Comparison of Ultrasonography and Conventional Radiography in the Diagnosis of Nasal Fractures

ORIGINAL ARTICLE Comparison of Ultrasonography and Conventional Radiography in the Diagnosis of Nasal Fractures

Oliver Thiede, MD; Jan-Hendrik Krömer, MD; Claudia Rudack, MD; Wolfgang Stoll, MD; Nani Osada, MD; Frank Schmäl, MD

Objective: To compare the diagnostic findings of ul- The results were analyzed by various statistical testing trasonography and radiography in nasal fractures. methods (for sensitivity, specificity, positive and negative predictive value, and accuracy). Design and Main Outcome Measures: In this prospective study, 63 patients (23 female and 40 male; mean Results: Assessment of the lateral nasal walls revealed age, 26.8 years) with clinical signs of a nasal bone frac- that ultrasonography was statistically superior (P=.04) ture were investigated. All patients underwent radiogra- to radiography. In contrast, assessment of the nasal dor- phy (lateral view of the nose plus occipitomental view) sum showed radiography to be statistically superior and ultrasonography (10-MHz ultrasound scanner) of the (P=.01) to ultrasonography. Assessment of the nasal pyra- nasal dorsum and the lateral nasal walls and a clinical mid revealed no statistical difference between radiogra- examination by 2 consultants. Thirty-six patients under- phy and ultrasonography (P=.91). went nasal fracture reduction. Two radiographs and 3 ultrasound images of each patient were analyzed by 2 Conclusion: In assessment of the nasal pyramid, ultra- experienced readers at different times. After assessing the sonography can be considered an alternative to radiog- nasal dorsum and lateral nasal walls in radiographs and raphy, with equivalent diagnostic performance. ultrasound images, they decided whether the nose was fractured or not or whether the results were uncertain. Arch Otolaryngol Head Neck Surg. 2005;131:434-439

ASAL FRACTURES ARE THE In addition to the clinical examina- most commonly ob- tion (crepitation, deviation from the mid- served fractures in the line, and dislocated fracture), the nasal trauma setting.1 A uni- bone fracture is often diagnosed by radi- form classification of these ography. The radiographs usually com- fractures has not been established, but there prise a lateral image of the nasal bone fo- N 2 have been some attempts to classify them. cused on the nasal dorsum and an The nasal pyramid is a complex structure occipitomental radiograph from which the consisting of the 2 nasal bones and the 2 lateral nasal walls can be assessed. The frontal processus of the maxillary bone. A need for radiography has been ques- nasal fracture can involve any part of the na- tioned by some authors,4 who find it an sal pyramid. The lateral nasal walls (cau- inappropriate means of assessing a nasal dal part of the nasal bones and parts of the fracture. However, adequate imaging of a frontal processus) and the nasal dorsum nasal fracture is often required because of (cranial part of the nasal bones), as well as legal consequences resulting from nasal the nasal septum, generally require the most fracture etiology. attention when assessing a fracture of the An alternative to radiography is ultra- nasal pyramid. sonography, a common and easy method Author Affiliations: If there are no further injuries, such as involving no additional radiation expo- Department of hematomas or expanded fractures of the na- sure. The value of ultrasonography as a di- Otorhinolaryngology, Head and sal septum, within 7 days after trauma the agnostic tool for the detection of “bone” Neck Surgery, University patient undergoes reduction of the nasal fractures (eg, in congenital hip disloca- Hospital (Drs Thiede, Krömer, fracture under local anesthesia or, in cases tion, fractures of the radial or clavicular Rudack, Stoll, and Schmäl), and of noncompliance, under general anes- bone, or fractures of the ribs or skull) has Department of Medical thesia.3 Afterward, the nose is fixed with been proven in various studies.5-8 Informatics and 9-12 Biomathematics, University a plaster cast. Some authors recommend To our knowledge, only 4 studies of Münster (Dr Osada), nasal packs to stabilize the bone frag- have dealt with the diagnosis of nasal frac- Münster, Germany. ments, especially in the case of a de- tures by ultrasound imaging. Danter et al10 Financial Disclosure: None. pressed fracture.3 used a special 20-MHz ultrasound head of

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 434

Figure 1. Defined positions of the ultrasound head. A, Right lateral wall. B, Left lateral wall. C, Nasal dorsum.

