Study Design
Between 1974 and 2002, all patients who received operative treatment for fractures of both bones of the forearm at our institution were included in a trauma fracture database and classified according to the AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) Comprehensive Classification of Fractures21. In this database we identified 157 patients who were treated for diaphyseal fractures of both forearm bones (AO/OTA 22 classification type A3, B3, or C1, C2, or C3) before 1999, allowing for at least ten years of follow-up22. Thirty-four of these patients were excluded: seven of the thirty-four patients had incomplete demographic data, seven were visitors (tourists or exchange students), and twenty had died.
The remaining 123 patients were invited for a long-term follow-up visit at our outpatient clinic under a protocol approved by our institutional review board. Thirteen of these patients were excluded: four of the thirteen patients were pregnant during the study period; three had complete brachial plexus injuries that did not resolve; three had severe central nervous system injuries; one was mentally challenged; and two patients had generalized musculoskeletal disorders. From the remaining 110 eligible patients, twelve (11%) declined participation, seven (6%) lived abroad, and twenty (18%) did not respond or could not be contacted after three written and three phone requests.
Seventy-one (65%) of the 110 patients were evaluated at a mean time of twenty-one years (range, thirteen to thirty-three years) after injury. Sixty-five were evaluated by an independent observer who was not involved in the care of the patients, four were evaluated at home, and two were unable to visit our university hospital but were evaluated by an orthopaedic surgeon at their local hospital. Informed consent was obtained from all patients in the study.
Three treatment strategies were identified: closed reduction and cast immobilization in thirty patients, external fixation in three patients, and open reduction and plate-and-screw fixation in thirty-eight patients, one of whom had initial temporary external fixation.
Patient Characteristics
There were forty-four men and twenty-seven women, with a mean age at the time of follow-up of forty-two years (range, nineteen to eighty-one years) (see Appendix). Skeletal maturity was determined on radiographs taken at the time of injury. Thirty-six patients were skeletally mature (mean age, thirty-one years) and thirty-five were skeletally immature (mean age, eleven years) at the time of injury. In the mature group, fifteen fractures were classified as simple (type-A3) fractures, fourteen were wedge (type-B3) fractures, and seven were complex (type-C) fractures. In the immature group, thirty-one fractures were classified as simple (type-A3) fractures, four were wedge (type-B3) fractures, and no fractures were complex (type-C) fractures. Twenty-four patients (34%) had open fractures; ten of these patients were in the skeletally immature group and fourteen were in the skeletally mature group23.
Thirty-two patients, including twenty-nine (81%) of the thirty-six patients in the mature group and three (9%) of the thirty-five patients in the immature group, sustained high-energy trauma. The mechanisms of injury in the mature group were a fall from a height (two patients), a motor-vehicle accident (twenty-one patients), industrial accidents (four patients, all with severe crush injuries), and a fireworks explosion (two patients). The mechanisms of injury in the immature group were a motor-vehicle accident (two patients) and an industrial accident involving a severe crush injury (one patient). The injuries of thirty-nine patients were sustained during a low-energy fall (thirty-two in the immature group and seven in the mature group). Twenty-two patients had multiple-system trauma (two in the immature group, and twenty in the mature group). The mature group had twenty-nine fractures of other limbs (thirteen upper-extremity fractures and sixteen lower-extremity fractures). In the immature group, one patient had ipsilateral fractures of both the upper and the lower extremity. Forty-nine patients sustained no other injury (sixteen in the mature group and thirty-three in the immature group). Eleven patients in the mature group had ipsilateral concomitant nerve injuries, including radial nerve injury in six, partial brachial plexus injuries in two, ulnar nerve injury in four, and median nerve injury in two patients. One patient in the immature group had a radial nerve lesion.
Treatment
We considered the term treatment to mean the treatment on the day of injury, and we considered the term complication to mean any deviation from that treatment. Of the thirty-five skeletally immature patients, twenty-five (71%) were treated with closed reduction and cast immobilization and ten were treated with plate-and-screw fixation. Of the thirty-six skeletally mature patients, twenty-eight (78%) were treated with plate-and-screw fixation, five were treated with closed reduction and cast immobilization, and three were treated with external fixation. In total, thirty patients were initially treated with closed reduction and cast immobilization for at least six weeks, with a maximum treatment time of twelve weeks (range, two to twelve weeks). Five of these thirty patients underwent open reduction and internal fixation with plates and screws after loss of reduction.
