, withhigh-velocity blunt force representing themajority.1,3,21–23 Concomitant facial fractures alongwith associated intracranial and bodily injuriesconfirm their severity.5–7,24

From the R Adams Cowley Shock Trauma Center, Universityof Maryland School of Medicine, Johns Hopkins School ofMedicine.Received for publication April 4, 2008; accepted May 6,2008.Presented at the 87th Annual Meeting of the AmericanAssociation of Plastic Surgeons, in Boston, Massachusetts,April 5 through 8, 2008.Copyright ©2008 by the American Society of Plastic Surgeons

DOI: 10.1097/PRS.0b013e31818d58ba

Disclosure: None of the authors has a financialinterest in any of the products, devices, or drugsmentioned in this article.

www.PRSJournal.com1850

Historically, plain radiographs were diagnos-tic, but surgery provided definitive evaluation ofthe extent of injury and nasofrontal outflow tractinvolvement.25–27 Computed tomographic scansimproved assessment of posterior wall and naso-frontal outflow tract injury, but some uncertaintystill exists. Function is difficult to predict based onradiographic or clinical fracture patterns alone.Nasofrontal outflow tract patency has served as theclinical substitute for determining frontal sinusfunction. With nasofrontal outflow tract injurynoted as 13 to 55 percent,8,10 preoperative deter-mination of patency is crucial to management.Computed tomographic scans often delineatefracture involvement and injury but are unable topredict function. Preoperative nasofrontal out-flow tract computed tomographic assessment var-ies between providers, but there are three indica-tors of injury: gross outflow tract obstruction,frontal sinus floor fracture, and anterior table me-dial wall fracture.28–30 One indicator is sufficient todiagnose nasofrontal outflow tract injury, but todate no one has described the importance of one,two, or all criteria, or the significance of one overthe others. Thus, although nasofrontal outflowtract injury has become the cornerstone of treat-ment algorithms, its degree of injury in relation tocomplications has not been well established.

Treatment strategies stem from Rohrich andHollier’s landmark 1992 article and our descrip-tion of panfacial fracture management.3,31–33 Roh-rich provided insight into fracture pattern, extentof injury, and nasofrontal duct involvement. Ulti-mately, treatment decisions depend on fracturetype, comminution, degree of posterior table frac-ture, nasofrontal duct injury, neurologic status,and cerebrospinal fluid leak.1,3,10,21,22,34–37

EXISTING TREATMENT OPTIONS1. Observation by serial computed tomography

and/or clinical evaluation.2. Reconstruction implies duct and mucosal preser-

vation with anterior wall reconstruction. Thismay include in situ fragment elevation or micro-plating of a comminuted fracture. Recently, en-doscopic management has been used; however,there are few data to support its benefit.38–40

3. Obliteration involves complete removal of sinusmucosa; burring the sinus walls to eliminatemucosal invaginations; plugging the nasalfrontal ducts; and filling the sinus cavity of thesinus with fat, muscle, bone, or alloplasts.

4. Osteoneogenesis in which the sinus cavity isstripped of mucosa by burring the walls, the

ducts are sealed, and the cavity is preserved.Spontaneous obliteration occurs in a delayedfashion by the slow process of scar tissue andbone formation in the empty cavity.

5. Ablation (or exenteration), as Reidel describedin 1898, with removal of the anterior wall,mucosa, supraorbital rims, and proximal nasalbones to allow skin involution against the pos-terior wall or dura.41 The subsequent delayedreconstruction is difficult because of commu-nication with the nasal cavity and a significantcosmetic defect. Presently, the only potentialindication for the Reidel procedure is severeacute infection, where collapse of the deadspace, removal of infected, nonvascularizedbone, and protection from both mucocele andinfection is provided.

6. Cranialization removes the ducts, posterior sinuswall, and mucosa; the nasofrontal outflow tract isblocked with bone, creating a partition betweenthe intracranial and nasal cavities. The area oc-cupied by the sinus, in fact, becomes a portion ofthe intracranial cavity and becomes occupied bythe expanding brain after several months.42

Each treatment method has its proponents,and there is much controversy regarding indica-tions, applications, and ultimate success in givensituations. Although observation is pursued fre-quently, complication rates related to fracture pat-terns have not been recognized. We propose astatistically validated treatment algorithm for fron-tal sinus fractures based on fracture pattern, de-gree of nasofrontal outflow tract involvement, andcomplications (Fig. 1).

PATIENTS AND METHODSAn institutional review board–approved retro-

spective study of patients with frontal sinus frac-tures admitted to the R Adams Cowley ShockTrauma Center from 1979 to 2005 was conducted.Associated injuries, demographics, Injury SeverityScore, Glasgow Coma Scale score, and mechanismof injury were evaluated. Computed tomographicscans were reviewed and categorized. Nasofrontaloutflow tract injury was defined by one or more ofthe following: outflow tract/ductal “obstruction,”frontal sinus floor fracture, or fracture of the medialaspect of the anterior table (Fig. 2). Coronal andsagittal views were inspected before axial cuts toavoid bias of nasofrontal outflow tract injury basedon fracture patterns. Treatments included observa-tion, reconstruction with outflow tract and mu-cosal preservation, obliteration, osteoneogen-esis, and cranialization.

Volume 122, Number 6 • Frontal Sinus Fracture Treatment

1851

Exclusion criteria included the following: pa-tients with incomplete radiographic or clinicalrecords, patients who died within 48 hours withouttreatment, and patients who underwent decom-pressive craniectomies without subsequent frontalsinus surgery. Acute major complications were de-fined as those occurring within 6 months needingoperative management, including cerebrospinal

fluid leak, abscess, sinusitis, meningitis, mucocele,and persistent pneumocephalus. Minor complica-tions included wound drainage, infections or cere-brospinal fluid leak that did not require surgicaltreatment, and late contour irregularity or plate re-moval for palpability or visibility. Unless noted as aminor complication, the term “complication” in thisarticle refers to acute major complications only.

Fig. 1. Frontal sinus fracture treatment algorithm. NFOT, nasofrontal outflow tract.

Fig. 2. Demonstration of computed tomographic diagnosis of nasofrontal outflow tract injury with(left) fracture of the sinus floor, (middle) fracture of the medial aspect of anterior table, and (right)frank obstruction.

Plastic and Reconstructive Surgery • December 2008

1852

Associations were based on Pearson’s correla-tion coefficients, and outcome measures were eval-uated by means of chi-square and Gini-square dis-tribution, recursive partition tree, and receiveroperator characteristic curve. Chi-square analysismeasured the degree of homogeneity (entropy orclassification error), and Gini-square analysis repre-sented the likelihood ratio (no complication dividedby complications). A recursive partition tree is usedas a statistical method for the multivariable anal-ysis of medical diagnostic tests and used in ouranalysis to distinguish patients with and withoutcomplications.43 This separates data into homoge-neous groups by creating smaller subsets based onvarious frontal sinus variables. Receiver operatingcharacteristic area under the curve plots the sensi-tivity (true classifications) against misclassifications(JMP 7.0.1 software; SAS Institute, Inc., Cary, N.C.).The receiver operating characteristic area under thecurve indicates the diagnostic accuracy of the deci-sion algorithm of frontal sinus fracture manage-ment. In other words, it statistically discriminatesbetween patients with and without complicationsundergoing treatment (e.g., observation, recon-struction, obliteration). A receiver operating char-acteristic value of greater than 0.75 indicates diag-nostic accuracy.