an ultrasound device to investigate nasal fractures in 18 DIAGNOSIS AND TREATMENT patients. The ultrasonographic results were compared with the clinical diagnosis (sensitivity, 83%; and specificity, Following clinical examination, a first-line imaging set of con- 50%) and with radiography findings (sensitivity, 94%; ventional radiographs (lateral view and occipitomental [or and specificity, 83%). The authors did not distinguish be- Waters] view of the nasal bones) was taken in every subject. tween the nasal dorsum and the lateral nasal walls. Fur- The radiologic management of patients was performed accord- ing to the guidelines of the German Society of Otorhinolaryn- thermore, a comparison of the radiographic findings with 11 gology, Head and Neck Surgery. The lateral view of the nose the clinical diagnosis was not performed. Kwon et al delineates fractures of the nasal dorsum, where the 2 nasal performed ultrasonography on 45 patients who were sus- bones are connected. The Waters view is used to detect frac- pected to have a nasal fracture. Ten of those patients also tures of the lateral nasal walls. Furthermore, an ultrasound underwent a computed tomographic scan. The authors examination in defined positions (Figure 1) was performed found that the ultrasound diagnosis showed a positive by 2 experienced otorhinolaryngologists (O.T. and J.-H.K.) correlation with the findings on computed tomographic using a 10-MHz ultrasound head (EUP-L34T, 3ϫ1.4 cm; scanning. Beck et al9 examined 21 patients using 5-MHz Hitachi Medical Corp, Tokyo, Japan) of a commonly used and 7.5-MHz ultrasound heads. These authors demon- ultrasound device (EUB-525RS, Hitachi Medical Corp). On strated that all fractures diagnosed by radiography also the basis of these diagnostic findings (clinical examination, radiography, and ultrasonography), a decision was made on could be detected by ultrasonography. However, they did whether there was a nasal fracture, and the treatment was not compare the ultrasonographic and radiographic re- then determined. sults with the clinical diagnosis. In 36 patients, repositioning of the fractured nasal bone was The aim of the present single-blinded study (2 inde- performed under local or general anesthesia. In most cases, na- pendent readers [F.S. and C.R.]) was to analyze the cor- sal reduction was performed by digital pressure and an eleva- relation of ultrasonography and radiography results with tor. The surgeons were asked to document all details about the the clinical and the intraoperative findings in a repre- nasal repositioning, especially the fracture location, occur- sentative number of patients. Furthermore, we ana- rence of crepitation, and type of stabilization (plaster cast or lyzed whether the 2 diagnostic methods showed differ- other). Furthermore, they were to document whether the frac- ent results based on different fracture locations. ture had been stably repositioned. After data collection for all 63 patients, 2 otorhinolaryngologists (F.S. and C.R.) who had more than 10 years’ clinical METHODS experience analyzed the radiographs (lateral view and occipitomental view of the nasal bones) and ultrasound images (right and left lateral nasal walls and nasal dorsum) at different times STUDY POPULATION without being aware of the patients’ history (single-blinded study). The sequence in which the diagnostic methods (ultra- Sixty-three patients (23 female and 40 male; mean age, 26.8 sonography and radiography) were analyzed was randomized. years) with clinical signs of a nasal fracture were enrolled in For each diagnostic method and for each fracture location (na- this prospective study. The study was carried out between Janu- sal dorsum and lateral nasal walls), the reader noted whether ary 2003 and November 2003 in accord with ethical standards there was a fracture, no fracture, or an uncertain result. The established in an appropriate version of the 1964 Declaration reader then stated whether the nasal pyramid was or was not of Helsinki. broken or whether the result was uncertain.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 435

Figure 2. Lateral view of the nose (A) and ultrasound image (B) in a patient with a fractured nasal dorsum. Arrow indicates fracture.

