- Email: [email protected]
Distal unicondylar fractures of the proximal phalanx
Distal unicondylar fractures of the proximal phalanx The records of 38 consecutive patients (38 fractures) who underwent treatment for distal unicondylar fractures of the proximai phalanx were reviewed to evaluate fracture characteristics, mechanism of injury, treatment options, and functional outcomes. Four classes of fracture pattern were defined radiographically. Most fractures occurred during bag sports and involved an axial splitting of extended digits, with the condyle closest to the midline of the hand fracturing most commonly. We believed that the fracture occurred as a result of tension loading due to a distraction force from the collateral ligament. All fractures healed. Follow-up examination averaged 3 years. Five of seven nondisplaced fractures treated with splinting and four of ten displaced fractures treated with reduction and single Kiischner wire fixation displaced. Fractures treated with multiple Kirschner wire fixation had the best final joint motion. Class IV fractures with a small pahnar coronal fragment had the poorest final motion. A short period of post-operative immobilization did not adversely affect final proximal interphahmgeal joint motion. We recommend multiple Kirschner wire or miniscrew iixation of these fractures as the most predictable method of treatment. Final proximal htterphahmgeal joint motion is not uniformly excellent in patients with these fractures. (J IIm Suac 1993;18A:594-9.)
Arnold-Peter C. Weiss, MD, Providence, Indianapolis, hf.
I
ntra-articular fractures of the proximal interphalangeal (PIP) joint often present significant difficulties in appropriate technical treatment and the attainment of a functional arc of motion after treatment has ended.’ The difficulty of successfully treating these small fractures is due to frequent comminution and the technical limitations of dealing with small articular fracture fragments. This group of fractures comprises specific subsets that have received little attention in the literature. Of these injuries, distal unicondylar fractures of the proximal phalanx are common in young, active patients and define a specific fracture pattern. Various
From the Department of Orthopaedics, Brown University School of Medicine, Rhode Island Hospital, Providence, RI., and the Indiana Hand Center, Indianapolis, Ind. Received for publication Sept. 23, 1992.
April 24, 1992; accepted
in revised form
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Reprint requests: Arnold-Peter C. Weiss, MD, Department of Orthopaedics, Brown University School of Medicine, Rhode Island Hospital, 593 Eddy St., Providence, RI 02903. 3iIl4399tl
594
THEJOURNAL
OF HAh’D SURGERY
R .I., and Hill Hastings II, MD,
treatment protocols for these fractures have been advocated; these include simple conservative splint treatment, closed reduction augmented by percutaneous Kirschner wire fixation, or miniscrews for fracture fixation.‘” We examined this particular subset of patients in an attempt to define fracture morphology and its mechanism of propagation, various treatment options and the overall functional outcome of the treatment of distal unicondylar proximal phalanx fractures. Materials
and methods
From 1981 to 1989, 38 consecutive cases (38 patients) of distal unicondylar proximal phalanx fractures were treated at the Indiana Hand Center and the Indiana University Medical Center. The patients had an average age of 24 years (range, 5 to 65); there were 28 male and 10 female patients. On questioning of patients, five main mechanisms of injury were noted: ball sports in 19 cases, torsional injury in four, direct blow/crush in 6, falls in 5, and fighting in 4. Standard anteroposterior, lateral, and oblique x-ray views of the involved proximal phalanx were taken to evaluate the fracture pattern and the displacement. After careful review of the unicondylar fractures in our series, an x-ray classification system of four
Vol. HA, No. 4 July 1993
Distal unicondylar fractures of proximal phalanx
CLASS I (Obllque Voler) N = 22
CLASS II (Long Segittal) N=8
CLASS Ill (Dorsal Coronal) N=4
595
CLASS IV (Volar Coronal) N=4
Fig. 1. Diagram demonstrating the fracture pattern noted in each of the four classes determined by radiographic examination and the number of clinical cases in each group represented in this study.
