Injury, Int. J. Care Injured (2007) 38, 931—936

www.elsevier.com/locate/injury

Paediatric femoral fractures: Factors influencinglength of stay and readmission rate

Nichola C. Wilson a, N. Susan Stott b,*

aDepartment of Orthopaedics, Palmerston North Hospital, Palmerston North, New ZealandbDepartment of Paediatric Orthopaedics, Starship Children’s Hospital, Auckland, New Zealand

Accepted 15 March 2007

KEYWORDSFemur;Fracture;Child;Length of stay;Hip spica;Operation

Summary

Aim: This report describes factors influencing length of stay and unplanned read-missions within 30 days of discharge for children presenting with femoral fractures toStarship Hospital.Methods: The charts and radiographs of all children admitted with femoral shaftfractures between January 1998 and December 2002 were reviewed, excludingpathological fractures. The type of management, the length of stay and the unex-pected readmission rate within 30 days were recorded. Ninety-three children with 95femoral fractures were included in the study.Results: Forty-six fractures were treated by application of immediate hip spica and49 by other surgical methods (21 by external fixation, 20 by flexible intramedullarynailing and 8 by other methods). Patients with isolated femur fractures treated byeither hip spica application or other operative fixation had median lengths of stay of 3days (range 1—10 days) and 6 days (range 2—15 days), respectively. Patients withassociated injuries, whose femur fractures were treated by either hip spica applica-tion or other operative fixation, had longer median lengths of stay of 5 days (3—15days) and 17 days (5—70 days), respectively. Nine of the 93 children had an unplannedreturn to the operating theatre during their initial hospital stay, with increased lengthof stay in 5 as a result. A further 21 children were readmitted within 30 days ofdischarge with problems of loss of fracture alignment and cast damage in the hip spicagroup; pin-site infections in the external fixation group and pain around the distal nailinsertion site in the flexible intramedullary nailing group. Despite these earlyproblems, fracture alignment was satisfactory at union in 89 of the 95 fractures,with satisfactory remodelling in the remaining six fractures at follow-up.Conclusions: The presence of associated injuries was a key factor in the length of theinitial hospital stay. However, the need for further operative interventions in the first

* Corresponding author. Tel.: +64 9 373 7599x82861.E-mail address: s.stott@auckland.ac.nz (N.S. Stott).

0020–1383/$ — see front matter # 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.injury.2007.03.013

932 N.C. Wilson, N.S. Stott

hospital stay also increased the length of stay in some patients. Overall, the operativemanagement of femoral fractures is resource intensive with a significant need forfurther minor operative intervention, regardless of the type of initial management.# 2007 Elsevier Ltd. All rights reserved.

Femoral shaft fractures represent approximately1.6% of fractures in the paediatric population andfollow a bimodal distribution, with peak incidencesat 2 and 12 years.2 Skin or skeletal traction inhospital has been the traditional method of man-agement for femoral fractures in children but isalso the most expensive, due to the length ofhospital stay.8 As a result, the management offemoral fractures in children has changed overthe last 15 years, with an emphasis on interven-tions that facilitate cost savings through earlierdischarge from hospital.4,6 These interventionsinclude early hip spica cast immobilization inyounger patients, and a variety of operative inter-ventions in older patients, including external fixa-tion, flexible intramedullary nailing or femoralplating. The choice of which operative treatmentis still controversial (particularly in the intermedi-ate age group of 6—10 years) but primary influen-cing factors include age, fracture configuration,associated injuries, family issues, local experienceand cost.1,3,7,17

The clinical and radiographic outcomes ofoperative femoral fracture management are gen-erally good.22 However, there is less informationregarding the impact of operative fracture man-agement on length of stay and early readmissionrate in a public hospital setting, where access tothe operating theatre is often constrained. Star-ship Children’s Health is a tertiary paediatric cen-ter that admits over 1500 children a year withacute fractures. The purpose of this study is tolook at factors that influence length of stay andreadmission rate following femoral fracture in ourclinical environment.

Patients and methods

The Research and Development Unit at AucklandHospital gave approval for this retrospectivereview. The operating theatre log database andthe written theatre logbooks were searched forthe diagnosis of femoral fracture between January1998 andDecember 2002. A total of 128 sets of noteswere then reviewed and 35 patients excluded for avariety of reasons including: pathological fracture(18), proximal femoral neck fracture (8), growthplate injury (3) and initial operativemanagement ofthe fracture at another hospital site (6). This gave a

total of 93 patients with 95 femoral fractures (twobilateral).