The ultrasonographic and radiographic results were com- be improved by reduction. Reduction of the nasal frac- pared with the diagnosis made in the outpatient clinic. The di- tures was performed under local anesthesia in 26 pa- agnosis was based on results from the clinical examination, ra- tients. In 4 patients, the nose deviated so much or these diographs, ultrasound images, and, in cases in which reduction 4 patients were so noncompliant that reduction was pos- of the nasal bone was performed, intraoperative findings. The findings were rated as “correct” if the radiography or the ul- sible only under general anesthesia. Because of a frac- trasonography result matched the clinical diagnosis. If they failed ture of the nasal septum in 5 patients, the nasal septum to match, they were rated as “incorrect.” When assessment of and the nasal fracture also had to be reconstructed. One the nose by radiography or ultrasonography was uncertain, the of these patients also had a septum hematoma. Reduc- cases were rated as “uncertain,” irrespective of whether the nose tion of the lateral nasal walls was possible in all but 1 pa- was broken. In addition, the radiographic and ultrasono- tient. In 3 patients, the nasal dorsum could not be re- graphic assessments of the 2 readers were compared. duced, but the lateral nasal wall was depressed and could be reduced. In 26 patients, reduction of the nasal frac- STATISTICAL ANALYSIS tures was stable. In 9 patients, reduction of the nasal fractures was not stable, so the nostrils were packed with ab- The statistical program used was Statistical Product and Service sorbent cotton for 3 days to stabilize the nasal pyramid. ␹2 Solutions 11.0 for Windows (SPSS Inc, Chicago, Ill). The test In all 36 patients, the nose was fixed with a plaster cast. and the ␬ test were used (namely, to compare the results between the 2 diagnostic tools [radiography and ultrasonography] Two radiographs and 3 ultrasound images of each of and between the 2 readers). The results of the ␬ test were inter- the 63 patients were analyzed by 2 different readers (F.S. preted as follows: ␬ values ranging from 0.21 to 0.40 indicated and C.R.). Figure 2 shows an example of a fractured fair correspondence; 0.41 to 0.60, moderate correspondence; 0.61 nasal dorsum, and Figure 3 shows an example of a frac- to 0.80, good correspondence; and 0.81 to 1.00, very good cor- tured lateral nasal wall. As no significant differences were respondence. PϽ.05 was considered a significant difference. Fur- found between the results of the 2 readers and because thermore, the sensitivity, specificity, positive and negative pre- the ␬ test showed a good correspondence in most com- dictive value, and accuracy were calculated. parisons between the readers’ findings (Table 1), the data of both readers were pooled for further analysis. RESULTS The descriptive statistical analysis (sensitivity, specificity, positive and negative predictive value, and accu- A nasal fracture was diagnosed in 42 (67%) of 63 pa- racy) of the comparison between the clinical and surgi- tients. In all 42 patients, a fracture of at least 1 lateral na- cal diagnosis and the radiographic and ultrasonographic sal wall was found. In 40 patients, a fracture line of the assessment of the nasal dorsum showed comparable or nasal dorsum was diagnosed. In 19 patients, the broken higher values for the radiographic examination (Table 2). nose was caused by punches, and the second leading cause In contrast, the comparison of the lateral nasal walls re- was falls, as seen in 15 patients. Thirty-six (86%) of 42 vealed that ultrasonography was superior to radiogra- patients with a broken nose underwent reduction of the phy. In the final comparison of the nasal pyramid, for nasal fracture. Six patients refused reduction, or the frac- most values, the ultrasonographic sensitivity, specific- ture lines were only minimally dislocated, so that the pa- ity, positive and negative predictive value, and accuracy tients were aware of virtually no nasal deformity that could were superior to the radiographic results.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 436

Figure 3. Occipitomental view of the nose (A) and ultrasound image (B) in a patient with a fractured lateral nasal wall. Arrow indicates fracture.

Table 1. Comparison of Findings of the 2 Readers

Reader 1 Reader 2

Localization Broken Not Broken Uncertain Broken Not Broken Uncertain ␹2 Value* ␬ Value Nasal dorsum Lateral nasal view 36 17 10 32 21 10 0.7 0.89 Ultrasound 23 23 17 17 33 13 4.9 0.76 Lateral nasal wall Occipitomental view 30 18 15 25 17 21 1.5 0.73 Ultrasound 36 21 6 28 26 9 4.3 0.71 Nasal pyramid Radiographs 46 9 8 38 16 9 2.8 0.65 Ultrasound 43 12 8 32 26 5 5.6 0.55

*df=2 for all. All differences were not significant.