RESULTSOne thousand ninety-seven frontal sinus frac-

tures were evaluated. One hundred fifty-three in-dividuals had insufficient data, and there were 87deaths, leaving 857 patients (Table 1). Caucasianpatients constituted 61 percent, African Americanpatients constituted 27 percent, Hispanics pa-tients constituted 11 percent, and Asians patientsconstituted 1 percent. The average length of con-valescence was 114 days before return to functionor work and the average follow-up was 2.8 months(range, 0 to 66 months) (Table 2). Most patientshad surgery within 24 hours of injury (59 percent)and only 19 percent had surgery after 72 hours.

The most common mechanism of injury wasmotor vehicle collision (42 percent), unrestrained

(60 percent) and restrained (22 percent), with 18percent unknown. Ten percent of the patients wereinvolved in motorcycle collisions, 8 percent werepedestrians, 14 percent were victims of assault, and7 percent had ballistic injuries. Environmental acci-dents accounted for the remaining 19 percent, ofwhich 70 percent included falls and miscellaneoustrauma. Associated injuries (Table 3) were dividedbased on nasofrontal outflow tract injury. Adjacentfacial fractures are also listed in Table 3, and naso-frontal outflow tract–injured patients had threetimes more concomitant facial fractures than thosewithout. Frontal sinus fractures were highly associ-ated with nasoorbitoethmoid and orbital roof inju-ries, and more so with nasofrontal outflow tract in-volvement.

Five hundred four patients (59 percent) under-went surgery, 94 percent had nasofrontal outflowtract involvement by one or more criteria, 77 percenthad obstruction as one of the criteria, and 80 percenthad two or more criteria. Cranialization was the mostcommon treatment, obliteration was second, andreconstruction was third. Most obliterations hadbone (34 percent), galea (25 percent), fat (23 per-cent), or temporoparietal fascia (7 percent). Onepatient had six procedures, one had five procedures,

Table 1. Total Patient Series with FrontalSinus Fractures

No. %

Total 1097 100Insufficient data 153 13.9Mortality 87 7.9Analyzed series 857 78.1Surgery 504 45.9Observation 353 32.2

Table 2. Patient Averages

Value

ISS 25.3 � 11.3Age (yr) 33.5 � 17GCS score 12.1 � 4.7LOS (days) 8.8 � 10.3Follow-up (mo) 2.8 � 2.1Mortality 7.9%ISS, Injury Severity Score; GCS, Glasgow Coma Scale; LOS, length ofstay (inpatient).

Table 3. Associated Injuries in Frontal SinusFractures with and without Nasofrontal OutflowTract Injury

InjuryWithout NFOT

Injury (%)With NFOTInjury (%)

Brain 31 76Cervical spine 7 14Upper extremity fracture 15 25Lower extremity fracture 13 23Pneumothorax 12 24Abdominal 7 13Orbital roof 13 40Orbital wall 7 13Orbital floor 2 7NOE 12 31Zygoma 8 18Le Fort 2 17Mandible 3 5NOE, nasoorbitoethmoid complex; NFOT, nasofrontal outflow tract.


1853

2.6 percent had four procedures, 4.9 percent hadthree procedures, and 7.1 percent had two proce-dures. Thus, a major complication requiring surgeryhad a 69 percent chance of having a third operation,and if so, an additional 53 percent risk of a fourth.

Fracture type, treatment, and complicationsare summarized in Tables 4 through 9. The mostcommon fracture pattern was simultaneous an-teroposterior displaced (38.4 percent) followed byanterior nondisplaced (21.6 percent). Side orcomminution did not play a role in any associa-tions more than displacement alone.

Our series had 61 major complications (7.1percent), and all except one had nasofrontal out-flow tract injury (1.6 percent). Likewise, there wasonly one complication in the nasofrontal outflowtract injury group that did not have obstruction bycriteria (1.6 percent). Patients lacking nasofrontaloutflow tract injury had no complications whenobserved and those with nasofrontal outflow tractinjury by obstruction had the lowest complicationrates with obliteration or cranialization. Further-more, patients with nasofrontal outflow tract in-jury diagnosed by obstruction plus another crite-rion treated with observation or reconstructionhad a 100 percent complication rate. Likewise,those managed by osteoneogenesis had a 56 percentrisk of complication compared with obliteration (10percent) and cranialization (9 percent). Seventy-one percent of all patients had nasofrontal outflow

tract injury, 26 percent had one criterion, 25 percenthad two criteria, and 49 percent met all criteria.Overall, 67 percent of those with nasofrontaloutflow tract injury had obstruction accordingto criteria, with a 14 percent complication rate.

The receiver operating characteristic area underthe curve of our algorithm equals 0.8621, represent-ing statistically significant diagnostic accuracy (Fig.3). Gini-square and chi-square analysis of those with-out nasofrontal outflow tract injury treated by ob-servation according to our algorithm was 0, creatingan undefined p value calculation. Those with naso-frontal outflow tract injury and obstruction treatedby cranialization or obliteration had a Gini-squareand chi-square value equal to 215.97, calculated as2[32 * ln (32/360) � 328 * ln (328/360)], and acorresponding value of p � 0.0001. Those with na-sofrontal outflow tract injury but no obstructionshould be treated by reconstruction if the fracture isdisplaced (p value undefined, chi-square � 0) andobservation if the fracture is nondisplaced (p �0.00124, chi-square � 7.35).

The frontal sinus fracture groups can be clas-sified as follows:

1. Anterior wall fractures.

A. Nondisplaced anterior wall fractures(185 patients) (Table 4).

B. Displaced anterior wall fractures (143patients) (Table 5).

Table 4. Nondisplaced Anterior Wall Fractures (21.6 Percent of the Series)

Complications

O* % O†Complications

O (%)‡O andOther§

ComplicationsO and Other

(%)��

CriteriaMet¶

Treatment No. No. % 1 2 3

ObservationWithout NFOT injury 151 0 0With NFOT injury 6 1 16.7 0 0 0 6

Reconstruction(sinus preserved)

Without NFOT injury 1 0 0With NFOT injury 8 0 0 0 0 0 7 1

OsteoneogenesisWithout NFOT injury 0 0 0With NFOT injury 2 1 50 2 100 50 1 100 1 1

ObliterationWithout NFOT injury 0 0 0With NFOT injury 13 1 7.6 12 92 8 8 13 5 7 1

CranializationWithout NFOT injury 0 0 0With NFOT injury 4 0 0 4 100 0 4 0 0 4 0

Total 185 3 1.6 18 55 11 13 15 19 13 1NFOT, nasofrontal outflow tract.*Obstruction criterion met.†Number obstructed from all with NFOT injury.‡Complications resulting from obstruction.§Patients obstructed with a second or all criteria met (fracture of the frontal sinus floor, fracture of anterior ethmoid cells).��Percentage of complications arising from this group.¶One or more of the following: obstruction of nasofrontal outflow tract, fracture of the frontal sinus floor, fracture of anterior ethmoid cells.


