84 Accid Emerg Med 1997;14:84-87

Double take-fracture fishing in accident andemergency practice

Patrick Hyland-McGuire, H R Guly, P M Hughes

AbstractObjective-To investigate conditionswhere, after initially negative plain x raysfollowing trauma, there subsequentlyproves to be fracture, and to explore waysin which the management might be im-proved.Design-A 16 month prospective study.Patient details were collected from acci-dent and emergency (A&E) review clinicsand returns, A&E ward admissions, cor-respondence from other services, and dis-cussions at a weekly clinicoradiologicalconference. The inclusion criteria com-prised A&E trauma patients with normalinitial plain x rays and proven fractures onsubsequent imaging for the same patientevent.Setting--A large A&E department seeing65 000 new attendances per annum withfull back up services.Results-55 cases were identified: 41 frac-tures were identified on subsequent plainx ray, six on bone scan, six on CAT scan,and two on MRI scan. The commonestregions involved were the wrist, pelvis/hip,ankle/foot, and leg. Follow up had not beenarranged at the initial attendance in 17instances and between two and 135 dayswere required for definitive fracture rec-ognition. All but nine patients requiredalteration in treatment because offracturedetection.Conclusions-Clinical suspicion of frac-ture at initial A&E attendance shouldprompt organised follow up even in theface ofnormal plain x rays. Considerationshould be given to alternative imagingtechniques which may have a higher reso-lution than plain x rays. Close corroborationbetween A&E and radiology departmentshas benefits in patient care in this group ofpatients and may lead to a reduction infunctional disability and litigation.(7Accid Emerg Med 1997;14:84-87)

Keywords: fractures; x ray diagnosis; accident andemergency department; missed diagnosis.

Radiology plays an important role in the iden-tification of fractures in acute trauma patientsin accident and emergency (A&E) practice.However, previous reports have highlightedpotential shortcomings with plain x rays. Thedifficulty in the case of scaphoid fractures hasbeen widely publicised, and the more wide-spread use of bone scans in this condition hasbeen advocated.' Similarly, the problem in

evaluating occult femoral neck fractures is alsowell known and the value of computerised axialtomography (CT) and magnetic resonanceimaging (MRI) has been explored.2 Delayeddiagnosis may lead to increased functional dis-ability and litigation. Against this background a16-month prospective study of proven frac-tures not visible on initial plain x rays was car-ried out in a large A&E department to estimatethe incidence and spectrum of delayed fracturerecognition and to examine ways in which theoverall management might be improved.

MethodsThe inclusion criteria selected for patients inthe study group was as follows:

* A&E attendees with a recent history oftrauma.* Plain x rays taken at the first attendance withthe following stipulations:(1) correct views for the particular fracture

concerned(2) good quality films(3) reported as normal by a consultant radi-

ologist* A definite fracture seen on subsequent diag-nostic imaging:(1) for the same injury(2) no history of interim trauma(3) reported as a fracture by a consultant radi-

ologist.Senior staff in the A&E department (consult-ants and senior registrars) and radiologistsattending a weekly A&E clinicoradiologicalconference (trauma radiologist and radiologytrainees) were informed of the study andrecruited as case collectors. A list of patientsfulfilling the inclusion criteria was compiled byliaison with the principal investigator from thefollowing sources:

(1) Daily A&E review clinics staffed by seniorA&E staff seeing all booked return pa-tients;

(2) Unbooked return patients to the A&Edepartment all ofwhom are seen by seniorstaff if at all possible;

(3) Patients admitted to the A&E ward run bysenior staff;

(4) x rays brought to the weekly conference byA&E seniors and radiologists;

(5) Correspondence letters from other depart-ments on patients who had attended theA&E department;

(6) Informal feedback from colleagues fromother specialities about patients who hadattended the A&E department.

Derriford Hospital,Plymouth PL6 8DH,United Kingdom:Accident andEmergencyDepartment,P Hyland-McGuireH R Guly

Department ofDiagnostic ImagingP M Hughes

Correspondence to:Mr Patrick Hyland-McGuire,senior registrar in A&E.