Fig. 2. X-ray films demonstrating fracture pattern orientation in class I (A), class II (B), class III (C), arid class IV (D).
main fracture patterns was developed (Fig. 1). Class I fractures (22 cases) involved an oblique pahnar pattern, with the fracture line extending from the proximal metaphyseal flare of the distal aspect of the proximal phalanx to the intracondylar depression at the distal artic-
ular surface of the proximal phalanx (Fig. 2, A). The plane of the fracture resides in neither the sagittal nor the coronal planes. The distal fracture fragment lies palmar to the proximal phalangeal shaft. Class II fractures (eight cases) involved a long oblique fracture line,
The Journal of 596
HAND SURGERY
Weiss and Hastings
h
s
ai L
T
I
II
100
T
III
MS
IV
fracture noted on preoperative x-ray films. Class IV fractures (palmar coronal) had the poorest final motion.
with the plane of the fracture sagittally oriented (Fig. 2, B). The last two fracture patterns represent similar injuries with opposite planar orientation. Class III fractures (four cases) involved a dorsal coronal fragment that was generally small in size (Fig. 2, C). Class IV (four cases) involved a palmar coronal fragment that, similarly, was small in size (Fig. 2, D). The condyle fractured (radial vs. ulnar) was noted in each digit. Five cases involved the thumb (two radial; three ulnar); five cases involved the index finger (one radial; four ulnar); three cases involved the middle finger (two radial; one ulnar); seven involved the ring finger (four radial; three ulnar); and 18 cases involved the proximal phalanx of the small finger (thirteen radial; five ulnar). Treatment distribution. Thirty-six fractures underwent operative reduction (28 open and 8 closed). In seven cases, the fractures were initially thought to be nondisplaced and were treated with splint immobilization only. With close follow-up, five of these fractures demonstrated displacement of 1 to 3 mm at the articular surface on radiographic evaluation and required surgical treatment. Fracture fixation methods involved miniscrew only in 10, miniscrews with Kirschner wire supplementation in four, single Kirschner wire fixation in 10 (six open, four closed), multiple Kirschner wires in 11 (seven open, four closed), and loop wire fixation in one case. In all patients a palmar plaster splint was placed postoperatively, and treatment included early active range-of-motion exercises with initiation dictated by the stability of fracture fixation as
SK
MK
Type of Fixation
Class of Fracture
Fig. 3. Final active range of motion compared with class of
MS & K
Fig. 4. Final active range of motion compared with type of fracture fixation used. Single Kirschner wire fixation of unicondylar fractures resulted in the poorest final PIP joint motion. MS, Miniscrew; MS&K, miniscrew and Kirschner wire; SK, single Kirschner wire; MK, multiple Kirschner wire.
judged by the surgeon intraopcratively. Active rangeof-motion exercises were begun an average of 15 days postoperatively (range, 1 to 28 days). Passive rangeof-motion exercises were initiated after the operating surgeon thought that sufficient fracture healing had occurred for subsequent displacement to be unlikely. If Kirschner wires were used for fracture fixation, these were left in place as long as could be appropriately tolerated or until fracture healing was complete.
Results All 38 patients were available for follow-up examination an average of 3 years (range, 1 to 8 years) after the initial procedure. At follow-up, the range of active motion at the PIP joint was measured. The average extension was 13 degrees (range, 0 to 35 degrees), and average flexion was 85 degrees (range, 60 to 115 degrees) in the index through the small fingers. In the thumb, the average extension was 16 degrees of hyperextension (range, 10 to 25 degrees) and an average flexion of 51 degrees (range, 30 to 70 degrees). At final follow-up, all fractures were found to have healed. Statistical analysis using one-way analysis of variance (ANOVA) corrected for mean data was performed on each group of data. Final active PIP range of motion, when correlated with class of fracture, was 71 2 16 degrees ( 2 standard deviation) in class I, 74 + 21 degrees in class II,
Vol. 18A, No. 4 July 1993
o-7
Distal unicondylar fractures of proximal phalanx
8 - 14
15 - 21
597
m
22 - 28
Type of Reduction (Kirschner wire fixation)
Interval to AROM (days)
Fig. 5. Interval from operative fracture fixation to onset of
Fig. 6. Diagram indicating that final outcome of PIP joint
active range of motion exercises in therapy when compared with final follow-up examination. PIP joint motion demonstrates a drop in final motion result if therapy is not initiated within 3 weeks postoperatively.
motion is not affected by open verstis closed reduction in those fractures fixed with Kirschner wires alone.
