Injury, Int. J. Care Injured 41 (2010) 388–395

How comparable is so-called standard fracture fixation with an identical implant?A prospective experience with the antegrade femoral nail in South Africa andEurope

Thomas Gross a,*, Tobias Huettl b, Laurent Audige b, Chris Frey c, Mauro Monesi d,1, Franz Josef Seibert e,Peter Messmer f,2

a Computer Assisted Radiology & Surgery (CARCAS), University Hospital Basel, Realpstrasse 54, CH-4057 Basel, Switzerlandb AO Clinical Investigation and Documentation, AO Foundation, Clavadelerstrasse 8, CH-7270 Davos Platz, Switzerlandc Chris Hani Baragwanath Hospital, PO Bertsham, Johannesburg 2013, South Africad Ospedale S. Maria Misericordia Udine, I-33100 Udine, Italye Medical University Graz, A-8036 Graz, Austriaf University Hospital Zurich, CH-8091 Zurich, Switzerland

A R T I C L E I N F O

Article history:

Accepted 12 October 2009

Keywords:

Fracture fixation

Femur

Intramedullary nailing

Outcome

Intercontinental comparison

Observational study

Developing countries

South Africa

Europe

Trauma centre

A B S T R A C T

Background: The utilisation and consequences of standardised operative procedures may importantly

differ between different healthcare systems. This is the first investigation comparing the treatment and

outcome of femoral shaft fractures stabilised with an identical implant between trauma centres in 2

continents (Europe, EU and South Africa, SA).

Methods: Following standardised introduction of the technique, the prospective, observational

multicentre study enrolled 175 patients who underwent intramedullary fracture fixation using the

antegrade femoral nail (AFN) for femoral shaft fractures. Eleven EU hospitals recruited 86 patients and 1

SA centre 89 patients in the study period. Comparison of epidemiologic data, operative characteristics as

well as subjective (e.g., pain, SF-36) and objective (e.g., X-ray, range of motion [ROM]) 3-month and 1-

year outcomes were performed (p < 0.05).

Results: Compared to EU centres, several significant differences were observed in SA: (1) on average,

patients operated on were younger, had less concomitant diseases and had more severe open fractures;

(2) operative stabilisation was more often undertaken by young, unsupervised residents, with shorter

operating and intraoperative fluoroscopy times; (3) mean hospital stay was shorter, with less recorded

complications, but a higher loss to follow-up rate. Non- or malunion rates and subjective outcomes were

similar for both groups, with the physical component of the SF-36 at the 1-year follow-up not fully

restoring to baseline values.

Conclusions: Our investigation demonstrates the importance of several major differences between 2

different regions of the world in the treatment of femoral shaft fractures, despite involving only high

level trauma centres and using an identical implant. The intercontinental comparison of results from

clinical studies should be interpreted very carefully considering the heterogeneity of populations and

clinical settings.

� 2009 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Injury

journa l homepage: www.e lsevier .com/ locate / in jury

Introduction

The healthcare disparity between developed and developingcountries can be very large. Differences between low- and middle-

* Corresponding author. Tel.: +41 61 381 5447.

E-mail address: tgross@uhbs.ch (T. Gross).1 Current address: Ospedale M. Bufalini, I-47023 Cesena (FC), Italy.2 Current address: Emergency and Trauma Center, Rashid Hospital, DOHMS,

Dubai, United Arab Emirates.

0020–1383/$ – see front matter � 2009 Elsevier Ltd. All rights reserved.

doi:10.1016/j.injury.2009.10.021

income countries (LMIC) and high-income countries (HIC) havebeen reported in type of trauma care offered from the prehospitalperiod until recovery of patients.30,35,44 In addition, there are alsodifferences in mechanisms of injury. In general, LMIC such asSouth Africa (SA) demonstrate a much higher burden of injuriesdue to homicide or interpersonal violence in comparison toEurope (EU).19,28,46 Some of the specific problems in surgical carefaced by LMIC are lack of equipment and resources,1,24,27 poorlystandardised classification of surgical operations,27 poor report-ing,24,27 poor follow-up rate,2,17 and lack of access to medicalliterature.26

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395 389

Intramedullary nailing of femoral fractures is a standardprocedure in orthopaedic surgery.9,10,16,34,36 However, typical forsurgical interventions, factors such as different injury types,patient populations, surgical experience and approaches or thetype of implant used may have an important impact on theoperation and the subsequent outcome of patients. This may beespecially valid in the comparison of interventions between highlydeveloped and developing countries.

There are very few publications specifically on the treatment offemoral shaft fractures in low- and middle-income countries(LMIC) and we could not find any intercontinental comparisonbetween LMIC and HIC in this context. Studies that compare datafrom different regions in the world can help to better understandclinical settings in terms of healthcare and study implementation.

The prospective study was designed to investigate thestabilisation of femoral shaft fractures with the antegrade femoralnail (AFN), which was a new implant at the time of the study andwas accordingly introduced in all centres. The objective of thisevaluation was to elucidate on important differences regarding thebasic characteristics of patients and fractures, their treatment and1-year objective and subjective outcomes between trauma centresin EU and SA, nevertheless executing a well-standardisedprocedure such as intramedullary nailing with an identicalimplant.

Patients and methods

The prospective multicentre study included patients from 11trauma centres in EU and 1 centre in Johannesburg, SA in theintroduction period of the antegrade femoral nail (AFN). It wasconducted according to Good Clinical Practice guidelines, approvedby the Institutional Review Boards of all collaborating centres, andmonitored by the Arbeitsgemeinschaft fur Osteosynthese – ClinicalInvestigation and Documentation (AOCID), Davos, Switzerland. Allparticipants provided written informed consent.