1854

2. Posterior wall fractures.

A. Nondisplaced posterior wall fractures(31 patients) (Table 6).

B. Displaced posterior wall fractures (28patients) (Table 7).

3. Anteroposterior wall fractures.

A. Nondisplaced anteroposterior wall frac-tures (141 patients) (Table 8).

B. Displaced anteroposterior wall frac-tures (329 patients) (Table 9).

Treatment effectiveness is summarized inTable 10.

DISCUSSIONFrontal sinus fractures represent a small por-

tion of maxillofacial fractures. Nevertheless, nu-merous classification schemes and treatment al-

Table 5. Displaced Anterior Wall Fractures (16.7 Percent of the Series)

Complications


O (%)‡O andOther§

ComplicationsO and other

(%)��

Criteria Met¶


ObservationWithout NFOT injury 13 0 0With NFOT injury 14 0 0 1 7 0 0 8 6 0


Without NFOT injury 14 1 7.1With NFOT injury 30 1 3.3 2 7 50 0 20 10 0

OsteoneogenesisWithout NFOT injury 1 0 0With NFOT injury 4 2 50 4 100 50 3 67 1 1 2

ObliterationWithout NFOT injury 7 0 0With NFOT injury 43 2 4.7 41 95 5 39 5 0 11 32

CranializationWithout NFOT injury 0 0With NFOT injury 17 1 5.9 17 100 6 17 6 8 9


Table 6. Nondisplaced Posterior Wall Fractures (3.6 Percent of the Series)

Treatment

Complications


% O‡O andOther§

ComplicationsO and Other��

CriteriaMet¶

No. No. % 1 2 3

ObservationWithout NFOT injury 14 0 0With NFOT injury 9 3 33.3 3 33 100 3 100 3 4 2

ObliterationWithout NFOT injuryWith NFOT injury 5 1 20 5 100 20 4 0 1 1 3

CranializationWithout NFOT injuryWith NFOT injury 3 0 0 3 100 0 3 0 3



1855

gorithms have been presented without statisticalsignificance. Lack of treatment uniformity con-tributes to suboptimal care and fails to followevidence-based medicine. Unlike most facialfractures, mismanagement of frontal sinus frac-tures leads to devastating and potentially fatalcomplications.5,44

Regardless of treatment, determination of na-sofrontal outflow tract obstruction remains the cor-

nerstone of management. Complications arise fromfailure of frontal sinus drainage, and recent naso-frontal outflow tract findings have contributed tofurther understanding.3,30,45–47 Until now, no onecould relate fracture patterns and degree of naso-frontal outflow tract injury to its complications. Wepresent a statistically significant treatment algorithmbased on degree of nasofrontal outflow tract injuryand its relationship to fracture patterns (Fig. 1).

Table 8. Nondisplaced Anterior and Posterior Wall Fractures (16.5 Percent of the Series)

Complications


% O‡O andOther§


(%)��

CriteriaMet¶


ObservationWithout NFOT injury 43 0 0With NFOT injury 60 3 5 5 8 60 2 100 42 18 0


Without NFOT injuryWith NFOT injury 6 1 16.7 1 16 100 0 6


CranializationWithout NFOT injuryWith NFOT injury 22 1 4.5 20 91 5 18 6 2 9 11


Table 7. Displaced Posterior Wall Fractures (3.3 Percent of the Series)

Treatment

Complications


% O‡O andOther§


(%)��

CriteriaMet¶

No. No. % 1 2 3

ObservationWithout NFOT injury 1 0 0With NFOT injury 12 0 0 2 16 0 1 0 11 1

OsteoneogenesisWithout NFOT injuryWith NFOT injury 1 0 0 1 100 0 1


CranializationWithout NFOT injury 1 0 0With NFOT injury 8 1 12.5 8 100 13 7 14 1 3 4



1856

The unique experience with high-energy fa-cial injuries at the R Adams Cowley Shock TraumaCenter allows one to evaluate options of frontalsinus fracture treatment. There was no particularsurgeon who had more complications if success ofa single treatment option was evaluated among allgroups. The treatment choice for a particular frac-ture pattern did not relate to personal surgicaljudgment or results achieved. Most patients weremanaged within 72 hours, because early treatmentreduces complications.10,16,48 Although we do notdescribe late complications because of their lowincidence, we do make inferences regarding in-dolent infections.

Plain radiographs do not adequately character-ize frontal sinus fractures and therefore have beenlargely supplanted by computed tomography.49–51

Importantly, one cannot assess nasofrontal outflowtract involvement from plain films alone. Nasofron-tal outflow tract involvement can be defined by com-puted tomography: obstruction, associated anteriorethmoid complex fracture, and frontal sinus floorfracture.28,29 Computed tomography is thereforemandatory for adequately assessing frontal sinusfracture and nasofrontal outflow tract function andfor planning surgery.

Most series report motor vehicle collisions asthe primary injury mechanism and our data con-firm this.5,7,12 The next most common mechanismsare assault and accidental injuries, confirming high-

energy trauma as the common denominator. Therarity with lower energy trauma reflects the struc-tural ability of the skull and face to absorb and dis-tribute forces along the stress-bearing framework,reducing the incidence of skull and brain damagedespite extensive facial fractures.52,53

Concomitant facial fractures were found inover 75 percent of patients. Interestingly, thosewith nasofrontal outflow tract involvement werealmost three times more likely to have other facialfractures than those without. Orbital floor, naso-orbitoethmoid complex, zygoma, and Le Fort frac-tures were most commonly associated. Similarly,intracranial injuries (epidural and subdural he-matomas) were over twice as likely with nasofrontaloutflow tract injury (76 percent). Studies have shownintracranial involvement (33 to 70 percent) withfrontal sinus fractures,5,12,16,52 confirming the rela-tionship of the nasofrontal outflow tract with thecranial base and injury severity. Rohrich and Hollierpresented a diagnostic and therapeutic algorithmestablishing a major emphasis on nasofrontal out-flow tract anatomy and drainage.2,3,45,46,54 A true ductmay be identified in only 15 percent of humans,3 andfor this reason, the term “nasofrontal outflow tract”was defined. Although the nasofrontal outflow tractanatomy was established, its relationship to injuryand complications was missing.

Some authors have argued that patients lackingnasofrontal outflow tract injury with isolated ante-

Table 9. Displaced Anterior and Posterior Wall Fractures (38.4 Percent of the Series)

Complications


% O‡O andOther§


(%)��

Criteria Met¶


ObservationWithout NFOT injuryWith NFOT injury 30 4 13.3 5 17 80 3 100 17 10 3


Without NFOT injuryWith NFOT injury 39 6 15.4 8 21 75 6 100 15 19 5

OsteoneogenesisWithout NFOT injuryWith NFOT injury 14 6 42.8 13 93 46 12 50 1 2 11

ObliterationWithout NFOT injuryWith NFOT injury 93 11 11.8 87 94 13 87 13 12 81

CranializationWithout NFOT injury 5 0 0With NFOT injury 148 14 9.5 144 97 10 137 10 7 21 120

Total 329 41 257 79 16 245 16 40 64 220NFOT, nasofrontal outflow tract.*Obstruction criterion met.†Number obstructed from all with NFOT injury.‡Complications resulting from obstruction.§Patients obstructed with a second or all criteria met (fracture of the frontal sinus floor, fracture of anterior ethmoid cells).��Percentage of complications arising from this group.¶One or more of the following: obstruction of nasofrontal outflow tract, fracture of the frontal sinus floor, fracture of anterior ethmoid cells.