Table I. Distribution of unicondylar 75 k 19 degrees in class III, and 57 2 18 degrees in class IV (not statistically different among the group mean; p = 0.41) (Fig. 3). Final active range of motion correlated with type of fracture fixation was 79 + 8 degrees in fractures undergoing miniscrew fixation, 68 + 11 degrees in fractures treated with miniscrews and supplemental Kirschner wires, 61 -C 16 degrees in fractures treated with single Kirschner wire fixation, and 89 -+ 18 degrees in fractures treated with multiple Kirschner wires (statistical difference among the group means; p = 0.001) (Fig. 4). The interval between operative fixation and the initiation of active range-ofmotion exercises correlated with final active range of motion was 76 + 16 degrees in fractures where exercises were begun between 1 and 7 days postoperatively, 80 f 20 degrees in those begun between 8 and 14 days postoperatively, 84 -+ 16 degrees in those begun between 15 and 21 days postoperatively, and 71 -+ 19 degrees in those begun between 22 and 28 days postoperatively (not statistically different among the group means; p = 0.49) (Fig. 5). In cases in which fractures underwent Kirschner wire fixation (either single or multiple), comparison of final range of motion with whether or not the fracture required open reduction was 77 + 4 degrees in fractures undergoing open reduction and 79 + 5 degrees for those treated with closed reduction (no statistical difference; p > 0.05) (Fig. 6). The final average PIP active motion for the various age groups was as follows: under 15 years, 86 2 17 de-
Mechanisms of injury
fractures
II
SpOltS Torsional Crush Falls
16 3 1 1
0 1 4 2
2 0 1 0
1 0 0 2
19 4 6 5
Fight Total
-! 22
1 8
1 4
1 4
_A 38
grees; 15 to 30 years, 72 k 24 degrees; 30 to 45 years, 65 -+ 21 degrees; over 45 years, 66 + 23 degrees (not statistically different among the group means; p =
0.28). Class of fracture was compared with the mechanism of injury as detailed in Table I. Complications. Complications occurred in the treatment of four cases. All four cases underwent initial single Kirschner wire fracture fixation and were found on early follow-up to have subsequent displacement requiring a secondary procedure for reduction. In two cases miniscrews were used for appropriate fracture fixation, and in two cases crossed Kirschner wires were used to obtain rigid fracture fixation. Four patients with poor range of motion underwent a secondary operative procedure consisting of dorsal capsulotomies of the PIP joint and extensor tendon tenolyses.
5%
Weiss and Hastings
LAlERAL “F lo ~vmh RoUtlo”,
Fii. 7. Most common mechanism of injury hypothesized for each of the four classes of distal unicondylar fractures.
Discussion Several authors have reviewed fractures of the proximal interphalangeal joint and commented on those involving only one distal condyle. London’ presented a classification system that represented all of these fractures as having a sag&al plane.. He believed that some of these fractures were stable and required only splint treatment whereas others required operative fixation. McCue et al. 3demonstrated that distal unicondylar fractures were frequently associated with sporting activities. In a review of twenty cases, they found that neither the radial nor the ulnar condyle was involved more frequently as the fracture fragment. Operative fracture fixation, most commonly with two Kirschner wires, was performed in all cases with an average of 93 degrees of PIP joint motion noted at final follow-up examination. Soeui’ believed that distal unicondylar fractures of the proximal or middle phalanx were caused by a trochlear shear mechanism rather than by compression. Our study points to several aspects regarding treatment of these fractures and their probable mechanism of injury. They appear to be more common after sporting injuries in which a ball comes between two slightly flexed, outstretched digits with subsequent high-velocity spread of these digits. This cause of injury was seen most commonly in class I fractures with an oblique, multiaxis plane of fracture pattern. We believe that class I fractures involve a tension load that imparts a distraction force on the condyle through the collateral ligament with an element of rotation. Class II fractures involving a sag&al orientation were more frequently seen in injuries that involved a pure anteroposterior force mechanism with a resulting splitting phenomenon
and longer fracture lines or in pure lateral stress injuries with no digital flexion at the time of injury. Class III and IV fractures occurred with pure hyperflexion and hyperextension of the proximal inter-phalangeal joint. We believe that these two classes represent knock-off fractures caused by a transient dorsal or palmar subluxation of the PIP joint combined with a flexion/extension component (Fig. 7). The fractured condyle tends to be on the outermost digits of the hand, as would be expected in a tension load with an avulsion fracture of the condyle through the collateral ligaments. If compression played a large role in this type of fracture, the long finger would be expected to share a high portion of the injuries because it is the most prominent digit and would be more likely to be injured in ball sports; this is not the case. On analysis of the condyle and digit involved in these fractures, the condyle toward the midline of the hand (i.e., long finger) is the one that is fractured in an overwhelming majority of cases. No explanation other than that the fracture occurs through a tension mechanism with a distraction force can support this observation. Although the overall distribution of radial versus ulnar condyle fractured is relatively even, as pointed out by McCue et al., the observation that the condyle fractured is more frequently toward the midline of the hand demonstrates the mechanism of injury in the majority of cases. In cases in which the condyle away from the midline of the hand is fractured, either a compression mechanism (with the finger deviating away from the midline) or a tension mechanism (with the finger deviating toward the midline) might be involved, although the former is more likely.