One hundred and seventy-five (n = 89 in SA; n = 86 in EU)patients (average age 37.5; range 18–95 years) who sustained afemoral shaft fracture (AO type 32), either as a monotrauma,within multiple fractures or multiple trauma cases and with anyseverity of soft-tissue injury (closed as well as any grade of openfractures) were enrolled in the study. Patients had to receivedefinitive surgical repair with the AFN within 7 days afteroccurrence of the fracture; conversion from emergency externalfixation to the AFN was allowed. Patients with existing implants ofthe fractured hip and femur, those younger than 18 years, unableor unlikely to cooperate, or patients who had participated in anyother clinical device or drug trial within the previous month wereexcluded.

Surgeons in the 12 centres were advised to perform theoperative procedure following a standard protocol prepared by theimplant manufacturer and the 2 principal investigators (TG andPM), illustrated by an educational video.23 Standard perioperativetreatment followed international AO as well as hospital guidelines,with the aim of achieving postoperative movement and weight-bearing as early as possible.

Patient data were collected at baseline, 6 and 12 weeks and 12months after surgery using identical datasheets in all participatingcentres. Baseline assessments included demographics, body massindex (BMI), American Society of Anaesthesiologists (ASA) riskscores, and fracture characteristics based on the AO classification.Soft-tissue injuries in closed fractures were graded according toTscherne and Oestern.42 Open fractures were classified accordingto Gustilo and Anderson.13 Patients underwent radiographicexamination of the femoral shaft with the adjacent joints in theanteroposterior and lateral views at study entry and duringscheduled follow-up visits. Details of the procedure including the

surgeon’s general (resident vs. consultant level) and specificexperience (number of previous interventions undertaken with theproximal femur nail [PFN], a similar implant), or need ofsupervision (intervention was instructed/assisted by a moreexperienced surgeon vs. independent responsibility withoutsupervision), open or closed reduction, reamed or unreamedinsertion, surgery and fluoroscopy times were documented.

Complications were recorded throughout the study follow-upof 1 year. All complications and radiographs were reviewed by theprincipal investigators (TG and PM). Distinction was madebetween local (e.g., surgical site infection [SSI], delayed, mal-and nonunion) and systemic complications. The definitions forfracture healing were adapted from McLaren et al.22, with thepresence of bridging callus and the ability of the patient to bear hisor her own weight without pain indicating union. Nonunion wasassumed in case of (a) a delayed union, i.e. insufficient healing inthe first 6 months after the AFN procedure where the surgeondecided for early reoperation; or (b) failed union/absence ofbridging callus after >6 months. Malunion was defined using Ricciet al.’s definition of malalignment.34 Further objective outcomeassessments included the length of hospital stay, any reoperationover time and the range of motion (ROM) at follow-up visits.

Subjective criteria included (a) patient-reported pain registeredby a visual analogue scale (VAS 0–10) as well as graduationfollowing the Harris Hip Score at 12 weeks and 1 year and (b) thequality of life assessment using the Short Form 36 (SF-36)questionnaire, comparing in particular patients’ self-assessmentof their preoperative status (interviewed during the initial hospitalstay) with the 12-month follow-up assessment.

Data management, including data entry, plausibility checks,and query generation was performed by AOCID. Data wereprocessed with Qualicare Version 9.1a25 database (Qualidoc,Trimbach, Switzerland) and linked with the digitised radiographicimages. All statistical analyses were conducted with the softwareIntercooled Stata Version 10 (StataCorp LP, College Station, TX,USA). EU and SA population baseline characteristics werecompared; continuous variables were described using means,standard deviations (SD) and ranges (or medians and inter-quartileranges as appropriate) and compared using the T-test; categoricalvariables were tabulated with absolute and relative frequenciesand compared using the Fisher’s exact test. Patient baseline factorswere compared between the group of patients examined 1 yearpostsurgery and the group of patients lost to follow-up at 1 year asabove; factors associated with 1-year loss to follow-up usingunivariable statistics at a significant level of 0.10 were examinedtogether in a multivariable logistic regression model with abackward selection procedure. The significant level for selection inthe final model was set to 0.05. Relative risks (RR) were derived bymultivariable binomial regression. Outcome parameters werecompared similarly between EU and SA patient groups; howevercontinuous outcomes were analysed using multivariable linearregression. The following potential influencing factors wereinvestigated: EU (vs. SA), age <65 years, BMI >30, single injury(vs. multiple fracture or trauma), open fracture, ASA score III–V (vs.ASA score I + II), surgeon experience, nail reaming, centre enrollingat least 10 patients, occurrence of a local complication andoccurrence of delayed or nonunion. All comparative analyses wereexploratory without a predetermined sample size or powercalculation; 95% confidence intervals (CI) were computed fordifference in proportions or means as appropriate to supportinterpretation of clinical relevance. In addition, outcome changesbetween 3-month and 1-year follow-up (e.g., ROM parameters) orbetween baseline and 1-year follow-up (e.g., SF-36) wereexamined by paired T-test. To examine potential biases in datacollection, all analyses were reimplemented by excluding EUcentres which have enrolled <10 patients. These results are

Table 1Baseline characteristics of patients.