1857

rior wall fractures can be safely observed becausethey rarely have complications.1–5,9,10,17,18,47,55,56 Fur-thermore, some hypothesize that complicationswere attributed to invasive operations. El Khatib etal. questioned whether the frontal sinus was a “vic-tim” or a “culprit.”56 A literature-based search an-swered the safety of observation.57 The authors con-cluded that most articles constituted level 4 evidence(poor), with an average complication rate of 9 per-cent (3 percent for those observed). However, therole of nasofrontal outflow tract involvement was notdefined. The authors ultimately concluded thatthose with more severe injuries benefit from surgerywith fewer short-term sequelae. They note that a

prospective randomized study could compare non-surgical to surgical treatment but is neither feasiblenor ethical.

The decision accuracy of our algorithm is rep-resented by the receiver operating characteristicarea under the curve of 0.8621 (Fig. 3). Receiveroperating characteristic curves are used extensivelyin medical research and frequently mentioned inconjunction with evidence-based medicine.58–61 Fur-thermore, these curves are beneficial at decisionanalysis with binomial outcomes (complication orno complication).60,62,63 The receiver operating char-acteristic area under the curve based on our decisionalgorithm of 0.8621 is quite good; it is close to theideal value of 1.0 and much larger than the worst-case value58; p values in recursive partition analysis(decision tree analysis) are weak measures of statis-tical significance.43,64 This is especially true in frontalsinus fractures because of the ratio of uncomplicatedto complicated cases. The underrepresentation ofcomplicated cases leads to Gini-square and chi-square values of 0 in two of our arms, creating anundefined p value. A p value is the probability offinding a chi-square value that is as large as or largerthan what we computed. Therefore, a receiver op-erating curve provides a better statistical metric forevaluating the most appropriate treatment algo-rithm for frontal sinus fractures.

Observation in patients who lack nasofrontaloutflow tract involvement is definitely safe asnoted by 222 patients with zero complications. Infact, there was only one complication among allthose without radiographic evidence of nasofron-tal outflow tract involvement (0.4 percent). Thispatient underwent anterior table reconstructionwith nasofrontal outflow tract stenting because ofa questionable intraoperative finding. Perhapsthis patient may have avoided a complication ifthey had been observed or not stented. Althoughsome advocate nasofrontal outflow tract recon-struction and stenting,65,66 it is plagued with failureand a high restenosis rate.3,15 One may say thatpatients with anterior table displacement who lacknasofrontal outflow tract injury still need opera-tive fixation for proper contouring. However, wefound that patients who had displaced anteriortables without nasofrontal outflow tract involve-ment were minimally displaced. Nasofrontal out-flow tract injury with obstruction is highly asso-ciated with worsening fracture type (Fig. 4).Ninety-two percent of patients with any form ofanterior displacement and 88 percent of poste-rior wall fractures (98 percent displaced) hadnasofrontal outflow tract involvement. Commi-nution played no role in our algorithm but was

Fig. 3. Receiver operating characteristic area under the curve(AUC) plots the sensitivity (true-positives) against false-positives(1 – specificity). The straight black line represents results if left to“chance alone.” The receiver operating characteristic area underthe curve indicates the diagnostic accuracy of the decision algo-rithm for frontal sinus fractures. In other words, it describes howwell the algorithm discriminates between patients with andwithout complications who underwent treatments (observa-tion, reconstruction, osteoneogenesis, cranialization, andobliteration) with regard to fracture pattern, displacement,obstruction, and nasofrontal outflow tract injury. Our receiveroperating characteristic area under the curve of 0.8621 indi-cates that the decision algorithm provides a diagnostic accu-racy of over 86 percent, which is regarded as statistically sig-nificant diagnostic accuracy.


1858

Tab

le1

0.

Fro

nta

lSin

us

Trea

tmen

tEf

fect

iven

ess

Com

plic

atio

ns

O*

%O

†C

ompl

icat

ions

%O

‡O

and

Oth

er§

Com

plic

atio

nsO

and

Oth

er(%

)��

Cri

teri

aM

et¶

Tre

atm

ent

No.

No.

%1

23

>1

(%)

Obs

erva

tion

Wit

hou

tN

FOT

inju

ry22

20

0W

ith

NFO

Tin

jury

131

118.

416

1263

910

087

386

34R

econ

stru

ctio

n(s

inus

pres

erve

d)W

ith

out

NFO

Tin

jury

151

6.7

Wit

hN

FOT

inju

ry83

89.

611

1373

610

048

305

42O

steo

neo

gen

esis

Wit

hou

tN

FOT

inju

ry1

00

Wit

hN

FOT

inju

ry21

942

.920

9545

1656

44

1381

Obl

iter

atio

nW

ith

out

NFO

Tin

jury

70

0W

ith

NFO

Tin

jury

169

15#

8.9*

*16

497

915

110

836

125

95C

ran

ializ

atio

nW

ith

out

NFO

Tin

jury

60

0W

ith

NFO

Tin

jury

202

178.

419

697

918

69

1045

147

95W

ith

out

NFO

Tin

jury

251

(29.

30%

)1

0.4

Wit

hN

FOT

inju

ry60

6(7

0.70

%)

609.

940

767

1436

815

157

153

296

74T

otal

857

617.

1N

FOT

,n

asof

ron

tal

outf

low

trac

t.*O

bstr

ucti

oncr

iter

ion

met

.†N

umbe

rob

stru

cted

from

all

wit

hN

FOT

inju

ry.

‡Com

plic

atio

ns

resu

ltin

gfr

omob

stru

ctio

n.

§Pat

ien

tsob

stru

cted

wit

ha

seco

nd

oral

lcr

iter

iam

et(f

ract

ure

ofth

efr

onta

lsi

nus

floo

r,fr

actu

reof

ante

rior

eth

moi

dce

lls).

��Per

cen

tage

ofco

mpl

icat

ion

sar

isin

gfr

omth

isgr

oup.

¶O

ne

orm

ore

ofth

efo

llow

ing:

obst

ruct

ion

ofna

sofr

onta

lout

flow

trac

t,fr

actu

reof

the

fron

tals

inus

floor

,fra

ctur

eof

ante

rior

ethm

oid

cells

.#O

blit

erat

ion

wit

hfa

t:n

ine

com

plic

atio

ns

41(2

2%).

**O

blit

erat

ion

com

plic

atio

ns

wit

hou

tfa

t:si

xof

128

(4.7

%).


1859

always associated with nasofrontal outflow tractinjury (100 percent) and highly associated withobstruction (97 percent). However, definitionof nasofrontal outflow tract injury depends notonly on fracture pattern and severity but also ondegree of injury.