Vol. 18A, No. 4 July 1993
The functional outcome of these fractures is affected by treatment methods, although to a lesser extent than expected. Class IV fractures result in the poorest final range of motion (although not statistically different from the other classes), most likely because of a bony block or increased cam stop imparted during flexion by the palmar fragment or reiatively shortened collateral ligaments, respectively, if reduction is not perfect. The use of splint immobilization for nondisplaced fractures and single Kirschner wire fixation for displaced fractures had a high incidence of subsequent displacement and required careful follow-up examination with a high index of suspicion for fracture movement. Final range of motion was poorest in those fractures treated with a single Kirschner wire because of those that displaced. Although this group represented the simplest fracture pattern without comrninution, instability at the fracture site represented an important component that is probably best addressed by multiple fixation sites (i.e., two or more Kirschner wires and/or screws). The palmar oblique fracture plane of class I injuries carries significant fixation implications. Kirschner wires that enter from the fractured condylar fragment in the central or dorsal portion will obtain purchase on only a very thin, weak cortical portion of the fragment and will be susceptible to rotation or displacement. The best fixation area in these fracture fragments is the distal palmar aspect. One should attempt to direct Kirschner wire fixation from the contralateral dorsal side of the shaft with the intact condyle into the palmar aspect of the fractured condyle. Occasionally fracture line orientation makes this approach difficult, and in these cases it is best to go from fracture fragment to intact condyle, thus avoiding numerous passes. At least one other Kirschner wire should be placed for control of alignment. With the use of a miniscrew, the same principle should be applied to maximize cortical purchase. In the placement of Kirschner wires or miniscrews, care should be taken to avoid transfixing the collateral ligaments, which might limit their excusion postoperatively. The area just proximal to the collateral Iigament origin is ideal for hardware placement.
Distal unicandylar fractures of proximal phalanx
599
In fractures treated with Kirschner wires alone, whether reduction was open or closed did not appear to influence final motion. With presumably less postoperative scarring, those treated with closed reduction should have better final motion, but this assumption may be neutralized by a more stable, exact reduction under open observation allowing earlier motion. One issue raised by this study is what factors involved in a lateral bending proximal interphalangeal injury dictate a unicondylar fracture versus a collateral ligament tear. A biomechanical study by Kiefhaber et al.’ of the lateral stability of the PIP joint to lateral stress, which loaded 78 joints in a tension fashion, demonstrated only one unicondylar fracture, with all the other injuries being ligamentous. The loading rate in this study was very low by clinical standards (60 to 90 degrees per second). A high loading rate, as would be seen in a ball sport injury, might predispose to fracture rather than ligament tear. A similar clinical situation occurs in the distal femoral unicondylar fracture, most frequently seen in pedestrian/motor vehicle accidents, in which a unicondylar fracture occurs from lateral bending/ tension failure of the bone by the collateral ligament as the car bumper strikes the leg at a high loading rate.
REFERENCES 1. Hastings HH II, Carrol C IV: Treatment of closed articular
2. 3.
4.