Characteristic South Africa (n = 89) Europe (n = 86) p-Value All patients (n = 175)

Age (years� SD) 33.4�11.2 41.7�23.3 0.003* 37.5�18.6

Gender (n, %)

Male 69 (78%) 63 (73%) 0.599 132 (75%)

Female 20 (22%) 23 (27%) 43 (25%)

Body mass indexa (mean� SD) 25.4�3.2 24.9�4.2 0.361 25.2�3.7

Number of concomitant diseases (n, %)

0 78 (88%) 26 (30%) <0.001* 104 (59%)

1 5 (6%) 36 (42%) 41 (23%)

2 1 (1%) 11 (13%) 12 (7%)

3 3 (3%) 5 (6%) 8 (5%)

4 2(2%) 6 (7%) 8 (5%)

5 0 2 (2%) 2 (1%)

ASA Classification (n, %)

ASA Class I 81 (91%) 50 (58%) <0.001* 131 (75%)

ASA Class II 4 (4%) 21 (24%) 25 (14%)

ASA Class III 2 (2%) 11 (13%) 13 (7%)

ASA Class IV 2 (2%) 2 (2%) 4 (2%)

ASA Class V 0 2 (2%) 2 (1%)

Trauma type (n, %)

Single fracture 56 (63%) 46 (54%) <0.001* 102 (58%)

Multiple fractures 1 (1%) 20 (23%) 21 (12%)

Multiple trauma 32 (36%) 20 (23%) 52 (30%)

Injury severity score (mean� SD) 39.7�11.4 33.1�8.6 0.032* 37.1�10.8

Fracture side (n, %)

Right 49 (55%) 42 (49%) 0.451 91 (52%)

Left 40 (45%) 44 (51%) 84 (48%)

AO classification (32) (n, %)

A 15 (17%) 39 (45%) <0.001* 54 (31%)

B 43 (48%) 29 (34%) 72 (41%)

C 31(35%) 18 (21%) 49 (28%)

Open fracture (n, %)

Yes 39 (44%) 9 (10%) <0.001* 48 (27%)

No 50 (56%) 77 (90%) 127 (73%)

Open injury classification (n, %)

I and II8 3 (8%) 7 (78%) <0.001* 10 (21%)

III8 36 (92%) 2 (22%) 38 (79%)

8Grade.* Significant.a BMI: SA n = 88; EU n = 85.

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395390

presented only to provide possible explanation in the observedgroup differences.

Results

Baseline characteristics of patients are shown in Table 1. Onaverage, the SA trauma centre recruited about 10 times as manypatients for the investigation as the EU hospitals (mean 8; range 2–20 cases). SA patients were found to be significantly younger(average 33.4; range 18–85 years) than EU patients (average 41.7;range 18–95 years) and healthier. Only 4% (n = 4) of patients in SAhad a high ASA risk class (III, IV or V) compared to 17% (n = 15) inthe EU centres (p = 0.007). The extension of trauma wassignificantly different between SA and EU centres with, on onehand more single fractures and more severe multiple trauma casesin SA, and on the other hand more multiple fractures in EU. Femurfractures were more often open and overall more severe in SA. Themajority of injuries in both continents were due to vehicularaccidents (n = 44 [49%] in SA; n = 59 [69%] in EU). However, thesecond most common injuries in SA were gunshot wounds (n = 32[36%]) followed by falls (n = 7 [8%]). In SA, 82% (32/39) of openfractures and 13% (4/32) of multiple trauma cases were due togunshots. Within the EU cases, falls accounted for 25% (n = 21) of

all injuries and no gunshot injuries were reported. Overall, 68% ofall open fractures and 8% of multiple trauma cases were due togunshots.

The AFN operation was executed in 88% of SA cases (n = 78) onthe day of or the day after the injury but in only 63% of EU cases(n = 54; p < 0.001). In SA, the surgeons who performed theprocedures in >9/10 of cases were residents (Table 2). This is incontrast to the EU centres where>4/5 of executing surgeons had aconsultant level. However, surgeons in SA executed their inter-ventions more often without any supervision (>9/10 of cases)compared to their EU counterparts (2/3 of cases, p = 0.002). Theaverage operative time was significantly shorter for the SA group(range 40–135 min) compared to the EU group (range 25–230 min)and correlated with the time needed for intraoperative fluoroscopy(r = 0.48, p < 0.001). The rate of reaming cases did not differbetween SA or EU centres and tended to lengthen the nailingprocedure on average by 5 min. Open fractures were 1.3 timesmore likely to get reamed than closed fractures (p = 0.06).Following multivariable analysis, 4 factors were found to beassociated with the operation time: (1) an intervention undertakenin an EU centre (vs. SA; 40 min longer, 95% CI 26–54, p < 0.001), (2)intraoperative reduction undertaken in a closed manner (23 minlonger, 95% CI 7–39, p = 0.006), (3) patient with an ASA score III or

Table 2Details of operative intervention.

Variable South Africa (n = 89) Europe (n = 86) p-Value All patients (n = 175)

Surgery time (min� SD) 61.8�14.4 96.2�41.4 <0.001* 78.7�35.2

Fluoroscopy time (min� SD) 1.3�0.6 6.8�7.8 <0.001* 3.9�6.1

Reduction (n, %)

Open 7 (8%) 7 (8%) 1.000 14 (8%)

Closed 82 (92%) 79 (92%) 161 (92%)

Reaming (n, %)

Yes 56 (63%) 43 (50%) 0.095 99 (57%)

No 33 (37%) 43 (50%) 76 (43%)

Iatrogenic fracture (n, %)

Yes 1 (1%) 3 (3%) 0.362 4 (2%)

No 88 (99%) 83 (97%) 171 (98%)

Surgeons’ experience (n, %)

Chief surgeon or consultant 8 (9%) 70 (81%) <0.001* 78 (45%)

Resident 81 (91%) 16 (19%) 97 (55%)

Surgery supervision (n, %)

Total or partiala 5 (6%) 19 (22%) 0.002* 24 (14%)

Unsupervised 84 (94%) 67 (78%) 151 (86%)

* Significant.a Teaching surgery.