The ␹2 test showed radiography to be significantly su- tients. For this reason, and because the 10-MHz ultra- perior to ultrasonography in assessment of the nasal dor- sound head is the most commonly used ultrasound de- sum, while ultrasonography was significantly superior vice in the examination of the head and neck in daily to radiography in assessment of the lateral nasal walls. routine, the 10-MHz head was used in the present study. In the final assessment of the nasal pyramid, there was The size of the ultrasound head used in this study was no significant difference between the findings of the ul- 3ϫ1.4 cm. This proved to be a disadvantage in the ex- trasonographic examination and the radiographic examination of children with a small nose, whereas it en- amination (Table 3). abled the nasal dorsum and the lateral nasal wall to be visualized in one image in adult patients. This was ad- COMMENT vantageous in the retrospective analysis of nasal trauma and especially in cases with potential legal conse- In contrast to the study by Danter et al,10 who used a quences, in which good documentation of the fracture special 20-MHz ultrasound head and a 7.5-MHz ultra- is important. If the nasal dorsum is swollen or if trauma sound head in 8 patients with nasal fractures, all patients in has led to an irregular surface of the nose, a water-filled the present study were examined with a 10-MHz ultra- rubber glove may be helpful in optimizing performance sound head. Although the uncommon 20-MHz ultra- of the ultrasound examination. However, this was not nec- sound head has a better surface resolution, Danter at al10 essary in any of the present patients. demonstrated that all nasal fractures that were detect- Ultrasonography is a dynamic procedure that allows able with a 20-MHz ultrasound head could also be the reader to make the diagnosis from an unlimited num- detected with a 7.5-MHz ultrasound head in all 8 pa- ber of pictures. This made it much easier for the 2 in-

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 437

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 2. Comparison of Radiography and Ultrasonography vs Clinical Diagnosis and Intraoperative Findings Among All 63 Patients

Clinical and Surgical Diagnosis Positive Negative Opinions of the Sensitivity, Specificity, Predictive Predictive Accuracy, 2 Readers Broken Not Broken Total % % Value, % Value, % % Radiographic lateral nasal view 82 69 90 87 75 Broken 61 7 68 Not broken 5 33 38 Uncertain 8 12 20 Total 74 52 126 Ultrasound image of nasal dorsum 49 69 90 55 57 Broken 36 4 40 Not broken 20 36 56 Uncertain 18 12 30 Total 74 52 126 Radiographic occipitomental view 52 43 78 54 49 Broken 43 12 55 Not broken 16 19 35 Uncertain 23 13 36 Total 82 44 126 Ultrasound image of lateral nasal wall 70 70 89 66 70 Broken 57 7 64 Not broken 16 31 47 Uncertain 9 6 15 Total 82 44 126 Ultrasound image of nasal pyramid 77 60 87 66 71 Broken 65 10 75 Not broken 13 25 38 Uncertain 6 7 13 Total 84 42 126 Radiographic view of nasal pyramid 80 38 80 64 66 Broken 67 17 84 Not broken 9 16 25 Uncertain 8 9 17 Total 84 42 126

Table 3. Comparison of Radiography and Ultrasonography Findings

Radiography Ultrasonography ␹ 2 Localization False Correct Uncertain False Correct Uncertain Value* P Value Nasal dorsum 12 94 20 24 72 30 8.9 .01 Lateral nasal wall 32 55 39 37 67 22 6.3 .04 Nasal pyramid 26 83 17 23 90 13 0.2 .91

*df=2 for all. All differences were not significant.

vestigators performing the ultrasound examination to de- phy reached a sensitivity of 77% and a specificity of 57%. cide whether there was a nasal fracture than for the 2 These figures are lower than those reported by Kwon et readers whose decisions were based on 3 ultrasound im- al,11 who located the nasal fracture in all patients using ages only. This suggests that results of an ultrasound ex- ultrasonography. amination are better when the procedure is performed In the study by Danter et al,10 assessment of the nasal by the same person reading the results. fracture yielded a sensitivity of 83% if the clinical diag- In this study, assessment of the nasal dorsum by ra- nosis was used as the reference and a sensitivity of 94% diography reached a sensitivity of 82% and a specificity if the radiography results were considered. These data of 69%. For the lateral nasal walls, a sensitivity of 52% are similar to the results reported in this study. and a specificity of 49% were achieved. The final assess- The present study revealed a significant advantage of ment of the nasal pyramid by radiography reached a sen- radiography with respect to the nasal dorsum. In par- sitivity of 80% and a specificity of 38%. These data par- ticular, a slightly dislocated fracture of the nasal bone allel results reported in the literature.13 was difficult to assess by ultrasonography. These results Assessment based on ultrasound images yielded a sen- are in contrast to those of Kwon et al,11 who found that sitivity of 49% and a specificity of 69% with respect to radiographic assessment of the nasal bone yielded a the nasal bone. The analysis of the lateral nasal walls sensitivity of only 60% to 70%, whereas analysis of the showed a sensitivity of 70% and a specificity of 70%. The nasal bone by ultrasonography yielded a sensitivity of final assessment of the nasal pyramid by ultrasonogra- 100%.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 438