Radiographic evidence of nasofrontal outflowtract injury (Fig. 2) was documented by specificcriteria (obstruction, anterior ethmoid cell frac-ture, or frontal sinus floor fracture). By evaluatingwhich criterion, combination of criteria, or num-ber of criteria were involved, a statistically signif-icant relationship with complications was found.Ninety-eight percent of complications in patients

with nasofrontal outflow tract injury had obstruc-tion as a criterion. The only complication withoutobstruction was a patient who was observed with ananterior nondisplaced fracture and resulting si-nusitis. No specific details of this patient’s hospitalcourse could explain this complication. The num-ber of criteria met did not play a role; however, 56of 59 complications with obstruction had frankobstruction with at least another criterion (95 per-cent). Overall, obstruction was found in 75 per-cent of anterior displaced fractures with nasofron-tal outflow tract involvement, in contrast to 59percent anterior nondisplaced and 78 percentposterior displaced versus 59 percent posterior

Fig. 4. Images of a frontal sinus fracture caused by a motor vehicle injury. This patient has a simultaneousanteroposterior displaced fracture seen on the axial cut (above, left) and a three-dimensional reconstruction(above, right). The fracture meets all criteria for nasofrontal outflow tract injury: fracture of the medial aspectof the anterior table, sinus floor fracture (f ), and outflow tract obstruction (o) (below, left). (Below, right)Photograph obtained intraoperatively showing comminution and obstruction around the nasofrontal out-flow tract.


1860

nondisplaced fractures. Ultimately, the finding ofobstruction in association with a frontal sinus frac-ture must persuade the provider toward oblitera-tion or cranialization.

Cranialization and obliteration have a 9 percentcomplication rate for any fracture with nasofrontaloutflow tract injury caused by obstruction (Table10). This is compared with complication rates of 63percent for observation, 73 percent for recon-struction, and 45 percent for osteoneogenesis. Ifone evaluates obstruction plus a second or thirdcriterion, then obliteration and cranialization stayconstant at complication rates of 10 percent and9 percent, whereas observation and reconstruc-tion increase to 100 percent and osteoneogenesisto 56 percent. After individual review of each an-atomical fracture pattern, cranialization and oblit-eration were comparable to the least complica-tions, whereas the other management strategieswere problematic. The only exception was a pa-tient with an anterior displaced fracture who wasobserved despite a diagnosis of obstruction but didnot have a complication. Therefore, a patient withnasofrontal outflow tract involvement caused byobstruction merits defunctionalizing to avoid se-rious sequelae. Importantly, the combination ofanother criterion in addition to obstructionshould drive the provider even further toward cra-nialization or obliteration.

Cranialization has several distinct advantages:wide exposure of the injured area allows assess-ment and repair of dural injury, access to thecranial base for complex facial fracture repair, andelimination of the sinus with its propensity forinfection and mucocele formation in a singlestage. Cranialization had formerly been reservedfor more serious injuries26,67,68 characterized bygrossly comminuted fractures with open skin andthose with fragments insufficient for reconstruct-ing the posterior wall. Some authors feel that cra-nialization is highly morbid; however, they did notmention the meticulous burring of the mucosa,careful plugging of the nasofrontal outflow tract,or reconstruction of the cranial base as prerequi-sites for success.37,44,69,70 Donald has been a majorproponent of cranialization26,44,67,68,71,72 and citesseveral advantages: immediate frontal contour res-toration, elimination of sinus and subsequent riskfor infection and mucocele, and the fact that thereis no need to rely on the variable take of a fat graftas a barrier between the central nervous systemand the sinonasal cavity.44 Its success relies onmeticulous mucosa removal and obliteration ofthe nasal communication. The frontal sinus liningis a tenacious mucoperiosteum that regeneratesfrom residual basilar mucosal cells if simple strip-ping alone is used.44 To eliminate the mucosa, thewalls must be burred to remove mucosal tails along

Fig. 5. Computed tomographic scans showing the sequelae of improper frontal sinus frac-ture management. This patient underwent management of panfacial fractures after a motorvehicle collision 4 years previously, which included obliteration of the frontal sinus with bonegraft. She presented to our clinic with months of progressive frontal pressure, headaches, andthrobbing. A mucopyocele formed after inadequate occlusion of the nasofrontal outflow tract(left). The direct sinonasal communication seen in the axial cut (right) resulted in persistent airand an indolent infection.


1861

the diploic veins of Breschet. Attention is thenturned to the anterior wall fragments and oblit-eration of the sinus. For less serious (nondisplacedor minimally displaced) anterior and posteriorwall fractures, obliteration may be used with amore limited exposure with equal success.

The choice of obliteration material is widely vari-able. Autogenous materials including fat, muscle,and bone have been used for decades.5,15,22,34,44,68–70,73–75

Alloplastic biomaterials including hydroxyapatite,bioactive glass, methylmethacrylate, calcium phos-phate bone cement, and oxidized regenerated cel-lulose have also been used.36,76–82 Supporters ofalloplastic materials note their possible antibacte-rial properties, lack of donor morbidity, and un-ending supply. Rohrich and Mickel70 and Wolfeand Johnson37 have separately addressed this con-troversy, and both advocate use of autogenous

materials. Local flaps provide a vascularizedsource to the injury site,1,83,84 and cancellous bonegrafts force the sinus to undergo ossification.85

Controversy has arisen over the use of fat for oblit-eration which, although proven safe for chronicsinusitis, is not safe or efficacious in sinuses withfractured walls.

Fat grafts have been less supported, as they relyon the viability where the damaged bony walls arepoorly vascularized. Although Montgomery notesthat fat is resistant to infection,35 our series indi-cates that fat obliteration results in high compli-cation rates. Overall, 41 fat obliterations were per-formed with nine complications (22 percent), andobliteration without fat had only six complications(5 percent). The degree of fat graft vascularizationdepends on the vascular bed; in comminuted frac-tures, there is a decreased blood supply, resulting

Fig. 6. Images of a 30-year-old patient treated 1 month previously with obliteration of the frontal sinus after sustaininga simultaneous anteroposterior fracture with nasofrontal outflow tract injury. Pus (arrows) is expressed from thesuperior orbital rim and left medial canthus (above, left). Complete removal of previous hardware and anterior table alongwith extensive burring of the sinus was performed (above, right). A vascularized free fibula osteomyocutaneous flap wasused to construct the frontal buttress and external contour, and the muscular segment obliterates the sinonasal com-munication (below, left). (Below, right) A lateral sinus view showing complete separation of the nasal cavity and frontalsinus after properly occluding the nasofrontal outflow tract.


1862

in resorption and fibrous replacement of the graft.Donald and Klotch offer experimental evidencethat fractures with significant bone loss that un-dergo sinus obliteration with fat result in up to 50percent reepithelialization with the risk of infec-tion and mucocele.26,44,67,68,86 Donald also relatedprevention of complications to the percentage ofposterior wall removed. The necessity of meticu-lous mucosal removal in obliteration techniqueshas been emphasized by Bergara and Itoiz, whodescribed successful fat grafting despite introduc-tion into septic cavities.25 Mucosal invaginationsalong the veins of Breschet constitute a source ofmucocele and abscess formation. Plugging theducts with bone seems important, as mucosa maytravel up the duct to reepithelialize the sinus aftersimple mucosal stripping.