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395 391

IV (15 min longer, 95% CI 1–30, p = 0.041), (4) surgeons with lessprevious PFN experience and were supervised during the inter-vention (30 min longer, 95% CI 4–54, p = 0.021). Intraoperativefluoroscopy times were on average 5.5 min longer in EU (95% CI4.8–8.2, p < 0.001) compared to SA, and on average 2.2 min shorter(95% CI 0.5–3.9, p = 0.012) in interventions undertaken by surgeonswith a consultant level.

Table 3Local and systemic complications.

Complications (no of events) South Africa

n Reoperated fem

Local

Complications due to the implant

Intraoperative crack of femur corticalis 1

Breakage of distal locking screw(s) 1

Migration/telescoping recon screw 2 1

Painful distal interlocking screw

Healing problems

Nonunion 1

Malunion 4

Surgical site infections

Superficial, treated conservatively 1

Deep infection associated with sepsis + nonunion 1 1

Deep infection with haematoma + osteomyelitis

Other

Compartment syndrome

Haematoma

Peroneal nerve palsy 1

Systemic

Thromboembolytic events

Deep vein thrombosis 1

Pulmonary embolism

Fat embolism

Infections

Pneumonia

Urogenital

Other infections

Others

Pressure sore 1

Breathing problem

Cardiac failure, suspected

Death due to gunshot wound, post operative 1

n = number of complication cases (with several patients suffering more than 1 complica 1 case was operated for malunion with breakage of interlocking screws and is cou

The overall complication rate (Table 3) was significantly higherin EU where 35 patients suffered from 43 complications vs. 13 SApatients with 15 complications (p = 0.006). When the probability ofthe occurrence of a local complication was tested with binomialregression analysis after controlling for location (EU vs. SA),reaming was associated with a half risk reduction of localcomplications (p = 0.035). For the most frequent complication,

Europe All

ur n Reoperated femur n Reoperated femur

3 4

9a 4 10 4

1 1 3 2

1 1 1 1

4 3 5 3

2a 6

1 2

1 1

1 1 1 1

1 1 1 1

1 1 1 1

3 4

2 3

1 1

1 1

6 6

3 3

2 2

1

1 1

3 3

1

ation).

nted as reoperation for the latter.

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395392

i.e. broken interlocking screws (n = 10), spontaneous approxima-tion of fracture ends could be observed in all cases. EU centresshowed a significantly higher rate (p = 0.034) of such episodes. Theincidence of broken interlocking screws did not depend onwhether nailing was undertaken in a reamed or unreamed manner(p = 0.7). Nevertheless, unreamed interventions appeared to have a3 times higher risk for a reoperation compared to those reamed (12vs. 4%; p = 0.078). Overall, 5 times as many EU patients (11/86,12.8%) had to be reoperated compared to SA patients (2/89, 2.2%;RR = 5.7; 95% CI 1.3-24.9; p = 0.009). For all reoperations, furtherreported healing was uneventful. In terms of systemic complica-tions, all thromboembolic events occurred under pharmacologicalthromboembolic prophylaxis and 1 case of fat embolism followingreamed nailing. All complications resolved by 1 year except 1peroneal nerve palsy that partially persisted. Five patients diedduring the 1-year follow-up: 4 elderly patients and 1 gunshotwound victim during the follow-up period; all deaths occurred atleast 2 months following the intervention and appeared to beunrelated to either accident or intervention.

Subjective 1-year outcomes did not demonstrate any differencewhether patients were operated in SA or EU (Table 4). Fifty percentof all patients (EU: 43%, SA: 60%; p > 0.1) declared having no painat all. Patients with a nonunion presented with a mean 15 pointslower physical component score of the SF-36 than other patients(95% CI 8.3–22; p < 0.001). The same patients were 24 times (95%CI 6.5–92) more likely to rate a higher (moderate or marked) meanpain level in the Harris Hip Scoring compared to patients withuneventful healing of the fracture (p = 0.001). Patients who

Table 4Outcomes.

South Africa

Follow-up (no. of patients, %)

At baseline 89 (100%)

At 3 months 72 (81%)

At 1 year 56 (63%)

Lost to follow-up 33 (37%)

Hospital stay (days) 11.8�10.8

Complications (no. of patients, %)8Local complications 11 (13%)

Nonunion 4 (5%)

Malunion 2 (2%)

Systemic complications 3 (3%)

Mortality (no. of patients, %)

After 30 days 0

After 1 year 1 (2%)

Subjective outcomes

Pain at fracture site (VAS 0–10� SD)a

At 3 months 2.1�1.1

At 1 year 0.7�1.0

Hip pain at 3 months (no. of patients, %)

None to slight 54 (76%)

Mild to moderate 15 (21%)

Marked to disabled 2 (3%)

Hip pain at 1 year (no. of patients, %)

None to slight 51 (93%)

Mild to moderate 4 (7%)

Marked to disabled 0

SF-36 scoresb (mean� SD)

Physical health at baseline 53.4�6.3

Mental health at baseline 52.3�5.0

Physical health at 1 year 51.9�4.2

Mental health at 1 year 52.0�3.0

8Total number of patients with complications might differ from sum of local and syste* Significant.a VAS pain at 3 months: n = 72 for SA; 81 for EU VAS at 1 year: n = 56 for SA; 72 forb SF-36 scores at 1 year: n = 56 for SA; 69 for EU.