Data on patient demographics, clinical detailsand treatment were then extracted from themedical records. The duration of hospital stayand events such as return to theatre within thefirst admission and early readmission within thefirst 30 days were noted. The location, type offracture and position at union were obtained fromreview of the patient radiographs and medicalrecords.

Statistical analyses were carried out using eitherparametric (ANOVA) or non-parametric (KruskalWallis test) statistics to compare means/mediansof groups. Correlations were tested using Pearson’scorrelation coefficients.

Results

Patient demographics

The mean age at injury was 6.7 � 3.7 years (range0.7—14 years). There were 35 females and 58males. The most common mechanism of injurywas a fall (46%), followed by pedestrians versuscar (25%) and then children involved in motor vehi-cle accidents (17%). Thirty-three of the children(36%) had multi-trauma, defined as one or moreadditional injuries. Associated injuries includedhead injuries in 19, spinal injuries in 3, chest,abdominal or pelvic injuries in 18 and other extre-mity injuries in 22.

Fracture management

Of the 95 fractures, 79 were located in the middlethird of the femoral shaft. Six were located in theproximal third and the remaining 10 occurred inthe distal third. All fractures were closed apartfrom four open fractures, all Gustilo and AndersonGrade I.

Forty-six fractures were treated by immediateclosed reduction and application of a hip spica cast.The other 49 fractures were treated by some form ofinternal or external fixation: flexible intramedullarynails in 20 fractures; external fixation in 21 frac-tures; and other methods of fixation (crossedK-wires, delta nails or open reduction and internalfixation with an AO plate) in 8 fractures.

Operative management of femoral fractures 933

Figure1 Boxplots of lengthof in-patient stay for childrenwith and without associated injuries treated by differentmethods of management. The graph shows the medianlength of stay (solid line), the 25th—75th quartiles (solidbox) and 5th—95th percentiles (whiskers). Note that boththe median length of stay and the range of lengths of stayare much less in the group without associated injuries.

Table 1 Patient demographics

Type of intervention(n = number of patients)

M:Fratio

Age in years(mean � S.D.)

Number of patients withassociated injuries (%)

Median days ofhospital stay (range)

Early hip spica (n = 46) 32:14 4.0 � 2.5 6 (13) 3 (1—15)

Operative fixation (n = 47)Flexible intramedullary nail (n = 20) 10:10 10.1 � 2.4 8 (40) 7 (3—52)External fixator (n = 20)a 10:10 8.3 � 2.5 16 (80) 21 (5—70)b

Other surgical fixation (n = 7)c 6:1 9.9 � 2.7 3 (43) 6 (2—20)

Total group (n = 93) 58:35 6.7 � 3.7 33 (36) 5 (1—70)a Twenty patients with 21 fractures.b One death at 24 h.c Seven patients with eight fractures.

Table 1 characterizes the gender ratio, age atinjury, associated injuries and the length of hospitalstay for children treated by different types of sur-gery. One patient with bilateral femur fracturesdied at 24 h, due to other injuries. Children treatedwith an early hip spica were significantly youngerthan all other groups ( p < 0.001). This group ofchildren were also the least likely to have associatedinjuries and had the shortest median length of stayin hospital ( p < 0.01).

The length of stay more than doubled with thepresence of associated injuries, regardless of themethod of management of the femoral fracture(Fig. 1). Patients with isolated femur fracturestreated either by hip spica application or by opera-tive fixation had median lengths of stay of 3 days(range 1—10 days) and 6 days (range 2—15 days),respectively. In contrast, patients who had femurfractures associated with other injuries treated byeither hip spica application or other operative fixa-tion hadmedian lengths of stay of 5 days (3—15 days)and 17 days (5—70 days), respectively. Length ofstay was moderately correlated with age in thosepatients with isolated femur fractures (r = 0.6365,p < 0.001) but not in the group of patients withassociated injuries (r = 0.07).

Table 2 outlines the time to surgical interven-tion for this cohort of 93 patients with 95 frac-tures. Eighteen patients waited more than 48 hafter injury for their definitive surgical manage-ment. The reasons for delay included inter-hospi-tal patient transfer, associated injuries takingprecedence and temporary equipment or operat-ing theatre non-availability. Overall, 9 of 93 chil-dren had an unplanned return to the operatingtheatre during their initial hospital stay (Table 3).In the hip spica group, three children, aged 6—8years, returned to theatre during their initialadmission. Two had a change of hip spica for castproblems (cracking and pressure) and one had anexternal fixator applied, for post-cast peronealnerve palsy. These three children were above

the 75th percentile for weight. Their lengths ofstay were increased to between 7 and 10 days, asa result.