Nadell and Kline have shown that contami-nated cranial vault and sinus fragments could bereturned to the wound if thoroughly debrided,without risk of infection.87 If the tenets of carefulmucosa removal, bone graft plugging of the na-sofrontal ducts, surgical debridement of involvedareas of the ethmoid sinus, and attention to duralintegrity are followed, cranialization is a suitable

alternative to obliteration, with a lower risk inanteroposterior displaced fractures. Althoughdata are not available to justify the use of crani-alization over other techniques, in simpler frontalsinus injuries, obliteration and cranialization havethe least morbidity when used according to ouralgorithm.

Isolated posterior wall fractures are rare, andour series does not have statistical power to de-lineate management. Some authors support ob-servation of nondisplaced posterior wall fractures,but Donald26,67,68 cites higher morbidity and ad-vocates obliteration.26,67,68 Some authors also notethat surgical management depends on posteriortable displacement of more than one tablewidth3,72 or cerebrospinal fluid involvement.8–10

Hybels and Newman evaluated 26 posterior tablefractures in a cat model and found that no com-plications arose unless the nasofrontal outflowtract was injured.86 We hypothesize that those wholack nasofrontal outflow tract involvement shouldbe observed (0 percent complications) and thosewith nasofrontal outflow tract injury should becranialized or obliterated regardless of displace-ment. We had three complications from observing

Fig. 7. Treatment algorithm for indolent infectious complications following frontal sinus fractures.


1863

those with nasofrontal outflow tract injury. How-ever, all three were nondisplaced fractures withoutflow obstruction. We can strongly concludethat one must use caution observing patients withisolated nondisplaced posterior wall fracture if ob-struction is present.

Chronic complications were not documented inour series, but some important findings must beaddressed. Although early diagnosis seems critical,patients often fail to return for follow-up before clin-ical manifestations of a missed nasofrontal outflowtract obstruction manifests. These injuries precipi-tate frontal sinusitis and mucocele formation.3,55,88

Both of these have indolent courses that create sim-ilar problems. In the face of inadequate sinus drain-age, mucous is produced continuously, resulting ul-timately in infection (Fig. 5).37,69,70

Mucoceles have an insidious course and usuallyresult from the regrowth of sinus mucosa in poorlymanaged frontal sinus fractures. Mucoceles usuallypresent years after the initial injury as an infectionwith pressure or mass-like symptoms.3,55,88 The indo-lent symptoms may begin late with visual distur-bances, nasal obstruction, and pain or may be oc-casioned by bone destruction and erosion;treatment is complicated and requires intracranial-extracranial approaches.72,89,90

We had a series of five patients who had per-sistent infection despite numerous treatments.The patients had sequelae of a slow indolent pro-cess noted by frontal bone resorption, eczematousskin changes, and cutaneous fistulas (Fig. 6). Allfive patients were managed successfully by meansof a novel strategy including a single-stage de-bridement and free vascularized tissue transfer forfinal obliteration.91 Therefore, persistent frontalsinus fracture infections that fail conventional sal-vage techniques can be treated according to oursupplementary algorithm (Fig. 7).

CONCLUSIONSWe developed a statistically significant treat-

ment algorithm for frontal sinus fractures. Pa-tients without radiographic evidence of nasofron-tal outflow tract involvement may be observed,whereas those with nasofrontal outflow tract injurywith obstruction must be treated by either oblit-eration or cranialization. There is no role for oblit-eration with fat or osteoneogenesis. Finally, a sin-gle-stage vascularized free tissue transfer holdspromise for treating persistent indolent infectiouscomplications.

Eduardo D. Rodriguez, M.D., D.D.S.R Adams Cowley Shock Trauma Center

22 South Greene StreetBaltimore, Md. [email protected]

REFERENCES1. Gerbino, G., Roccia, F., Benech, A., and Caldarelli, C. Anal-

ysis of 158 frontal sinus fractures: Current surgical manage-ment and complications. J. Craniomaxillofac. Surg. 28: 133,2000.

2. Rohrich, R. J., and Hollier, L. The role of the nasofrontalduct in frontal sinus fracture management. J. Craniomaxillo-fac. Trauma 2: 31, 1996.

3. Rohrich, R. J., and Hollier, L. H. Management of frontalsinus fractures: Changing concepts. Clin. Plast. Surg. 19: 219,1992.

4. Stanley, R. B., Jr., and Becker, T. S. Injuries of the nasofrontalorifices in frontal sinus fractures. Laryngoscope 97: 728, 1987.

5. Wallis, A., and Donald, P. J. Frontal sinus fractures: A reviewof 72 cases. Laryngoscope 98: 593, 1988.

6. Nahum, A. M. The biomechanics of maxillofacial trauma.Clin. Plast. Surg. 2: 59, 1975.

7. Nahum, A. M. The biomechanics of facial bone fracture.Laryngoscope 85: 140, 1975.

8. Yavuzer, R., Sari, A., Kelly, C. P., et al. Management of frontalsinus fractures. Plast. Reconstr. Surg. 115: 79e, 2005.

9. Bell, R. B., Dierks, E. J., Brar, P., Potter, J. K., and Potter, B. E.A protocol for the management of frontal sinus fractures em-phasizing sinus preservation. J. Oral Maxillofac. Surg. 65: 825,2007.

10. Chen, K. T., Chen, C. T., Mardini, S., Tsay, P. K., and Chen,Y. R. Frontal sinus fractures: A treatment algorithm andassessment of outcomes based on 78 clinical cases. Plast.Reconstr. Surg. 118: 457, 2006.

11. Dickenson, J. T. C. J., and Kameron, D. B. Principles offrontal sinus reconstruction. Laryngoscope 98: 593, 1988.

12. Gonty, A. A., Marciani, R. D., and Adornato, D. C. Manage-ment of frontal sinus fractures: A review of 33 cases. J. OralMaxillofac. Surg. 57: 372, 1999.

13. Ioannides, C., and Freihofer, H. P. Fractures of the frontalsinus: Classification and its implications for surgical treat-ment. Am. J. Otolaryngol. 20: 273, 1999.

14. Ioannides, C., Freihofer, H. P., Vrieus, J., and Friens, J. Frac-tures of the frontal sinus: A rationale of treatment. Br. J. Plast.Surg. 46: 208, 1993.

15. Luce, E. A. Frontal sinus fractures: Guidelines to manage-ment. Plast. Reconstr. Surg. 80: 500, 1987.

16. Shen, B. H. F. R. H., and Lin, J. T. Management of frontalsinus fractures. J. Plast. Reconstr. Surg. Assoc. R.O.C. 6: 25,1997.

17. Stanley, R. B., Jr. Management of frontal sinus fractures.Facial Plast. Surg. 5: 231, 1988.

18. Tiwari, P., Higuera, S., Thornton, J., and Hollier, L. H. Themanagement of frontal sinus fractures. J. Oral Maxillofac.Surg. 63: 1354, 2005.