suffered at least 1 complication rated their mean pain on average1 point higher than patients without complications (VAS 0–10; 95%CI 0.3–1.8; p = 0.005) at 1 year postinjury. In both SA and EU, themean level of pain diminished between the 3-month and 1-yearvisit (VAS: p < 0.0001; Harris Hip: p = 0.002) down to an average of<VAS 1. We observed a tendency of lower physical but highermental SF-36 values in EU patients at baseline as well as at the 1-year follow-up. In both study groups the physical component of theSF-36 at 1 year did not fully restore to baseline values (SA: p < 0.05,EU: p < 0.001). Comparing preinjury and 1-year detailed functionparameters of the SF-36, the capacity to walk longer distances or tolift weights did not appear to be negatively influenced by the injuryand operation in either of the study groups. In contrast, on averagepatients of both study groups described a higher body pain (SA:p = 0.0001; EU: p = 0.002) and a negative impact on their workingcapacity following the injury (SA: p = 0.0001; EU: p = 0.008). At the1-year follow-up, 10 out of 12 of ROM measurements showedsignificant differences between the 2 study groups, about half ofthem with better results in SA or EU (Table 5). Between the 3-month and 1-year follow-up visits there was no significantimprovement in overall ROM measurements. Apart from thegeographic location (SA vs. EU), no other factors such as age >60,BMI >30, open fracture type, or the occurrence of complicationsshowed any relevant impact on the ROM measurements.

The overall 1-year follow-up rate was 73% and was significantlylower (p = 0.002) among SA (63%) compared to EU patients (84%).Following multivariable analysis, the patient group with missing1-year follow-up (n = 47) only differed significantly from those

Europe p-Value All patients

86 (100%) 175 (100%)

82 (95%) 154 (88%)

72 (84%) 128 (73%)

14 (16%) 0.002* 47 (27%)

21.4�20.4 <0.001* 16.5�16.9

21 (24%) 0.078 32 (19%)

1 (1%) 0.21 5 (3%)

4 (4%) 0.68 6 (4%)

19 (22%) <0.001* 22 (13%)

0 0

4 (6%) 0.385 5 (4%)

1.9�1.9 0.361 2.0�1.6

1.0�1.5 0.209 0.9�1.3

52(64%) 0.259 106 (70%)

26 (32%) 41 (27%)

3 (4%) 5 (3%)

63 (88%) 0.744 114 (90%)

8 (11%) 12 (9%)

1 (1%) 1 (1%)

51.7�10.3 0.199 52.6�8.5

53.8�7.4 0.130 53.0�6.3

49.7�9.0 0.086 50.7�7.3

53.6�6.8 0.105 52.9�5.4

mic complications as some patients may have both.

EU.

Table 5Range of motion (ROM).

Factors South Africa (n = 56) Europe (n = 71) p-Value All patients (n = 127)

Mean� SD Mean� SD Mean� SD

ROM hip

Flexion 140.5�5.9 115.6�21.9 <0.001* 126.6�20.9

Extension 0�0 8.4�11.3 <0.001* 4.7�9.4

Abduction 33.4�6.1 43.9�12.0 <0.001* 39.3�11.1

Adduction 20.2�2.3 25.6�8.0 <0.001* 23.2�6.7

Abduction at 908 knee flexiona 43.2�6.6 53.9�12.4 <0.001* 49.2�11.5

Adduction at 908 knee flexiona 16.8�4.7 19.6�10.4 0.059 18.4�8.4

External rotation 38.8�3.3 36.0�8.0 0.017* 37.2�6.5

Internal rotation 29.1�5.1 34.8�11.3 0.001* 32.3�9.5

External rotation at 908 knee flexion 37.5�4.8 36.0�8.7 0.245 36.7�7.3

Internal rotation at 908 knee flexion 16.9�7.4 34.7�11.3 <0.001* 26.9�13.2

ROM knee

Flexion 142.0�15.3 126.5�15.6 <0.001* 133.3�17.2

Extension 4.7�1.1 2.5�3.6 <0.001* 3.5�3.0

* Significant.a n for SA = 70.

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395 393

with complete follow-up in terms of geographic location (higherloss to follow-up in SA, p < 0.001) and age (mean 43 � 22.8 yearsfor patients lost to follow-up vs. 36 � 16.5 years for patients whocompleted the study; p = 0.015). Binomial regression analysis showedthat EU patients were 1.5 times more likely to complete the studycompared to SA patients (95% CI 1.2–1.7, p < 0.001) and patientsunder 60 years were also 1.5 times more likely to complete the studycompared to older patients (95% CI 1.1–21, p = 0.015). As the caseloadof the centres participating in the study differed significantly, we alsoperformed a separate analysis for the main parameters comparingonly the subgroup of patients from the 3 EU centres providing �10cases (n = 46) with patients from the SA centre. The subgroup analysisshowed that the patient characteristics, the operative procedure aswell as the outcome of patients in these EU centres did notsignificantly differ compared to SA patients (data not shown).

Discussion

As far as we are aware of, this study is the first to compare thecharacteristics and outcomes of an operative standard fracturemanagement between high level trauma centres in EU and SA.From a sociocultural perspective, several big differences in thehealthcare systems of European countries and those of LMIC suchas SA have been reported.5,11,21,29,35,41 In our investigation, thestandard intervention demonstrated to be effective in SA and EUpatients, with operation and fluoroscopy times as well as healingrates being well within the accepted ranges reported in theliterature.9,37,38,43 When comparing the 2 study groups, however,we observed several significant and interesting differencesbetween SA and EU at baseline as well as in subjective andobjective outcomes during the follow-up period. Most obvious, inthis multicentre study, the SA trauma centre recruited as manypatients as all participating EU centres together. The worldwidedifference in acute trauma caseload is well known with 90% ofsevere fractures occurring in LMIC.47 In LMIC, there exist within-country disparities in quality of care depending on geographiclocation (city vs. countryside) and medical facilities in large urbanareas like Johannesburg are overwhelmed by high trauma case-loads due to injuries incurred through violence and vehicularaccidents.41 In the province of Gauteng where the SA study centreis located, the annual mean number of trauma cases is 11,023 perfacility, the highest in the country.41 This is 3.6 times higher thanthose reported at Queens Medical Centre in Nottingham (UK),apparently EU’s largest teaching hospital.7 Indeed, the densedistribution of high quality hospitals in most of EU adds to theresulting lower number of trauma and emergency cases even in

