Classification of trochanteric fracture of the proximal femur: a study of the reliability of current systems

Classification of trochanteric fracture of the proximal femur: a study of the reliability of current systems

Injury, Int. J. Care Injured 33 (2002) 713–715 Classification of trochanteric fracture of the proximal femur: a study of the reliability of current s...

61KB Sizes 0 Downloads 52 Views

Injury, Int. J. Care Injured 33 (2002) 713–715

Classification of trochanteric fracture of the proximal femur: a study of the reliability of current systems Humayon Pervez, Martyn J. Parker∗ , Glyn A. Pryor, Lennel Lutchman, Nishan Chirodian Orthopaedic Department, Peterborough District Hospital, Thorpe Road, Peterborough PE3 6DA, UK Accepted 8 April 2002

Abstract Five observers using the Jensen modification of the Evans classification and the AO classification (with and without subgroups) classified the radiographs of 88 trochanteric hip fractures. Each observer classified the radiographs independently on two occasions 3 months apart. Kappa statistical analysis was used for determination of intra- and inter-observer variation. For the Jensen classification, the mean kappa value was 0.52 (range: 0.44–0.60) for intra-observer variation and 0.34 (range: 0.17–0.38) for inter-observer variation. For the AO system with subgroups, the mean kappa value was 0.42 (range: 0.20–0.65) for intra-observer variation and 0.33 (range: 0.14–0.48) for inter-observer variation. For the AO classification system without subgroups, the mean kappa value was 0.71 (range: 0.60–0.81) for intra-observer variation and 0.62 (range: 0.50–0.71) for inter-observer variation. We recommend classifying trochanteric fractures into three groups as that of the AO system without the subgroups. For ease of use, these three groups may be termed stable trochanteric, unstable trochanteric and trans-trochanteric. Neither the Jensen classification nor the AO classification with subgroups is an acceptable classification system for trochanteric hip fractures. © 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction Hip fractures are primarily classified into intra- and extra-capsular according to their relationship to the capsular attachment of the hip. Extra-capsular fractures comprise about half of all the hip fractures and may be subdivided into those in the trochanteric region (also termed per- or inter-trochanteric) and those below the level of the lesser trochanter (subtrochanteric) [1]. Various classification systems have been used to classify extra-capsular hip fractures. The most frequently used system is the Jensen and Michaelsen’s modification of Evans classification [2–4]. More recently, the AO classification system has been advocated [5,6]. The ideal classification system should be useful in planning treatment and predicting the outcome. It should be appropriate for clinical practice, audit and research. An ideal classification system should produce the same classification each time the same patient data is reviewed by one practitioner (intra-observer reliability or repeatability), and different practitioners should agree on the classification of the data for each particular patient (inter-observer reliability). The ∗ Corresponding author. Tel.: +44-1733-874193; fax: +44-1733-875013. E-mail address: [email protected] (M.J. Parker).

aim of this study was to determine intra- and inter-observer reliability of current fracture classification methods used.

2. Patients and methods The anterior–posterior and lateral radiographs of 88 patients were selected from the database of the Peterborough Hip Fracture Project to give a range of extra-capsular fracture types. All fractures were classified by five observers (one consultant, specialist hip fracture surgeon, two specialist registrars and one SHO), using the Jensen modification of Evans classification (Fig. 1), and AO classification system for extra-capsular fractures (Fig. 2). Each observer independently classified these fractures on two occasions 3 months apart. They were not allowed to see how the fractures were treated or to discuss their observations with other investigators. Each observer had reference to Jensen and AO classification systems. After 3 months of the initial assessment, the observers re-classified the radiographs presented to them in a different order. They submitted their classifications on both occasions without keeping a record and these were held in sealed envelopes until the end of the study. Eight sets of radiographs could not be classified for the second time as these had been required for clinical use and could not be obtained, leaving 80 sets of radiographs for re-classification.