19. Weber, S. C., and Cohn, A. M. Fracture of the frontal sinusin children. Arch. Otolaryngol. 103: 241, 1977.

20. Xie, C., Mehendale, N., Barrett, D., Bui, C. J., and Metzinger,S. E. 30-year retrospective review of frontal sinus fractures:The Charity Hospital experience. J. Craniomaxillofac. Trauma6: 7, 2000.

21. Schultz, R. C., and de Camara, D. L. Athletic facial injuries.J.A.M.A. 252: 3395, 1984.


1864

22. Wilson, B. C., Davidson, B., Corey, J. P., and Haydon, R. C.,III. Comparison of complications following frontal sinus frac-tures managed with exploration with or without obliterationover 10 years. Laryngoscope 98: 516, 1988.

23. Raveh, J., Laedrach, K., Vuillemin, T., and Zingg, M. Man-agement of combined frontonaso-orbital/skull base frac-tures and telecanthus in 355 cases. Arch. Otolaryngol. HeadNeck Surg. 118: 605, 1992.

24. Duvall, A. J., III, Porto, D. P., Lyons, D., and Boies, L. R., Jr.Frontal sinus fractures: Analysis of treatment results. Arch.Otolaryngol. Head Neck Surg. 113: 933, 1987.

25. Bergara, A. R., and Itoiz, A. O. Present state of the surgicaltreatment of chronic frontal sinusitis. A.M.A. Arch. Otolaryn-gol. 61: 616, 1955.

26. Donald, P. J., and Bernstein, L. Compound frontal sinusinjuries with intracranial penetration. Laryngoscope 88: 225,1978.

27. Lynch, R. C. The technique of radical frontal sinus operationwhich has given me the best results. Laryngoscope 31: 1, 1921.

28. Harris, L., Marano, G. D., and McCorkle, D. Nasofrontalduct: CT in frontal sinus trauma. Radiology 165: 195, 1987.

29. Heller, E. M., Jacobs, J. B., and Holliday, R. A. Evaluation ofthe frontonasal duct in frontal sinus fractures. Head Neck 11:46, 1989.

30. Landsberg, R., and Friedman, M. A computer-assisted ana-tomical study of the nasofrontal region. Laryngoscope 111:2125, 2001.

31. Kelly, K. J., Manson, P. N., Vander Kolk, C. A., et al. Se-quencing LeFort fracture treatment (organization of treat-ment for a panfacial fracture). J. Craniofac. Surg. 1: 168, 1990.

32. Manson, P. N., Crawley, W. A., and Hoopes, J. E. Frontalcranioplasty: Risk factors and choice of cranial vault recon-structive material. Plast. Reconstr. Surg. 77: 888, 1986.

33. Manson, P. N., Markowitz, B., Mirvis, S., Dunham, M., andYaremchuk, M. Toward CT-based facial fracture treatment.Plast. Reconstr. Surg. 85: 202, 1990.

34. Larrabee, W. F., Jr., Travis, L. W., and Tabb, H. G. Frontalsinus fractures: Their suppurative complications and surgicalmanagement. Laryngoscope 90: 1810, 1980.

35. Montgomery, W. W. Surgery of the frontal sinuses. Otolaryn-gol. Clin. North Am. 4: 97, 1971.

36. Snyderman, C. H., Scioscia, K., Carrau, R. L., and Weissman,J. L. Hydroxyapatite: An alternative method of frontal sinusobliteration. Otolaryngol. Clin. North Am. 34: 179, 2001.

37. Wolfe, S. A., and Johnson, P. Frontal sinus injuries: Primarycare and management of late complications. Plast. Reconstr.Surg. 82: 781, 1988.

38. Barone, C. M., Boschert, M. T., and Jimenez, D. F. Usefulnessof endoscopy in craniofacial trauma. J. Craniomaxillofac.Trauma 4: 36, 1998.

39. Forrest, C. R. Application of endoscope-assisted minimal-access techniques in orbitozygomatic complex, orbital floor,and frontal sinus fractures. J. Craniomaxillofac. Trauma 5: 7,1999.

40. Shumrick, K. A. Endoscopic management of frontal sinusfractures. Facial Plast. Surg. Clin. North Am. 14: 31, 2006.

41. Reidel, R., and Schenke, H. Ueber die Stimhohlen und ihreErkrankungen (inaugural dissertation). Jena, Germany,1898.

42. Spinelli, H. M., Irizarry, D., McCarthy, J. G., Cutting, C. B.,and Noz, M. E. An analysis of extradural dead space afterfronto-orbital surgery. Plast. Reconstr. Surg. 93: 1372, 1994.

43. Breiman, L. Classification and Regression Trees. Boca Raton,Fla.: Chapman & Hall/CRC, 1984.

44. Donald, P. J., and Ettin, M. The safety of frontal sinus fatobliteration when sinus walls are missing. Laryngoscope 96:190, 1986.

45. Lee, D., Brody, R., and Har-El, G. Frontal sinus outflowanatomy. Am. J. Rhinol. 11: 283, 1997.

46. McLaughlin, R. B., Jr., Rehl, R. M., and Lanza, D. C. Clinicallyrelevant frontal sinus anatomy and physiology. Otolaryngol.Clin. North Am. 34: 1, 2001.

47. Turgut, S., Ercan, I., Sayin, I., and Basak, M. The relationshipbetween frontal sinusitis and localization of the frontal sinusoutflow tract: A computer-assisted anatomical and clinicalstudy. Arch. Otolaryngol. Head Neck Surg. 131: 518, 2005.

48. Lee, T. T., Ratzker, P. A., Galarza, M., and Villanueva, P. A.Early combined management of frontal sinus and orbital andfacial fractures. J. Trauma 44: 665, 1998.

49. Kreipke, D. L., Moss, J. J., Franco, J. M., Maves, M. D., andSmith, D. J. Computed tomography and thin-section tomog-raphy in facial trauma. A.J.R. Am. J. Roentgenol. 142: 1041,1984.

50. Nahser, H. C., and Lohr, E. Possibilities of high resolutioncomputer tomography in the diagnosis of injuries of thefacial skull (in German). Radiologe 26: 412, 1986.

51. Rowe, L. D., Miller, E., and Brandt-Zawadzki, M. Computedtomography in maxillofacial trauma. Laryngoscope 91: 745,1981.

52. Olson, E. M., Wright, D. L., Hoffman, H. T., Hoyt, D. B., andTien, R. D. Frontal sinus fractures: Evaluation of CT scans in132 patients. A.J.N.R. Am. J. Neuroradiol. 13: 897, 1992.

53. Wright, D. L., Hoffman, H. T., and Hoyt, D. B. Frontal sinusfractures in the pediatric population. Laryngoscope 102: 1215,1992.

54. Kennedy, D. W., and Senior, B. A. Endoscopic sinus surgery:A review. Otolaryngol. Clin. North Am. 30: 313, 1997.

55. Manolidis, S., and Hollier, L. H. Management of frontal sinusfractures. Plast. Reconstr. Surg. 120: 32S, 2007.

56. El Khatib, K., Danino, A., and Malka, G. The frontal sinus:A culprit or a victim? A review of 40 cases. J. Craniomaxillofac.Surg. 32: 314, 2004.