Accepted for publication18 September 1996

84

on Decem

ber 18, 2021 by guest. Protected by copyright.

http://emj.bm

j.com/

J Accid E

merg M

ed: first published as 10.1136/emj.14.2.84 on 1 M

arch 1997. Dow

nloaded from

Missedfractures in A&E

Table 1 Age characteristics ofpatients in the study group

Age group Number ofpatients

<5 46-16 1017-30 1 131-50 961-65 6>66 15Range 1-94 yearsMean 41 yearsMale:female 6:5

Table 2 Delays from initialA&E presentation todefinitive fracture diagnosis

Delay in days Number ofpatients

0-7 68-14 1515-30 1731-50 12>50 5Mean 27Range 2-135

ResultsDuring the study period 67 822 new patientswere seen in the A&E department, 10 014 ofwhom were diagnosed as having a fracture atthe first attendance. A total of 57 patientsfulfilling the inclusion criteria for the study wastrawled from this caseload. Further casesundoubtedly did slip through the net but wereoutside the scope of the collection methods.From these 55 patients, case notes and x ray

Table 3 Point offracture detection

Fractures FracturesFollow up arranged found Follow up not arranged found

A&E review clinicFirst review clinic 15 GP referral back to A&E 5Second review clinic 7 Returned in pain to A&E 12Third review clinic 6Fracture clinic 1

AdmissionsA&E wards 6Orthopaedics 1Medical 2*

* One patient with follow up arranged through her GP was subsequently admitted to Care of theElderly where a fracture was found.

Table 4 Anatomical distribution offractures

24 Scaphoid

Distal radius

Lunate9 Neck of femur

Greater trochanterPubic ramusSacrumCoccyx

7 Ankle

CalcaneumFoot

7 ShaftLowerUpperLower femurDistal humerusRadial headClavicleRibLumbar spineTransverse process

15 WaistTubercleDistal poleProximal pole

8 UndisplacedIntra-articular

14 Undisplaced

Displaced12114 Lateral malleolus

Talar dome12 Metatarsal

Proximal phalanx43

l

l

1

733271

31

22

11

folders were retrieved and examined threemonths after the end date of the study andanalysed for 27 relevant variables.

Table 1 shows the age characteristics of the55 patients in the study group. The grade ofdoctor seeing the patient at first presentation tothe A&E department was a senior house officer(SHO) in 46 cases, a senior registrar in four,and a consultant in five. The injury occurredon the day of initial attendance in 34 cases, onthe previous day in nine, two days before inseven, three days before in three, and 10 daysbefore in two. The mechanism of injuryincluded falls (38), road traffic accidents (6:motorcycle 4; pedal cycle accidents 2); sportsinjuries (10); crush injuries (3); assault (1); anda bad military parachute landing (1).A comment noting clinical suspicion of frac-

ture was recorded in 38 case notes. Initialtreatment included admission in eight in-stances; application of plaster of paris in 14;support bandage in 24; analgesia alone inthree; aspiration of a joint in one; and nospecific treatment in five. Follow up arrange-ments as decided at initial consultation was asfollows: 28 were given A&E review clinicappointments and one a fracture clinic ap-pointment; one was sent for review to his fam-ily doctor; eight were admitted; three wereasked to return to the A&E department ifsymptoms persisted; and 14 were given no for-mal follow up advice.The time to definitive follow up, defined as

the delay between initial presentation to theA&E department and the time of positive diag-nosis of a fracture by repeat diagnosticimaging, ranged from two to 135 days. Thedistribution of delays is illustrated in table 2.

Table 3 shows the mechanism and locationof fracture recognition. These are divided intothose in the left hand column where follow uphad been arranged at initial attendance at theA&E department and those in the right handcolumns where no formal follow up had beenarranged. Of the 17 patients not given followup arrangements, 12 returned to the A&Edepartment with persisting pain and five werereferred back by their family doctor.The commonest symptoms prompting re-

peat imaging were persisting pain (36 patients)and restriction of joint movement (six cases).The commonest signs were persisting tender-ness (37 cases), decreased range of movement(eight cases) and swelling (six cases).The anatomical distribution of fractures is

illustrated in table 4.The imaging mode involved in definitive

fracture recognition was plain x ray in 41instances, bone scan in six cases; CT in sixcases (including one CT arthrogram), andMRI in two cases. Representative examples areshown in figs 1 and 2.The alteration in treatment resulting from

fracture detection included instigation of a newplaster of paris in 11 patients and a splint orsupport bandage in 12 others. Five patientsrequired operative intervention, comprisingtwo therapeutic arthroscopies, two dynamichip screws, and one hemiarthroplasty of the hip. Afurther three patients required admission for bed