0020-1383/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 0 - 1 3 8 3 ( 0 2 ) 0 0 0 8 9 - X

714

H. Pervez et al. / Injury, Int. J. Care Injured 33 (2002) 713–715

Kappa statistical analysis was used for determination of intra- and inter-observer variation. The kappa test was introduced by Cohen to determine the level of reliability. Altman interpreted this as ‘the chance-corrected proportional agreement’ [7]. Kappa is a coefficient of agreement which has a value varying from +1, representing perfect agreement, through 0 representing an agreement no better chance than −1, absolute disagreement. There is no precise definition of the level of agreement which is acceptable but the guidelines of Svanholm et al. [8] are that values of 0.75 or more representing excellent agreement, 0.75–0.5 is good, and 0.5 or less is poor agreement and reliability.

3. Results

Fig. 1. The Jensen modification of the Evans classification.

For the Jensen system, the mean kappa values were 0.34 (range: 0.17–0.38) for inter-observer variation and 0.52 (range: 0.44–0.60) for intra-observer variation (Table 1). For the AO classification with subgroups, the mean kappa values were 0.33 (range: 0.14–0.48) for inter-observer variation and 0.42 (range: 0.20–0.65) for intra-observer variation (Table 2). For the classification of fractures using the AO system into three groups (A1, A2 and A3), but not using the subgroups, mean kappa values were 0.62 (range: 0.50–0.71) for inter-observer variation and 0.71 (range: 0.60–0.81) for intra-observer reliability (Table 3). Table 1 Inter- and intra-observer kappa values for the Jensen classification Observer

MJP

GAP

LL

HP

NC

Intra-observer

MJP GAP LL HP NC

– 0.37 0.38 0.35 0.38

– – 0.34 0.17 0.41

– – – 0.35 0.33

– – – – 0.35

– – – – –

0.56 0.60 0.48 0.52 0.44

Table 2 Inter- and intra-observer kappa values for the AO classification with subgroups Observer

MJP

GAP

LL

HP

NC

Intra-observer

MJP GAP LL HP NC

– 0.32 0.36 0.14 0.34

– – 0.48 0.25 0.43

– – – 0.35 0.33

– – – – 0.31

– – – – –

0.65 0.48 0.45 0.20 0.33

Table 3 Inter- and intra-observer kappa values for the basic AO classification

Fig. 2. The AO classification of trochanteric fractures.

Observer

MJP

GAP

LL

HP

NC

Intra-observer

MJP GAP LL HP NC

– 0.58 0.63 0.50 0.59

– – 0.57 0.62 0.68

– – – 0.68 0.71

– – – – 0.68

– – – – –

0.78 0.71 0.81 0.60 0.69

H. Pervez et al. / Injury, Int. J. Care Injured 33 (2002) 713–715

4. Discussion Any fracture classification system should be reproducible among different observers as well as by the same observer on separate occasions. In addition, the classification should suggest treatment and/or outcome [9]. The kappa statistic provides a method of determining the diagnostic accuracy of a classification system with the calculation adjusting for random matches. For a hip fracture, the anatomical configuration, and therefore its classification generally determines the management. Previous studies have assessed the reliability of either the AO or Jensen classification methods of trochanteric fractures. Schipper et al. [10] studied the AO classification system of trochanteric fractures for 20 X-rays reviewed by 15 observers. They reported a mean intra-observer kappa value of 0.48 and inter-observer values of 0.33 and 0.34 for the classification with subgroups. For the AO system without subgroups, kappa values were 0.78 for intra-observer and 0.67 and 0.63 for inter-observer. An earlier study of five trochanteric hip fractures also found the AO classification to be unreliable [11]. Newey et al. [12] found the AO classification system to be unnecessarily complicated and fell short of playing a useful role in planning of treatment. For the Evans classification intra-observer kappa values of 0.66–0.92 and 0.69–0.81 have been reported. Values for inter-observer variation were 0.41–0.68 and 0.41–0.77 [13,14]. The results of this study confirm the unacceptability of both the full AO and Jensen classification systems. This study does indicate that the simpler classification of trochanteric fracture into three AO groups (31A1, 31A2, 31A3) is acceptable. For those who find the numerical terminology confusing, we recommend that these three groups be may be termed as stable trochanteric, unstable trochanteric and trans-trochanteric. The confusing and frequently misused terms of per- and inter-trochanteric are thereby avoided. AO group A1 fractures are two part trochanteric fractures, which may be displaced or undisplaced and are equivalent to Jensen classification types 1 and 2. AO group A2 fractures are comminuted and unstable, and equivalent to Jensen types 3, 4 and 5. AO group A3 fractures are at the level of the lesser trochanter and may be transverse, oblique or reversed. For reversed fracture lines, the fracture line runs distally in a medial to lateral direction. This group of fractures were originally separated in the Evans classification as type 2 fractures, but were incorporated within the other groups in