well-established trauma centres.15,45 In contrast, and given its highquality infrastructure, the participating SA trauma centre appearsto be more representative of high volume centres with animportant percentage of high energy trauma, such as those inthe US.4,20,43 In fact, the SA patient population was much younger,typical for the age groups associated with homicide and inter-personal violence19,41 and presented with less concomitantdiseases than their EU counterparts. The mechanisms and patternof injury were also significantly different, with SA injuries (highincidence of gunshot wounds due to interpersonal violence)resembling those described in Africa or the Americas,19,28,41 butalso in the US.4,20,43 The incidence of homicide/interpersonalviolence in SA is 113.4 per 100,000 males which is 8.2 times higherthan the global rate of 13.9.28 In contrast, Western Europeancountries have a very low incidence of injuries due to interpersonalviolence,19,46 with a European Union average of 1.7 homicide casesper 100,000 inhabitants between 1998 and 2000.3

In general, surgeons in developing countries have beendescribed as ‘‘innovative and technically proficient’’, as they haveto be able to do more with less, such as inserting implants evenwithout the aid of C-arm imaging.1,47 On the other side, some LMIChave specialised high volume urban hospitals with an infrastruc-ture that may be well comparable to developed countries2,24 suchas the SA centre in this study. Interestingly, in contrast toEU centres, femoral nailing in SA was almost always undertakenby unsupervised trainees. Despite this fact, they performedfracture stabilisation with relevantly shorter operating andintraoperative fluoroscopy times compared to board-certified EUtrauma surgeons. Given the comparable outcome of patients, ourinvestigation therefore provides evidence that standard stabilisa-tion of femoral shaft fractures such as with the AFN may be moreefficient in the hands of young residents with good hands-ontraining. Indeed, a more simple fracture region such as the femoralshaft can be expected to be more tolerant for such an approach, incontrast to more sophisticated fractures, e.g. comminutedintraarticular fractures of the proximal or distal tibia.

Concerning objective outcomes, the overall infection ratecompared well to current literature,34,47 and so did the nonunionand malunion rates, which were similar between EU and SA.37,43

Surprisingly, even though SA fractures were significantly moreoften open and injuries more complex, local infection andnonunion rates were not higher than in EU centres. Due to thelimited statistical power, we cannot comment whether the rate ofSSIs was in effect even lower in SA. We observed a significantlyhigher rate of healing with reamed nailing. Reaming was notassociated with a lower reoperation rate, but correlated with a

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395394

lower rate of local complications. In our study, a large number offractures were fixed unreamed. Interestingly, nonreaming was notcorrelated with open fracture treatment, an option favoured bymany surgeons, especially in EU as to avoid additional surgicaldamage for fracture healing. In contrast, we observed a tendency ofmore reaming in open fractures both in SA and EU centres. Thequestion of whether to ream or not to ream is controversial.Clinical studies have shown varying results in healing rates with ahigher level of evidence arguing for better results followingreamed nailing.6,8–10,31,40

Despite the comparable healing rates, SA patients had asignificantly lower complication and reoperation rate comparedto EU patients. One possible reason for the lower percentage ofsystemic complications in the SA group would be the younger ageand the healthier status of the patients. The reporting rate mightalso be lower due to the shorter hospital stay and higher loss tofollow-up rate in this centre. The trend of a higher localcomplication rate in EU centres was mostly due to a higher rateof breakage of interlocking screws and associated with a highernumber of reoperations. X-ray case analysis of the reoperationcases revealed that the breakage of distal interlocking screwscould be interpreted as a natural self-dynamisation of the fracturezone, in cases where intraoperative reduction could have beenbetter and/or earlier operative dynamisation would have avoidedthe breakage. In general, an open fracture situation makesintraoperative reduction easier. However, the expected highernonunion rate usually associated with open fractures14 was notobserved in the SA group. From our data, the decision not to reambefore fixation of the fracture with the AFN remained the majorexplanation for a higher reoperation rate, even though theincidence of broken interlocking screws per se did not depend onwhether nailing was undertaken in a reamed or unreamedmanner. It could be argued, that the high SA caseload and itspossible effect on surgeons’ experience may be responsible forbetter outcomes in SA patients. However, statistically, thiscaseload effect could only be partially supported by our datadue to the limited number of EU centres with high caseloads. The3 EU centres with the highest caseloads showed a trend of lowerrate of interlocking screw breakage or reoperations compared tothe other EU centres. Still, about 4 times as many patients had tobe reoperated in these 3 EU centres compared to the SA centre,although this difference was not statistically significant.

Concerning subjective outcomes, no relevant differences wereobserved between the 2 study groups. Both nonunion andcomplication rates correlated with subjective outcomes (i.e. SF-36physical dimension and pain). Having a complication was associatedwith a worse subjective outcome in both SF-36 and pain scores. Inboth study groups, the mean physical component of the SF-36 1 yearafter the accident did not fully restore to baseline (=preinjury) valuesbut compared well with results reported from other groups.16,39 Ithas been reported that sociocultural background can influencepatient-assessed outcomes. People in LMIC are adapted to havinglittle resources and have lower expectations from their healthcaresystems compared to people in HIC.12 From our data, we can onlyspeculate that SA patients may have lower expectations than EUpatients as reflected in the trend of higher SF-36 physical but lowermental component scores at baseline and at 1 year in SA patients.