Twenty-eight skeletally mature patients and eight skeletally immature patients were treated with open reduction and internal fixation of both the ulna and radius. Two skeletally immature patients were treated with only one plate, which was affixed to the radius in one patient and to the ulna in the other.
Four patients with open fractures had initial external fixation but later underwent conversion to plate-and-screw fixation at sixteen days, three months, four months, and six months, respectively, after the injury.
Plate-and-screw fixation had been performed by at least twenty different orthopaedic trauma and general trauma surgeons, and no standard protocols were used. The volar approach of Henry (for the radius) and the subcutaneous approach (for the ulna) were used in the majority of patients24. In six patients, the surgical approach was not documented.
Internal fixation was performed with 3.5-mm dynamic compression plates (Synthes, Zeist, The Netherlands) in thirty-three patients. In seven patients, a one-third tubular plate was used, and, in one patient, a 3.5-mm dynamic compression plate was used for the radius and a one-third tubular plate was used for the ulna. No bone-grafting was performed primarily. The postoperative protocol was not standardized, but it included short-term immobilization for less than one week in an above-the-elbow splint with the forearm in the neutral position, followed by active-assisted range-of-motion exercises according to a similar protocol to that used in the study by Droll et al.4.
Complications and Subsequent Procedures
Four skeletally immature patients and one skeletally mature patient who were initially treated with cast immobilization had loss of reduction while wearing the cast and were subsequently treated with open reduction and internal fixation (three skeletally immature patients and one skeletally mature patient) and/or intramedullary nail fixation (one skeletally immature patient) within two to thirty-six days after injury.
The skeletally immature patient who was treated with intramedullary nail fixation after initial cast treatment developed a deep infection that was treated with debridement, removal of the intramedullary nail, and oral antibiotics. Another skeletally immature patient developed an infection within the pin-track of an external fixator, which was treated with removal of the external fixator, debridement of the pin site, and oral antibiotics. Three weeks later, plates and screws were placed in this patient.
Two skeletally mature patients and one immature patient had a deep wound infection along with osteomyelitis after sustaining high-energy open fractures. A skeletally mature patient with an external fixator had a deep infection that was treated with removal of the fixator, debridement, and placement of antibiotic-impregnated methylmethacrylate beads. Plates and screws were applied ten days later with autogenous cancellous bone-grafting. A deep infection developed five months after this procedure, with the fracture united, and was treated with debridement, removal of implants, and open wound-packing. After this procedure, the patient remained free of infection.
Another deep infection was treated successfully with serial debridement and parenteral antibiotics. One skeletally immature patient developed osteomyelitis with a draining sinus, which resolved with parenteral antibiotic treatment followed by oral antibiotic treatment.
Other complications included two patients (one skeletally mature and one skeletally immature patient) who had pain and stiffness that were disproportionate to what was merited and who were diagnosed with complex regional pain syndrome. One skeletally immature patient was diagnosed with an exertional compartment syndrome in the forearm twenty years after the original injury and was treated surgically.
There were five refractures after closed reduction and cast immobilization. One patient in the mature group had a refracture seven months after injury and was treated with plate-and-screw fixation. Three patients in the skeletally immature group had a refracture after three, three, and twelve months, respectively, and were treated with repeat closed reduction and casting. One of these patients had a second manipulation for a 30° malalignment of the radius. Another patient in the skeletally immature group had two refractures (four and eleven months after the index injury), the first of which was treated with casting and the second of which was treated with open reduction and internal fixation. All refractures healed.
Seven operatively treated patients (including the one with the infected nonunion, described above) had a second surgery to address delayed fracture-healing an average of five months (range, four to seven months) after injury. All seven fractures healed after plate-and-screw fixation and autogenous cancellous bone-grafting.
One patient had extensor tenolysis and three patients (two skeletally mature and one skeletally immature) had tendon transfers to compensate for substantial muscle loss after high-velocity open fractures.
Of the total of forty-six patients who were treated with open reduction and internal fixation primarily or in a subsequent procedure, twenty-seven patients (59%) had implant removal a mean of twenty-five months (range, thirteen to forty-eight months) after surgery. No refractures occurred in this group.