Wrist

Pelvis/hip

Anklelfoot

Tibia

Miscellaneous

85

on Decem

ber 18, 2021 by guest. Protected by copyright.

http://emj.bm

j.com/

J Accid E

merg M

ed: first published as 10.1136/emj.14.2.84 on 1 M

arch 1997. Dow

nloaded from

Hyland-McGuire, Guly, Hughes

Plain x ray (left) showing healing fracture of the left tibia taken three months after a crushinjury between a boat and a pontoon in a 37 year old male. Initial x ray (right) wasnormal and the patient was thought not to have a fracture. The pati'ent was referred back tothe A&E department for x ray by his general practitioner with continuing tenderness and apalpable mass. The injury was initially treated with support bandage. Outcome was goodwith six weeks ofplaster ofparis treatment following diagnosis offracture.

................................................................

Bone scan showing high uptake across the pertrochanteric line in keeping with apertrochanteric fracture in a 94 year oldfemale performed five days after a fall. The patienthad initial normal x ray and was admitted to the A&E wardfor mobilisation. Mobilisationwas unduly slow so bone scan was done. The patient had a repeat x ray at 10 days andsubsequently had a dynamic hip screw for a displaced basicercical fracture. Outcome wasgood.

rest. Only nine patients had no alteration in treat-ment, the remaining patients having physiotherapyor having their existing plasters or supportbandage treatnent prolonged.

DiscussionPlain x rays currently form the mainstay ofdiagnostic imaging used in A&E practice.3However, initial plain radiography has its limi-tations. Delays in fracture diagnosis may leadto increased functional disability.4 Missedorthopaedic injuries remain the leading causeof malpractice claims in emergency medicine.5

In this study we examined the incidence (withlimitations caused by the logistics of collectingmethods) and spectrum of fractures not visibleon initial adequate plain x rays in a repres-entative A&E practice. The true incidence isundoubtedly greater than that shown.Nevertheless, the problem is highlighted andsome of the strengths and failings of the systemin dealing with the problem are shown.The value of a readily accessible system for

planned follow up ofA&E patients in the formof A&E review clinics (28 fractures found) isillustrated. Clinical responsibility for patientswith a differential diagnosis of soft tissue injuryor fracture traditionally rests with the A&Edepartment until referral can be made to theorthopaedic service when a fracture has beenclearly demonstrated. Review clinics provide auseful method of dealing with this importantgroup of conditions. Education of SHOs (46patients seen at first attendance) at the start oftheir A&E posts on the appropriate referral ofclinically suspicious categories of injuries tothese clinics will reduce the number of patients(17) not sent for follow up. Follow up clinicsshould optimally be manned by senior staff, asfindings on repeat x rays in the study wereoften subtle and decisions regarding the appro-priateness of alternative and often costly imag-ing may have to be made. In the study, defini-tive fracture diagnosis took anything betweentwo and 135 days. This level of service is possi-ble in larger departments and provides a usefulforum for audit and feedback for A&E staff,but in smaller units a definite decision mayhave to be made at an early stage to refer on tofracture clinics.

Patients returning to the A&E departmentwith persisting symptoms (17 in the study) arebest seen by senior staff who can plan furthermanagement at this opportune time anddecide on the suitability of alternative imaging.If possible, patients should be encouraged toreturn to A&E departments during normalworking hours when senior staff are moreavailable and when full imaging facilities areaccessible.The benefit of an A&E ward run by senior

A&E doctors for admission of patients for fur-ther observation and investigation is shown (sixfractures diagnosed). This service proved espe-cially useful for the difficult group of x raynegative hip/pelvic injuries in the elderly, whoare unable to mobilise-a group difficult toplace with other services. Those who could notbe mobilised after a period of observation hadother imaging, and six fractures were found.The benefit to patient care of radiographic

audit has been shown before.6 Most of the 55cases were discussed at the weekly clinicoradio-logical conference which provided a usefulforum for discussion and definitive opinion onx rays, together with advice on the appropriate-ness of further imaging. Informal audit of bothdepartments was implicit in this arrangement.