5.
fractures of the metacarpophalangeal and proximal interphalangeal joints. Hand Clin 1988;4:503-27. London PS. Sprains and fractures involving the interphalangeal joints. Hand 1971;3:155-8. McCue FC, Honner R, Johnson MC Jr, Gieck JH. Athletic injuries of the proximal interphalangeal joint requiring surgical treatment. J Bone Joint Surg 1970;521\:937-55. Soeur R. Fractures of the limbs: the relationship between mechanism and treatment. Brussels: La Clinique Orthopedique, 1981542-3. Kiefhaber TR, Stem PJ, Grood ES. Lateral stability of the proximal interphdangeal joint. J HANDSURG 1986; 1 lA:661-9.
Arnold-Peter C. Weiss, MD, Providence, Indianapolis, hf.
I
ntra-articular fractures of the proximal interphalangeal (PIP) joint often present significant difficulties in appropriate technical treatment and the attainment of a functional arc of motion after treatment has ended.’ The difficulty of successfully treating these small fractures is due to frequent comminution and the technical limitations of dealing with small articular fracture fragments. This group of fractures comprises specific subsets that have received little attention in the literature. Of these injuries, distal unicondylar fractures of the proximal phalanx are common in young, active patients and define a specific fracture pattern. Various
From the Department of Orthopaedics, Brown University School of Medicine, Rhode Island Hospital, Providence, RI., and the Indiana Hand Center, Indianapolis, Ind. Received for publication Sept. 23, 1992.
April 24, 1992; accepted
in revised form
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Reprint requests: Arnold-Peter C. Weiss, MD, Department of Orthopaedics, Brown University School of Medicine, Rhode Island Hospital, 593 Eddy St., Providence, RI 02903. 3iIl4399tl
594
THEJOURNAL
OF HAh’D SURGERY
R .I., and Hill Hastings II, MD,
treatment protocols for these fractures have been advocated; these include simple conservative splint treatment, closed reduction augmented by percutaneous Kirschner wire fixation, or miniscrews for fracture fixation.‘” We examined this particular subset of patients in an attempt to define fracture morphology and its mechanism of propagation, various treatment options and the overall functional outcome of the treatment of distal unicondylar proximal phalanx fractures. Materials
and methods
From 1981 to 1989, 38 consecutive cases (38 patients) of distal unicondylar proximal phalanx fractures were treated at the Indiana Hand Center and the Indiana University Medical Center. The patients had an average age of 24 years (range, 5 to 65); there were 28 male and 10 female patients. On questioning of patients, five main mechanisms of injury were noted: ball sports in 19 cases, torsional injury in four, direct blow/crush in 6, falls in 5, and fighting in 4. Standard anteroposterior, lateral, and oblique x-ray views of the involved proximal phalanx were taken to evaluate the fracture pattern and the displacement. After careful review of the unicondylar fractures in our series, an x-ray classification system of four
Vol. HA, No. 4 July 1993
Distal unicondylar fractures of proximal phalanx
CLASS I (Obllque Voler) N = 22
CLASS II (Long Segittal) N=8
CLASS Ill (Dorsal Coronal) N=4
595
CLASS IV (Volar Coronal) N=4
Fig. 1. Diagram demonstrating the fracture pattern noted in each of the four classes determined by radiographic examination and the number of clinical cases in each group represented in this study.
Fig. 2. X-ray films demonstrating fracture pattern orientation in class I (A), class II (B), class III (C), arid class IV (D).
main fracture patterns was developed (Fig. 1). Class I fractures (22 cases) involved an oblique pahnar pattern, with the fracture line extending from the proximal metaphyseal flare of the distal aspect of the proximal phalanx to the intracondylar depression at the distal artic-
ular surface of the proximal phalanx (Fig. 2, A). The plane of the fracture resides in neither the sagittal nor the coronal planes. The distal fracture fragment lies palmar to the proximal phalangeal shaft. Class II fractures (eight cases) involved a long oblique fracture line,
The Journal of 596
HAND SURGERY
Weiss and Hastings
h
s
ai L
T
I
II
100
T
III
MS
IV
fracture noted on preoperative x-ray films. Class IV fractures (palmar coronal) had the poorest final motion.