In the flexible intramedullary nail group, threepatients required application of a single leg hip spicafor early loss of reduction. Two of the three patientshad increased length of stay as a result (13 and 14days, respectively). Three children in the externalfixator group required adjustment of the externalfixator within the first 7 days, to correct fracturemalalignment and excessive femoral shortening.

934 N.C. Wilson, N.S. Stott

Table 2 Time between admission and operative intervention

Type of intervention (n = number of patients) Time between admission and operative intervention

Within 24 h Between 24 and 48 h Greater than 48 h

Early hip spica (n = 46) 38 3 5

Operative fixation (n = 47)Flexible intramedullary nail (n = 20) 9 5 6External fixator (n = 20)a 10 4 6Other surgical fixation (n = 7)b 5 1 1

Total (n = 93) 62 13 18a Twenty patients with 21 fractures.b Seven patients with eight fractures.

Table 3 Hospital stay data

Type of intervention(n = number of patients)

Number of patients returningto operating room duringfirst admission (%)

Number of patientsreadmitted withinfirst 30 days (%)

Early hip spica (n = 46) 3 10 (22)

Operative fixation (n = 47)Flexible intramedullary nail (n = 20) 3 4 (20)External fixator (n = 20)a 3 5 (25)Other surgical fixation (n = 7)b 0 2 (29)

Total (n = 93) 9 (10%) 21 (23)a Twenty patients with 21 fractures.b Seven patients with eight fractures.

The adjustment of the external fixator did notimpact on their length of stay.

Readmission within the first 30 days

Readmissionwithin thefirst 30dayswaschosenas thisis a benchmark commonly used in NZ as a perfor-mance indicator for District Health Boards.14 Twenty-one of the 93 children (23%) had an unplannedreadmission within 30 days of injury (Table 3).

In the hip spica group, 10 children (22%) had anunplanned readmission in the first 30 days. Seven ofthe 10 children were admitted for loss of fractureposition at 7—10 days and the other three childrenrequired early change of the hip spica cast, two fordamage and one for pain. In the operative group,seven children were readmitted due to superficialinfections either of external fixator pin-sites (fivepatients) or of K-wire insertion sites (two patients).These infections respondedto intravenousantibioticsand did not require surgery. A further four childrenwho had undergone flexible intramedullary nailingwere admitted for early removal of a prominentwire.

Fracture position at union

An acceptable position at the time of fracture unionwas defined as less than 108 of valgus, less than 58 of

varus and less than 108 of antecurvatum or recur-vatum. Using this definition, the position at time ofunion was acceptable in all but 6 of the 95 fractures(6%). Four fractures in the hip spica group, one in theflexible intramedullary nail group and one in theexternal fixator group fell outside the definition ofacceptable position.

Subsequent review of these six fractures at anaverage of 18 months after the fracture showedsatisfactory remodelling, with no patients requiringa corrective osteotomy.

Discussion

The data from this study highlight the two verydifferent groups of children that are commonlyadmitted to our institution with a femoral fracture.Representative of the first group is the young child,less than 6—7 years of age, who presents with anisolated femur fracture as the result of a low velo-city injury. Management of these patients byimmediate hip spica was relatively expeditious with89% having their definitive treatment within 2 daysof admission and almost all being discharged homewithin 2 days of hip spica application. The secondgroup comprises mainly older children, who com-monly present with a femur fracture as part of a

Operative management of femoral fractures 935

constellation of other injuries. The timing of opera-tive fixation of the femoral fracture did not appearto have a major impact on the length of stay in thisgroup of children, who often had a prolonged in-patient hospital stay of greater than 10 days. In bothgroups, there was a significant incidence of earlycomplications that required further operative inter-vention, both during the first admission and withinthe first 30 days after discharge.