57. Chuang, S. K., and Dodson, T. B. Evaluation and manage-ment of frontal sinus injuries. In R. J. Fonseca, R. Walker, N.Betts, M. P. Powers, and H. D Barber (Eds.), Oral and Max-illofacial Trauma, Vol. 2. Philadelphia: Saunders, 2004. P. 721.

58. Simon, S. STATS: Steve’s attempt to teach statistics (webpage). Available at: http://www.childrensmercy.org/stats/ask/roc.asp. Accessed February 21, 2008.

59. Lasko, T. A., Bhagwat, J. G., Zou, K. H., and Ohno-Machado,L. The use of receiver operating characteristic curves inbiomedical informatics. J. Biomed. Inform. 38: 404, 2005.

60. Obuchowski, N. A. Receiver operating characteristic curvesand their use in radiology. Radiology 229: 3, 2003.

61. Zweig, M. H., and Campbell, G. Receiver-operating charac-teristic (ROC) plots: A fundamental evaluation tool in clin-ical medicine. Clin. Chem. 39: 561, 1993.

62. Pepe, M. S. The Statistical Evaluation of Medical Tests for Clas-sification and Prediction. New York: Oxford University Press,2003.

63. Spackman, K. A. Signal detection theory: Valuable tools forevaluating inductive learning. In Proceedings of the Sixth Inter-national Workshop on Machine Learning. San Francisco: Calif.,Morgan Kaufman, 1989.

64. Metz, C. E. Basic principles of ROC analysis. Semin. Nucl. Med.8: 283, 1978.

65. Dedo, H. H., Broberg, T. G., and Murr, A. H. Frontoeth-moidectomy with Sewall-Boyden reconstruction: Alive andwell, a 25-year experience. Am. J. Rhinol. 12: 191, 1998.


1865

66. Romain, P., Legros, M., Hannion, X., Schmidt, P.,Chalumeau, F., and Bazin, A. Frontal sinus injuries: Aproposof 67 cases. Rev. Laryngol. Otol. Rhinol. (Bord.) 111: 5, 1990.

67. Donald, P. J. Frontal sinus ablation by cranialization: Reportof 21 cases. Arch. Otolaryngol. 108: 142, 1982.

68. Donald, P. J. Obliteration of compressed frontal sinus. Plast.Reconstr. Surg. 78: 832, 1986.

69. Mickel, T. J., Rohrich, R. J., and Robinson, J. B., Jr. Frontalsinus obliteration: A comparison of fat, muscle, bone, andspontaneous osteoneogenesis in the cat model. Plast. Recon-str. Surg. 95: 586, 1995.

70. Rohrich, R. J., and Mickel, T. J. Frontal sinus obliteration: Insearch of the ideal autogenous material. Plast. Reconstr. Surg.95: 580, 1995.

71. Donald, P. J. Frontobasal approach for trauma and tumor.Minim. Invasive Neurosurg. 37: 37, 1994.

72. Donald, P. J., Montgomery, W. W., and Calcaterra, T. Frontalbone defect with frontal sinus mucopyocele. Head Neck Surg.10: 59, 1987.

73. Goodale, R. L. Obliteration of the frontal sinus. Ann. Otol.Rhinol. Laryngol. 74: 831, 1965.

74. Mosher, H. A method of obliterating the naso-frontal ductand catheterizing the frontal sinus. Laryngoscope 21: 946,1911.

75. Weber, R., Draf, W., Keerl, R., et al. Osteoplastic frontal sinussurgery with fat obliteration: Technique and long-term re-sults using magnetic resonance imaging in 82 operations.Laryngoscope 110: 1037, 2000.

76. Baker, S. B., Weinzweig, J., Kirschner, R. E., and Bartlett,S. P. Applications of a new carbonated calcium phosphatebone cement: Early experience in pediatric and adult cranio-facial reconstruction. Plast. Reconstr. Surg. 109: 1789, 2002.

77. Kalavrezos, N. D., Gratz, K. W., Oechslin, C. K., and Sailer, H.F. Obliteration of the frontal sinus with lyophilized cartilage infrontal fractures. Mund. Kiefer. Gesichtschir. 2(Suppl. 1): S66,1998.

78. Kalavrezos, N. D., Gratz, K. W., Warnke, T., and Sailer, H. F.Frontal sinus fractures: Computed tomography evaluation ofsinus obliteration with lyophilized cartilage. J. Craniomaxil-lofac. Surg. 27: 20, 1999.

79. Peltola, M., Aitasalo, K., Suonpaa, J., Varpula, M., andYli-Urpo, A. Bioactive glass S53P4 in frontal sinus oblitera-tion: A long-term clinical experience. Head Neck 28: 834,2006.

80. Petruzzelli, G. J., and Stankiewicz, J. A. Frontal sinus oblit-eration with hydroxyapatite cement. Laryngoscope 112: 32,2002.

81. Verret, D. J., Ducic, Y., Oxford, L., and Smith, J. Hydroxy-apatite cement in craniofacial reconstruction. Otolaryngol.Head Neck Surg. 133: 897, 2005.

82. Weber, R., Draf, W., Kahle, G., and Kind, M. Obliteration ofthe frontal sinus: State of the art and reflections on newmaterials. Rhinology 37: 1, 1999.

83. Parhiscar, A., and Har-El, G. Frontal sinus obliteration withthe pericranial flap. Otolaryngol. Head Neck Surg. 124: 304,2001.

84. Thaller, S. R., and Donald, P. The use of pericranial flaps infrontal sinus fractures. Ann. Plast. Surg. 32: 284, 1994.

85. Grahne, B. Chronic frontal sinusitis treated by autogenousosteoplasty. Acta Otolaryngol. 72: 215, 1971.

86. Hybels, R. L., and Newman, M. H. Posterior table fracturesof the frontal sinus: I. An experimental study. Laryngoscope87: 171, 1977.

87. Nadell, J., and Kline, D. G. Primary reconstruction of de-pressed frontal skull fractures including those involving thesinus, orbit, and cribriform plate. J. Neurosurg. 41: 200, 1974.

88. Weitzel, E. K., Hollier, L. H., Calzada, G., and Manolidis, S.Single stage management of complex fronto-orbital muco-celes. J. Craniofac. Surg. 13: 739, 2002.

89. Abrahamson, I. A., Jr., Baluyot, S. T., Tew, J. M., Jr., andScioville, G., Jr. Frontal sinus mucocele. Ann. Ophthalmol. 11:173, 1979.

90. LaRossa, D. D., Noone, R. B., and Jackson, P. Facial deformityfrom frontal sinus mucocele: Single stage surgical correction.Case report. Plast. Reconstr. Surg. 60: 917, 1977.

91. Bluebond-Langner, R., Jackowe, D., and Rodriguez, E. D.Simultaneous obliteration and treatment of infected frontalsinus fractures: Novel use of the fibula flap. J. Craniofac. Surg.18: 680, 2007.


1866

26 yr experience treating frontal sinus fractures_jh paper

Documents