with the plane of the fracture sagittally oriented (Fig. 2, B). The last two fracture patterns represent similar injuries with opposite planar orientation. Class III fractures (four cases) involved a dorsal coronal fragment that was generally small in size (Fig. 2, C). Class IV (four cases) involved a palmar coronal fragment that, similarly, was small in size (Fig. 2, D). The condyle fractured (radial vs. ulnar) was noted in each digit. Five cases involved the thumb (two radial; three ulnar); five cases involved the index finger (one radial; four ulnar); three cases involved the middle finger (two radial; one ulnar); seven involved the ring finger (four radial; three ulnar); and 18 cases involved the proximal phalanx of the small finger (thirteen radial; five ulnar). Treatment distribution. Thirty-six fractures underwent operative reduction (28 open and 8 closed). In seven cases, the fractures were initially thought to be nondisplaced and were treated with splint immobilization only. With close follow-up, five of these fractures demonstrated displacement of 1 to 3 mm at the articular surface on radiographic evaluation and required surgical treatment. Fracture fixation methods involved miniscrew only in 10, miniscrews with Kirschner wire supplementation in four, single Kirschner wire fixation in 10 (six open, four closed), multiple Kirschner wires in 11 (seven open, four closed), and loop wire fixation in one case. In all patients a palmar plaster splint was placed postoperatively, and treatment included early active range-of-motion exercises with initiation dictated by the stability of fracture fixation as
SK
MK
Type of Fixation
Class of Fracture
Fig. 3. Final active range of motion compared with class of
MS & K
Fig. 4. Final active range of motion compared with type of fracture fixation used. Single Kirschner wire fixation of unicondylar fractures resulted in the poorest final PIP joint motion. MS, Miniscrew; MS&K, miniscrew and Kirschner wire; SK, single Kirschner wire; MK, multiple Kirschner wire.
judged by the surgeon intraopcratively. Active rangeof-motion exercises were begun an average of 15 days postoperatively (range, 1 to 28 days). Passive rangeof-motion exercises were initiated after the operating surgeon thought that sufficient fracture healing had occurred for subsequent displacement to be unlikely. If Kirschner wires were used for fracture fixation, these were left in place as long as could be appropriately tolerated or until fracture healing was complete.
Results All 38 patients were available for follow-up examination an average of 3 years (range, 1 to 8 years) after the initial procedure. At follow-up, the range of active motion at the PIP joint was measured. The average extension was 13 degrees (range, 0 to 35 degrees), and average flexion was 85 degrees (range, 60 to 115 degrees) in the index through the small fingers. In the thumb, the average extension was 16 degrees of hyperextension (range, 10 to 25 degrees) and an average flexion of 51 degrees (range, 30 to 70 degrees). At final follow-up, all fractures were found to have healed. Statistical analysis using one-way analysis of variance (ANOVA) corrected for mean data was performed on each group of data. Final active PIP range of motion, when correlated with class of fracture, was 71 2 16 degrees ( 2 standard deviation) in class I, 74 + 21 degrees in class II,
Vol. 18A, No. 4 July 1993
o-7
Distal unicondylar fractures of proximal phalanx
8 - 14
15 - 21
597
m
22 - 28
Type of Reduction (Kirschner wire fixation)
Interval to AROM (days)
Fig. 5. Interval from operative fracture fixation to onset of
Fig. 6. Diagram indicating that final outcome of PIP joint
active range of motion exercises in therapy when compared with final follow-up examination. PIP joint motion demonstrates a drop in final motion result if therapy is not initiated within 3 weeks postoperatively.
motion is not affected by open verstis closed reduction in those fractures fixed with Kirschner wires alone.