715

the Jensen modification. Clinical studies have indicated a marked increase in risk of fixation failure for these fracture patterns [15] and intra-medullary fixation of these fractures has been suggested. Therefore, a separate classification is appropriate. The term inter-trochanteric has also been used to describe these fractures [16]; however, this term has since been used for trochanteric fractures, therefore an alternative term is necessary for which the term trans-trochanteric is suggested. This term therefore applies to fractures in which the fracture line traversing the femur is predominately found within the area of femur between the upper and lower borders of the lesser trochanter.

References [1] Pynsent PB, Fairbank JCT, Carr AJ. Classification of musculoskeletal trauma. Oxford: Butterworth, 1999. [2] Evans EM. The treatment of trochanteric fractures of the femur. J Bone Joint Surg Br 1949;31(B):190–203. [3] Jensen JS. Classification of trochanteric fractures. Acta Orthop Scand 1980;51:803–10. [4] Jensen JS, Michaelsen M. Trochanteric femoral fractures treated with McLaughlin osteosynthesis. Acta Orthop Scand 1975;46:795–803. [5] Colton CL. Telling the bones. J Bone Joint Surg Br 1991;73(B):362. [6] Muller ME, Nazarian S, Koch P, et al. The comprehensive classification of fractures of long bones. Berlin: Springer, 1990. [7] Altman DG. Practical statistics for medical research. 3rd ed. Chapman and Hall, 1995, pp. 403–9. [8] Svanholm H, Starklint H, Gundersen HJG, et al. Reproducibility of histomorphologic diagnosis with special reference to kappa statistic. APMIS 1989;97:689–98. [9] Burstein AH. Fracture classification systems: do they work and are they useful? J Bone Joint Surg Am 1993;75(A):1743–4 [editorial]. [10] Schipper IB, Steyerberg EW, Castelein RM, vanVugt AB. Reliability of the AO/ASIF classification for pertrochanteric femoral fractures. Acta Orthop Scand 2001;72:36–41. [11] De Boeck H. Classification of hip fractures. Acta Orthop Belg 1994;60(Suppl 1):106–9. [12] Newey ML, Ricketts D, Roberts L. The AO classification of long bone fractures: an early study of its use in clinical practice. Injury 1993;24:309–12. [13] Andersen E, Jorgensen LG, Hededam LT. Evans’ classification of trochanteric fractures: an assessment of the interobserver and intraobserver reliability. Injury 1990;21:377–8. [14] Gerchen PM, Nielsen JO, Olesen B. Poor reproducibility of Evans’ classification of trochanteric fracture; assessment of 4 observers in 52 cases. Acta Orthop Scand 1993;64:71–81. [15] Haidukewych GJ, Israel TA, Berry DJ. Reverse obliquity fractures of the intertrochanteric region of the femur. J Bone Joint Surg Am 2001;83(A):643–50. [16] Moore JR. Fractures of the upper end of the femur including fracture dislocations at the hip joint. Am J Surg 1939;44:117–34.