This study has several limitations. Overall follow-up rate of 73%is comparable to literature,33,36–38,43 but SA follow-up rate (2/3)was lower than in EU (4/5). On one hand, poor follow-up rates canhinder objective assessments from a statistical point of view andmay be a sign of low data quality.28,44 The investigation wasmonitored by an independent organisation using a detailed,uniform protocol and a central control for all participatinghospitals, but could not prevent missing data in the SA centre.On the other hand, high loss to follow-up appears to be a common

problem particularly in LMIC2,17,47 and has been attributed to thepatients’ sociocultural beliefs, the distance of patients from theregional health services, lack of resources, and the cost oftreatment.2 In these environments, patients rarely return forfollow-up as soon as they are physically fit resulting in follow-uprates of only about 30%. Even large nailing databases with morethan 18,000 interventions from over 140 hospitals in 49 countrieshave been unable to perform a comprehensive analysis of union orcomplication rates.47 Despite the comparable size of the studygroups within the observed time period, EU patients wererecruited from 11 hospitals, whereas SA patients only from 1centre, a fact that can be interpreted as another limitation of thisstudy. We have no further information about hospital cases thatcould have been recruited for this study but in effect underwentanother procedure. Nevertheless, the observed difference may wellreflect the actual intercontinental difference in trauma incidenceand caseload per centre.

In conclusion, our investigation demonstrated some majordifferences in the characteristics and the treatment of femoralshaft fractures between trauma centres in EU and SA, despite thestandardised use of an identical implant. Any intercontinentalcomparison or study design has to adequately consider theaforementioned confounding factors in order to prevent prematureinterpretation of raw data.

Conflicts of interest

The authors report on no conflict of interest.

Acknowledgements

The authors thank Ms. R. Billiones for writing assistance withthe manuscript. The execution of the study was partially financedby the Arbeitsgemeinschaft fur Osteosynthese – Clinical Investiga-tion and Documentation (AO-CID), Davos, Switzerland and bySynthes, Oberdorf, Switzerland. This support at no time had anyinfluence on the analysis undertaken or the results reported.

Appendix A

Centres involved in the study:

1. Charite University Medical Center, Berlin, Germany.2. Chris Hani Baragwanath Hospital, Johannesburg, South Africa.3. Maggiore Hospital, Bologna, Italy.4. Municipal Clinical Hospital Braunschweig, Braunschweig,

Germany.5. Municipal Clinical Hospital St. Georg Leipzig, Leipzig, Germany.6. Offenbach Clinical Hospital, Offenbach, Germany.7. S. Maria Misericordia Hospital Udine, Udine, Italy.8. University Hospital Basel, Basel, Switzerland.9. University Hospital Geneva, Geneva, Switzerland.

10. University of Graz, Graz, Austria.11. University of Innsbruck, Innsbruck, Austria.12. University Hospital Zurich, Zurich, Switzerland.

References

1. Aiyer S, Jagiasi J, Argekar H, et al. Closed antegrade interlocked nailing offemoral shaft fractures operated up to 2 weeks postinjury in the absence ofa fracture table or C-arm. J Trauma 2006;61:457–60.

2. Alatise OI, Lawal OO, Adesunkanmi AK, et al. Surgical outcome of abdomino-perineal resection for low rectal cancer in a Nigerian tertiary institution. World JSurg 2009;33:233–9.

3. Barclay G, Tavares C. International comparisons of criminal justice statistics2000. Issue 05/02. UK home office. Available at http://www.homeoffice.gov.uk/rds/pdfs2/hosb502.pdf; 12 July 2002 [accessed 16 May 2009].

T. Gross et al. / Injury, Int. J. Care Injured 41 (2010) 388–395 395

4. Bergen G, Chen LH, Warner M, et al. Injury in the United States 2007 Chartbook.National Center for Health Statistics. Centers for Disease Development andPrevention; 2008.

5. Brysiewicz P. Trauma in South Africa. Int J Trauma Nurs 2001;7:129–32.6. Canadian Orthopaedic Trauma Society. Nonunion following intramedullary

nailing of the femur with and without reaming. Results of a multicenterrandomized clinical trial. J Bone Joint Surg Am 2003;85-A:2093–6.

7. Chalmers A, Nandakumar A. A snapshot survey of the emergency theatre caseloadat Queens Medical Centre, Nottingham, United Kingdom. 4th World Congress ofAnaesthesiologists 2–7 March 2008, Cape Town, South Africa. Available at http://www.wca2008.com/images/abstractWCA/Trauma.pdf [accessed 10 February2009].

8. Clatworthy MG, Clark DI, Gray DH, et al. Reamed versus unreamed femoralnails. A randomised, prospective trial. J Bone Joint Surg Br 1998;80:485–9.

9. El Moumni M, Leenhouts PA, Ten Duis HJ, et al. The incidence of non-unionfollowing unreamed intramedullary nailing of femoral shaft fractures. Injury2008;40:205–8.

10. Forster MC, Aster AS, Ahmed S. Reaming during anterograde femoral nailing: isit worth it? Injury 2005;36:445–9.

11. Goosen J, Bowley DM, Degiannis E, et al. Trauma care systems in South Africa.Injury 2003;34:704–8.

12. Graham C. Happiness and health: lessons—and questions—for public policy.Health Aff (Millwood) 2008;27:72–87.

13. Gustilo RB, Anderson JT. Prevention of infection in the treatment of onethousand and twenty-five open fractures of long bones: retrospective andprospective analyses. J Bone Joint Surg Am 1976;58:453–8.