Table I. Distribution of unicondylar 75 k 19 degrees in class III, and 57 2 18 degrees in class IV (not statistically different among the group mean; p = 0.41) (Fig. 3). Final active range of motion correlated with type of fracture fixation was 79 + 8 degrees in fractures undergoing miniscrew fixation, 68 + 11 degrees in fractures treated with miniscrews and supplemental Kirschner wires, 61 -C 16 degrees in fractures treated with single Kirschner wire fixation, and 89 -+ 18 degrees in fractures treated with multiple Kirschner wires (statistical difference among the group means; p = 0.001) (Fig. 4). The interval between operative fixation and the initiation of active range-ofmotion exercises correlated with final active range of motion was 76 + 16 degrees in fractures where exercises were begun between 1 and 7 days postoperatively, 80 f 20 degrees in those begun between 8 and 14 days postoperatively, 84 -+ 16 degrees in those begun between 15 and 21 days postoperatively, and 71 -+ 19 degrees in those begun between 22 and 28 days postoperatively (not statistically different among the group means; p = 0.49) (Fig. 5). In cases in which fractures underwent Kirschner wire fixation (either single or multiple), comparison of final range of motion with whether or not the fracture required open reduction was 77 + 4 degrees in fractures undergoing open reduction and 79 + 5 degrees for those treated with closed reduction (no statistical difference; p > 0.05) (Fig. 6). The final average PIP active motion for the various age groups was as follows: under 15 years, 86 2 17 de-
Mechanisms of injury
fractures
II
SpOltS Torsional Crush Falls
16 3 1 1
0 1 4 2
2 0 1 0
1 0 0 2
19 4 6 5
Fight Total
-! 22
1 8
1 4
1 4
_A 38
grees; 15 to 30 years, 72 k 24 degrees; 30 to 45 years, 65 -+ 21 degrees; over 45 years, 66 + 23 degrees (not statistically different among the group means; p =
0.28). Class of fracture was compared with the mechanism of injury as detailed in Table I. Complications. Complications occurred in the treatment of four cases. All four cases underwent initial single Kirschner wire fracture fixation and were found on early follow-up to have subsequent displacement requiring a secondary procedure for reduction. In two cases miniscrews were used for appropriate fracture fixation, and in two cases crossed Kirschner wires were used to obtain rigid fracture fixation. Four patients with poor range of motion underwent a secondary operative procedure consisting of dorsal capsulotomies of the PIP joint and extensor tendon tenolyses.
5%
Weiss and Hastings
LAlERAL “F lo ~vmh RoUtlo”,
Fii. 7. Most common mechanism of injury hypothesized for each of the four classes of distal unicondylar fractures.
Discussion Several authors have reviewed fractures of the proximal interphalangeal joint and commented on those involving only one distal condyle. London’ presented a classification system that represented all of these fractures as having a sag&al plane.. He believed that some of these fractures were stable and required only splint treatment whereas others required operative fixation. McCue et al. 3demonstrated that distal unicondylar fractures were frequently associated with sporting activities. In a review of twenty cases, they found that neither the radial nor the ulnar condyle was involved more frequently as the fracture fragment. Operative fracture fixation, most commonly with two Kirschner wires, was performed in all cases with an average of 93 degrees of PIP joint motion noted at final follow-up examination. Soeui’ believed that distal unicondylar fractures of the proximal or middle phalanx were caused by a trochlear shear mechanism rather than by compression. Our study points to several aspects regarding treatment of these fractures and their probable mechanism of injury. They appear to be more common after sporting injuries in which a ball comes between two slightly flexed, outstretched digits with subsequent high-velocity spread of these digits. This cause of injury was seen most commonly in class I fractures with an oblique, multiaxis plane of fracture pattern. We believe that class I fractures involve a tension load that imparts a distraction force on the condyle through the collateral ligament with an element of rotation. Class II fractures involving a sag&al orientation were more frequently seen in injuries that involved a pure anteroposterior force mechanism with a resulting splitting phenomenon
and longer fracture lines or in pure lateral stress injuries with no digital flexion at the time of injury. Class III and IV fractures occurred with pure hyperflexion and hyperextension of the proximal inter-phalangeal joint. We believe that these two classes represent knock-off fractures caused by a transient dorsal or palmar subluxation of the PIP joint combined with a flexion/extension component (Fig. 7). The fractured condyle tends to be on the outermost digits of the hand, as would be expected in a tension load with an avulsion fracture of the condyle through the collateral ligaments. If compression played a large role in this type of fracture, the long finger would be expected to share a high portion of the injuries because it is the most prominent digit and would be more likely to be injured in ball sports; this is not the case. On analysis of the condyle and digit involved in these fractures, the condyle toward the midline of the hand (i.e., long finger) is the one that is fractured in an overwhelming majority of cases. No explanation other than that the fracture occurs through a tension mechanism with a distraction force can support this observation. Although the overall distribution of radial versus ulnar condyle fractured is relatively even, as pointed out by McCue et al., the observation that the condyle fractured is more frequently toward the midline of the hand demonstrates the mechanism of injury in the majority of cases. In cases in which the condyle away from the midline of the hand is fractured, either a compression mechanism (with the finger deviating away from the midline) or a tension mechanism (with the finger deviating toward the midline) might be involved, although the former is more likely.