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 By contrast, in the present study, assessment of the rhinolaryngology, Head and Neck Surgery, University lateral nasal wall was significantly better when based on Hospital, Kardinal von Galen Ring 10, D-48129 Mün- ultrasound examination than on the occipitomental ra- ster, Germany ([email protected]). diograph. Specifically, the swelling of soft tissue led to difficulties in analyzing the radiograph, while this prob- lem did not occur in the ultrasound examination except REFERENCES in rare cases of open bone fracture. In the final assess- 1. Alvi A, Doherty T, Lewen G. Facial fractures and concomitant injuries in trauma ment of the nasal pyramid, no significant statistical dif- patients. Laryngoscope. 2003;113:102-106. ferences were found. 2. Murray JA, Maran AG, Busuttil A, Vaughan G. A pathological classification of nasal fractures. Injury. 1986;17:338-344. 3. Green KM. Reduction of nasal fractures under local anaesthetic. Rhinology. 2001; CONCLUSIONS 39:43-46. 4. Nigam A, Goni A, Benjamin A, Dasgupta AR. The value of radiographs in the man- The examination of nasal bone fractures by ultrasonog- agement of the fractured nose. Arch Emerg Med. 1993;10:293-297. raphy is a common method involving no radiation ex- 5. Klinger M, Danter J, Siegert R. Sonographie von Orbitabodenfrakturen: eine Al- ternative zur Computertomographie? Laryngorhinootologie. 1996;75:242-246. posure. Compared with the radiographic occipitomen- 6. Katz R, Landman J, Dulitzky F, Bar-Ziv J. Fracture of the clavicle in the newborn: tal view, ultrasonography is a more adequate diagnostic an ultrasound diagnosis. J Ultrasound Med. 1988;7:21-23. tool for assessment of lateral nasal wall fractures. In con- 7. Jank S, Strobl H, Emshoff R, et al. Bildgebende Diagnostik der medialen und lat- trast, nasal dorsum fractures were represented much more eralen Orbitawandfraktur: Sonographie versus Computertomographie. Mund Kiefer Gesichtschir. 2003;7:208-213. clearly in the lateral radiographic view of the nose. In the 8. Poussaint TY, Moeller KK. Imaging of pediatric head trauma. Neuroimaging Clin final assessment of the nasal pyramid, radiography and NAm. 2002;12:271-294, ix. ultrasonography were considered equivalent methods, al- 9. Beck A, Maurer J, Mann W. Sonographische diagnose von nasenbeinfrakturen: though ultrasonography showed slight advantages in the otorhinolaryngologie. In: Verhandlungsbericht der Deutschen Gesellschaft für Hals- descriptive statistical analysis. Therefore, ultrasonogra- Nasen-Ohrenheilkunde, Kopf Hals-Chirurgie. Stuttgart, Germany: Thieme- Verlag; 1992:68. phy can be considered an alternative to radiography, with 10. Danter J, Klinger M, Siegert R, Weerda H. Ultrasound imaging of nasal bone equivalent diagnostic performance in assessing nasal pyra- fractures with a 20-MHz ultrasound scanner [in German]. HNO. 1996;44:324- mid fractures. 328. 11. Kwon TK, Cha JH, Kim YW, et al, eds. The Role of Ultrasound in the Diagnosis of Nasal Bone Fracture. New York, NY: Kugler Publications Amsterdam; 1995. Submitted for Publication: September 2, 2004; final revi- 12. Friedrich RE, Heiland M, Bartel-Friedrich S. Potentials of ultrasound in the diag- sion received January 19, 2005; accepted February 2, 2005. nosis of midfacial fractures. Clin Oral Investig. 2003;7:226-229. Correspondence: Oliver Thiede, MD, Department of Oto- 13. Illum P. Legal aspects in nasal fractures. Rhinology. 1991;29:263-266.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 131, MAY 2005 WWW.ARCHOTO.COM 439