14. Gustilo RB, Gruninger RP, Davis T. Classification of type III (severe) openfractures relative to treatment and results. Orthopedics 1987;10:1781–8.

15. Haas N. The trauma centre: now and in the future. J Bone Joint Surg Br2002;84:627–30.

16. Helmy N, Jando VT, Lu T, et al. Muscle function and functional outcomefollowing standard antegrade reamed intramedullary nailing of isolatedfemoral shaft fractures. J Orthop Trauma 2008;22:10–5.

17. Hill AG, Perakath B, Bissett IP. The management of rectal cancer in a resourcepoor environment—a review. Int J Surg 2006;4:127–30.

19. Krug EG, Dahlberg LL, Mercy JA, et al. World report on health and violence.Geneva: World Health Organization; 2002.

20. Krug EG, Powell KE, Dahlberg LL. Firearm-related deaths in the United Statesand 35 other high- and upper-middle income countries. Int J Epidemiol1998;27:214–21.

21. Mansson A, Al Amin M, Malmstrom PU, et al. Patient-assessed outcomes inSwedish and Egyptian men undergoing radical cystectomy and orthotopicbladder substitution—a prospective comparative study. Urology 2007;70:1086–90.

22. McLaren AC, Roth JH, Wright C. Intramedullary rod fixation of femoral shaftfractures: comparison of open and closed insertion techniques. Can J Surg1990;33:286–90.

23. Messmer P, Gross T. Intramedullary nailing with the Antegrade Femoral Nail(AFN). AO International teaching Video 40097. Available at http://www.aovideo.ch/published/player.aspx?id=40097eem0398; 2005.

24. Mock C, Nguyen S, Quansah R, et al. Evaluation of trauma care capabilities infour countries using the WHO-IATSIC Guidelines for Essential Trauma Care.World J Surg 2006;30:946–56.

26. Noordin S, Wright JG, Howard AW. Global relevance of literature on trauma.Clin Orthop Relat Res 2008;466:2422–7.

27. Nordberg E. Surgical operations in eastern Africa: a review with conclusionsregarding the need for further research. East Afr Med J 1990;67:1–28.

28. Norman R, Matzopoulos R, Groenewald P, et al. The high burden of injuries inSouth Africa. Bull World Health Organ 2007;85:695–702.

29. Ozgediz D, Jamison D, Cherian M, et al. The burden of surgical conditions andaccess to surgical care in low- and middle-income countries. Bull World HealthOrgan 2008;86:646–7.

30. Padhi NR, Padhi P. Use of external fixators for open tibial injuries in the ruralthird world: panacea of the poor? Injury 2007;38:150–9.

31. Reynders PA, Broos PL. Healing of closed femoral shaft fractures treated withthe AO unreamed femoral nail. A comparative study with the AO reamedfemoral nail. Injury 2000;31:367–71.

33. Ricci WM, Devinney S, Haidukewych G, et al. Trochanteric nail insertionfor the treatment of femoral shaft fractures. J Orthop Trauma 2005;19:511–7.

34. Ricci WM, Schwappach J, Tucker M, et al. Trochanteric versus piriformis entryportal for the treatment of femoral shaft fractures. J Orthop Trauma 2006;20:663–7.

35. Roudsari BS, Nathens AB, Arreola-Risa C, et al. Emergency Medical Service(EMS) systems in developed and developing countries. Injury 2007;38:1001–13.

36. Salem KH, Maier D, Keppler P, et al. Limb malalignment and functional outcomeafter antegrade versus retrograde intramedullary nailing in distal femoralfractures. J Trauma 2006;61:375–81.

37. Sanders DW, MacLeod M, Charyk-Stewart T, et al. Functional outcome andpersistent disability after isolated fracture of the femur. Can J Surg 2008;51:366–70.

38. Starr AJ, Hay MT, Reinert CM, et al. Cephalomedullary nails in the treatment ofhigh-energy proximal femur fractures in young patients: a prospective, ran-domized comparison of trochanteric versus piriformis fossa entry portal.JOrthop Trauma 2006;20:240–6.

39. Thomson AB, Driver R, Kregor PJ, et al. Long-term functional outcomes afterintra-articular distal femur fractures: ORIF versus retrograde intramedullarynailing. Orthopedics 2008;31:748–50.

40. Tornetta P, Tiburzi D. Reamed versus nonreamed anterograde femoral nailing.J Orthop Trauma 2000;14:15–9.

41. Trunkey DD. The medical world is flat too. World J Surg 2008;32:1583–604.42. Tscherne H, Oestern HJ. Eine neue Klassifikation des Weichteilschadens bei

offenen und geschlossenen Frakturen. Unfallheilkunde 1982;85:111–5.43. Tucker MC, Schwappach JR, Leighton RK, et al. Results of femoral intrame-

dullary nailing in patients who are obese versus those who are not obese:a prospective multicenter comparison study. J Orthop Trauma 2007;21:523–9.

44. Udosen AM, Ogbudu S. The use of external fixators: a review of literature andexperiences in a developing world. Niger J Med 2006;15:115–8.

45. Uranues S, Lamont E. Acute care surgery: the European model. World J Surg2008;321605–12.

46. World Health Organization. Injuries and violence in Europe: why they matterand what can be done. Geneva: WHO; 2005.

47. Zirkle Jr LG. Injuries in developing countries—how can we help? The role oforthopaedic surgeons. Clin Orthop Relat Res 2008;466:2443–50.