Vol. 18A, No. 4 July 1993
The functional outcome of these fractures is affected by treatment methods, although to a lesser extent than expected. Class IV fractures result in the poorest final range of motion (although not statistically different from the other classes), most likely because of a bony block or increased cam stop imparted during flexion by the palmar fragment or reiatively shortened collateral ligaments, respectively, if reduction is not perfect. The use of splint immobilization for nondisplaced fractures and single Kirschner wire fixation for displaced fractures had a high incidence of subsequent displacement and required careful follow-up examination with a high index of suspicion for fracture movement. Final range of motion was poorest in those fractures treated with a single Kirschner wire because of those that displaced. Although this group represented the simplest fracture pattern without comrninution, instability at the fracture site represented an important component that is probably best addressed by multiple fixation sites (i.e., two or more Kirschner wires and/or screws). The palmar oblique fracture plane of class I injuries carries significant fixation implications. Kirschner wires that enter from the fractured condylar fragment in the central or dorsal portion will obtain purchase on only a very thin, weak cortical portion of the fragment and will be susceptible to rotation or displacement. The best fixation area in these fracture fragments is the distal palmar aspect. One should attempt to direct Kirschner wire fixation from the contralateral dorsal side of the shaft with the intact condyle into the palmar aspect of the fractured condyle. Occasionally fracture line orientation makes this approach difficult, and in these cases it is best to go from fracture fragment to intact condyle, thus avoiding numerous passes. At least one other Kirschner wire should be placed for control of alignment. With the use of a miniscrew, the same principle should be applied to maximize cortical purchase. In the placement of Kirschner wires or miniscrews, care should be taken to avoid transfixing the collateral ligaments, which might limit their excusion postoperatively. The area just proximal to the collateral Iigament origin is ideal for hardware placement.
Distal unicandylar fractures of proximal phalanx
599
In fractures treated with Kirschner wires alone, whether reduction was open or closed did not appear to influence final motion. With presumably less postoperative scarring, those treated with closed reduction should have better final motion, but this assumption may be neutralized by a more stable, exact reduction under open observation allowing earlier motion. One issue raised by this study is what factors involved in a lateral bending proximal interphalangeal injury dictate a unicondylar fracture versus a collateral ligament tear. A biomechanical study by Kiefhaber et al.’ of the lateral stability of the PIP joint to lateral stress, which loaded 78 joints in a tension fashion, demonstrated only one unicondylar fracture, with all the other injuries being ligamentous. The loading rate in this study was very low by clinical standards (60 to 90 degrees per second). A high loading rate, as would be seen in a ball sport injury, might predispose to fracture rather than ligament tear. A similar clinical situation occurs in the distal femoral unicondylar fracture, most frequently seen in pedestrian/motor vehicle accidents, in which a unicondylar fracture occurs from lateral bending/ tension failure of the bone by the collateral ligament as the car bumper strikes the leg at a high loading rate.
REFERENCES 1. Hastings HH II, Carrol C IV: Treatment of closed articular
2. 3.
4.
5.
fractures of the metacarpophalangeal and proximal interphalangeal joints. Hand Clin 1988;4:503-27. London PS. Sprains and fractures involving the interphalangeal joints. Hand 1971;3:155-8. McCue FC, Honner R, Johnson MC Jr, Gieck JH. Athletic injuries of the proximal interphalangeal joint requiring surgical treatment. J Bone Joint Surg 1970;521\:937-55. Soeur R. Fractures of the limbs: the relationship between mechanism and treatment. Brussels: La Clinique Orthopedique, 1981542-3. Kiefhaber TR, Stem PJ, Grood ES. Lateral stability of the proximal interphdangeal joint. J HANDSURG 1986; 1 lA:661-9.