The length of in-patient hospital stay is oftenused as a benchmark of performance in the healthcare sector, with reduced length of stay being onegoal of treatment.6 Our data show that a key factorinfluencing the length of stay is not the type ofoperative fixation or age of the patient but ratherthe presence of associated injuries. The presence ofone or more associated injuries more than doublesthe length of in-hospital stay, regardless of the typeof fracture management. For children with isolatedfemur fractures, a younger age did correlate with areduced length of stay, due to an earlier dischargeafter surgical intervention. However, in childrenwith associated injuries, age did not correlate withlength of stay. These data have important implica-tions for how outcomes are measured by the healthcare system, particularly when statistics for lengthof stay are used as a benchmark to compare differ-ent providers or different types of treatment.

Overall, the length of stay for younger childrenmanaged by immediate hip spica application hasdecreased in our unit over the last 10 years. Inthe early 1990s, the length of stay averaged 7 days5

but now is a median of 3 days. This reduction inlength of stay has been paralleled by a similarreduction in the upper age limit for children treatedby immediate hip spica application, which has fallenfrom 10 years in the previous study5 to 7 years in thecurrent study. These changes in our practice mirrorcurrent international trends, with increasing sur-geon preference for alternative methods of man-agement of femoral fractures in children over 6years of age.10,18

The majority of children managed by immediatehip spica application were discharged home the dayafter spica application, suggesting that earlieraccess to the operating theatre could reduce thelength of in-patient stay in this group. Early com-plications of hip spica management did significantlydelay the discharge of three patients. Two of thesethree children were older, at 7 years of age, and allthree of these children were on or above the 75thpercentile for weight relative to their chronologicage. Recent orthopaedic literature has highlightedthat obese children appear to have more complica-tions after femoral fracture management than chil-dren who are not obese.12 Our findings support the

concept that, for children aged over 6 years, thechoice of intervention for femoral fracture needs totake into account not only the age but also theweight of the child.

The median length of stay for older children withisolated femur fractures treated by operative fixa-tion was only 6 days. This length of stay is consistentwith other studies that report median length of staybetween 5 and 8 days for children with isolatedfemur fractures treated either by flexible intrame-dullary nailing11,15 or external fixation.8,16 Delays of1 to 2 days before definitive surgery were common inthis group of patients but it was difficult to deter-mine from this retrospective review how significantthis factor was in their overall length of stay.

Six of the 47 patients with femoral fracturestreated by operative fixation required further sur-gical intervention during the first admission, butonly two had an increased length of in-patient stayas a result. Unsatisfactory fracture alignment and/or excessive femoral shortening following applica-tion of an external fixator were seen in three cases.In retrospect, long leg films taken in theatre inaddition to image intensifier views may have helpedto detect these problems at the time of the firstoperation. Although revision of the external fixatorwas required, this did not adversely impact onhospital stay as all three patients had prolongedlengths of stay due to other injuries from multi-trauma.

The other three patients had early loss of reduc-tion at the fracture site following flexible intrame-dullary nailing, requiring a single leg hip spica toaugment fracture fixation. The fracture patternsseen in these three patients could be classified as‘length unstable’,19 with butterfly fragments andgreater than 25% comminution at the fracture site.It is now recognised that these types of fracturepatterns are associated with high rates of failure offixation by flexible intramedullary nails.15,19,20

These fractures would probably have been bettermanaged in other ways.

Almost one quarter of our patients were read-mitted at some point during the first 30 days afterdischarge. Common reasons for readmission wereloss of fracture position or cast damage in the hipspica group; pin-site infection in the group treatedby external fixation (external fixator or Steinmanpins); and pain from the nail tip in the flexibleintramedullary nailing group. Most of these admis-sions are probably not completely avoidable. Forexample, the incidence of pin-site infections follow-ing external fixation of the femur are anywhere from37 to 73%.9,13 Likewise change in fracture alignmentduring hip spica treatment is common, requiringeither wedging of the cast or remanipulation.5,21,23

936 N.C. Wilson, N.S. Stott

The one clearly avoidable problem was pain due todistal prominence of a flexible intramedullary nail,which is a technical complication. The incidence ofthis problem may now have been reduced by sub-sequent changes in surgical technique, which do notleave the wires as prominent as previously15.

Conclusion

In summary, this paper highlights the trends infemoral fracture management in our institution.These include increased utilization of internal orexternal fixation of fractures in children over theage of 6 years and a high incidence of associatedinjuries leading to a prolonged in-patient stay,despite early fixation of the femoral fracture. Over-all, the operative management of femoral fracturesremains resource intensive, with a high incidence offurther minor interventions both at the first admis-sion and within the first 30 days to produce asatisfactory outcome.