Certain anatomical regions yielded the high-est numbers of fractures. Careful follow up isadvised in injuries to the wrist, pelvis/hip,ankle/foot, and the leg. Though the problem ofoccult fractures of the scaphoid is well known,

86

on Decem

ber 18, 2021 by guest. Protected by copyright.

http://emj.bm

j.com/

J Accid E

merg M

ed: first published as 10.1136/emj.14.2.84 on 1 M

arch 1997. Dow

nloaded from

Missedfractures inA&E 87

the study highlighted distal radial fractures(eight in the study), particularly undisplacedfractures, as an area for particular vigilance.Two toddler's7 fractures were identified andpatients in this group should be followed up byrepeat x ray at two weeks if still not weightbearing. Two troublesome ankle injuries wereresolved by therapeutic arthroscopies followingMRI demonstration of osteochondral talardome fractures.Fourteen fractures were picked up on imag-

ing modes other than plain x ray. The superior-ity of other imaging techniques has previouslybeen shown in the case of bone scanning forscaphoid fracture,8 and a role for magneticresonance imaging9 and bone scanning'0 hasbeen advocated in the A&E setting. Thesetechniques are becoming increasingly availablein A&E. In our study we found bone scan use-ful in scaphoid fractures and fractured neck offemur: CT in neck of femur and scaphoid frac-tures, and MRI in ankle fractures.

CONCLUSIONSNegative initial plain x rays are not always con-clusive evidence of the absence of a fracture inA&E trauma patients. Persistent organisedreview by experienced A&E staff with directinvolvement of the radiology department in theface of continuing clinical suspicion will helpto detect these occult fractures. Other imaging

techniques should be considered in consulta-tion with the radiologists. The responsibilityfor patients in this group rests with the A&Edepartment until a fracture is positively identi-fied. Review clinics run by A&E senior staff, afollow up protocol for junior A&E staff, andclinicoradiological meetings and audit will helpto reduce disability in these patients andreduce the likelihood of litigation.

1 Tiel-van Buul MM, van Beek EJ, Broekhuizen AH,Nooitgedacht EA, Davids PH, Baker AJ. Diagnosingscaphoid fractures: radiographs cannot be used as a goldstandard. Injury 1992;23:77-9.

2 Evans PD, Wilson C, Lyons K. Comparison of MRI withbone scanning for suspected hip fracture in elderly patients.J Bone Joint Surg Br 1994;76:158-9.

3 Laasonen EM, Kivioja A. Delayed diagnosis of extremityinjuries in patients with multiple injuries. J Trauma1991;31:257-60.

4 Born CT, Ross SE, Iannacone WM, Schwab CW, De LongWG. Delayed identification of skeletal injury in multisys-tem trauma: the "missed" fracture. J Trauma 1989;29:1643-6.

5 Moore MN. Orthopaedic pitfalls in emergency medicine.South Med J 1988;81:371-8.

6 Thomas HG, Mason AC, Smith RM, Ferguson CM Valueof radiograph audit in an accident and emergency depart-ment. Injury 1992;23:47-50.

7 Shravat BP, Harrop SN, Kane TP. Toddler's fracture. JAccid Emerg Med 1996; 13:59-61.

8 Murphy D, Eisenhauer M. The utility of a bone scan in thediagnosis of clinical scaphoid fracture. J Emerg Med 1994;12:709-12.

9 Feldman F, Staron R, Zwass A, Rubin S, Haramati N. MRimaging; its role in detecting occult fractures. Skel Radiol1994;23:439-44.

10 Heinrich SD, Gallagher D, Harris M, Nadell JM. Undiag-nosed fractures in severely injured children and youngadults. J Bone Joint Surg Am 1994;76:561-74.

Referees for the Journal ofAccident & Emergency Medicine

All papers submitted for publication in the Journal of Accident &Emergency Medicine undergo peer review. As a result of the continuing risein the number of papers received the Journal seeks additional referees.

This is an interesting and stimulating activity. The Editorial Office ensuresthat the workload for referees is not onerous and guidelines are provided toallow a structured critique of each paper. Referees are expected to returncomments within three weeks of receipt of the manuscript.

Please contact the Editor, Journal of Accident & Emergency Medicine atBMA House, Tavistock Square, London WC 1H 9JR, telephone 0171-383-6795, fax. 0171-383-6668, stating your present appointment and any areasof special expertise. Reviewers are particularly welcome from other special-ties with an interest in Emergency Medicine and from outside the U.K.

on Decem

ber 18, 2021 by guest. Protected by copyright.

http://emj.bm

j.com/

J Accid E

merg M

ed: first published as 10.1136/emj.14.2.84 on 1 M

arch 1997. Dow

nloaded from