Conflicts of interest statement

We declare that there are no conflicts of interest.

References

[1] Al-Habdan I. Diaphyseal femoral fractures in children:should we change the present mode of treatment? Int Surg2004;89:236—9.

[2] Beaty J, Rockwood C, Wilkins K, Kasser J. Rockwood andWilkins’ fractures in children. Philadelphia: Lippincott Wil-liams & Wilkins; 2000. p. 893.

[3] Buckley SL. Current trends in the treatment of femoral shaftfractures in children and adolescents. Clin Orthop Relat Res1997;60—73.

[4] Buechsenschuetz KE, Mehlman CT, Shaw KJ, et al. Femoralshaft fractures in children: traction and casting versuselastic stable intramedullary nailing. J Trauma 2002;53:914—21.

[5] Ferguson J, Nicol RO. Early spica treatment of pediatricfemoral shaft fractures. J Pediatr Orthop 2000;20:189—92.

[6] Gallagher C. Applying quality improvement tools to qualityplanning: pediatric femur fracture clinical path develop-ment. J Healthc Qual 1994;16:6—14.

[7] Gardner MJ, Lawrence BD, Griffith MH. Surgical treatment ofpediatric femoral shaft fractures. Curr Opin Pediatr2004;16:51—7.

[8] Hedin H, Borgquist L, Larsson S. A cost analysis of threemethods of treating femoral shaft fractures in children: acomparison of traction in hospital, traction in hospital/homeand external fixation. Acta Orthop Scand 2004;75:241—8.

[9] Hedin H, Hjorth K, Rehnberg L, Larsson S. External fixation ofdisplaced femoral shaft fractures in children: a consecutivestudy of 98 fractures. J Orthop Trauma 2003;17:250—6.

[10] Heyworth BE, Galano GJ, Vitale MA, Vitale MG. Managementof closed femoral shaft fractures in children, ages 6 to 10:national practice patterns and emerging trends. J PediatrOrthop 2004;24:455—9.

[11] Houshian S, Gothgen CB, Pedersen NW, Harving S. Femoralshaft fractures in children: elastic stable intramedullarynailing in 31 cases. Acta Orthop Scand 2004;75:249—51.

[12] Leet AI, Pichard CP, Ain MC. Surgical treatment of femoralfractures in obese children: does excessive body weightincrease the rate of complications? J Bone Joint Surg Am2005;87:2609—13.

[13] Miner T, Carroll KL. Outcomes of external fixation of pedia-tric femoral shaft fractures. J Pediatr Orthop 2000;20:405—10.

[14] Ministry of Health. Toward clinical excellence: an introduc-tion to clinical audit, peer review and other clinical practiceimprovement activities. http://www.moh.govt.nz/moh.nsf; 2006.

[15] Narayanan UG, Hyman JE, Wainwright AM, et al. Complica-tions of elastic stable intramedullary nail fixation of pedia-tric femoral fractures, and how to avoid them. J PediatrOrthop 2004;24:363—9.

[16] Nork SE, Hoffinger SA. Skeletal traction versus externalfixation for pediatric femoral shaft fractures: a comparisonof hospital costs and charges. J Orthop Trauma 1998;12:563—8.

[17] Ozdemir HM, Yensel U, Senaran H, et al. Immediate percu-taneous intramedullary fixation and functional bracing forthe treatment of pediatric femoral shaft fracture. J PediatrOrthop 2003;23:453—7.

[18] Sanders JO, Browne RH, Mooney JF, et al. Treatment offemoral fractures in children by pediatric orthopedists:results of a 1998 survey. J Pediatr Orthop 2001;21:436—41.

[19] Sink EL, Gralla J, Repine M. Complications of pediatric femurfractures treated with titanium elastic nails: a comparisonof fracture types. J Pediatr Orthop 2005;25:577—80.

[20] Slongo TF. Complications and failures of the ESIN technique.Injury 2005;36(Suppl 1):A78—85.

[21] Sugi M, Cole WG. Early plaster treatment for fractures of thefemoral shaft in childhood. J Bone Joint Surg Br 1987;69:743—5.

[22] Wright JG. The treatment of femoral shaft fractures inchildren: a systematic overview and critical appraisal ofthe literature. Can J Surg 2000;43:180—9.

[23] Wright JG, Wang EE, Owen JL, et al. Treatments for pae-diatric femoral fractures: a randomised trial. Lancet2005;365:1153—8.