Evaluation
Arm-specific disability was evaluated with use of the validated Dutch form of the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire25. Scores range between zero and one hundred points, with a higher score indicating worse upper-extremity-specific disability and pain. Pain was scored as the mean of three different questions on the DASH, concerning pain at rest, pain with activity, and sleeping difficulties. The maximum possible pain score was five (severe pain resulting in sleep difficulties) and the minimum was one (no pain).
Depression during the previous week was quantified with use of the validated Dutch form of the Center for Epidemiologic Studies-Depression scale (CES-D)26. Scores range between zero and sixty points, with a higher score indicating a higher level of depression26. A mean CES-D of 13 was measured in a healthy Dutch population (scores above 16 indicate a depressive disorder)26.
The validated Dutch version of the Pain Catastrophizing Scale (PCS) was used to measure misinterpretation or overinterpretation of nociception (i.e., the tendency toward excessively negative interpretation of nociception as indicating harm or hopelessness)27,28. Nociception is the physiological response to actual or potential tissue damage, and pain is the interpretation of nociception. Nociception occurs on a spectrum between healthy (e.g., stretch pain or post-workout pain) and unhealthy (e.g., a burn from a hot stove). The concept of "catastrophizing" captures the tendency to interpret nociception as more unhealthy than merited. The PCS contains thirteen questions regarding over-interpretation and misinterpretation of nociception. Scores range between 0 and 62 points, with a higher score indicating a more exaggerated negative orientation toward pain28. Scores between 12.1 and 13.9 are considered normal27,28.
Range of motion was measured in a standardized fashion with a handheld goniometer by an independent observer not involved in the patient's care. Forearm rotation was measured with the elbow in 90° of flexion and with the arm at the patient's side. Elbow and wrist flexion and extension were measured, as was wrist radial and ulnar deviation. Grip strength was measured with use of a Jamar dynamometer (Asimow Engineering, Santa Monica, California) with the elbow at 90° of flexion and the forearm in neutral rotation. Grip strength was also calculated as a percentage of the unaffected arm. For patients with contralateral injuries, normative measurements, adjusted for age and sex29, were used as controls.
Standardized anteroposterior and lateral radiographs of both bones of the injured forearm were made, with the forearm in neutral position, at the long-term follow-up visit. The radiographs were evaluated by two independent observers (J.C.G. and A.G.J.B.) in a blinded fashion. The amount and location of the radial bow were measured on the anteroposterior radiograph with use of the method described by Schemitsch and Richards7. First, a line is drawn from the bicipital tuberosity to the most ulnar aspect of the radius at the wrist7. Next, the point of maximum radial bow is connected to this line with a perpendicular line; the length of this line is the maximum radial bow in millimeters7. The location of the radial bow is the distance from the bicipital tuberosity to the perpendicular line, divided by the length of the total line from elbow to wrist7.
The result was rated according to the method of Anderson et al.1: the result was considered excellent when union occurred with <10° loss of wrist flexion-extension and <25% loss of pronation-supination; satisfactory when union occurred with <20° loss of wrist flexion-extension and <50% loss of pronation-supination; unsatisfactory when union occurred with >30° loss of wrist flexion-extension and >50% loss of pronation-supination; and a failure when there was nonunion with or without loss of motion.
Statistical Analysis
Continuous data were presented as the mean when normally distributed; otherwise, the medians were reported. Post-hoc power analysis with use of the Student t test indicated that a minimum of thirty patients per group provided 77% power to detect a 0.71 standard deviation difference in the arc of forearm rotation of skeletally mature and immature patients.
In bivariate analysis, the association between the DASH score with continuous independent variables was investigated with use of the Spearman correlation. The association between the DASH score with dichotomous independent variables (presence of associated arm injuries, compound fractures, sex, injury to dominant arm, occupation/heavy work [white/blue collar]) and skeletal maturity at the time of injury was investigated with use of the Mann-Whitney U test. The Kruskal-Wallis test for ordinal data was used for analysis of AO fracture type (type A3, B3, C1, C2, and C3). Associations with a p value of <0.05 were considered significant.
The number of independent variables that can be included in a multivariate model is limited by the overall sample size of the study. Therefore, we inserted seven variables (one variable for every ten patients) that were significant (p < 0.03) in the multivariate analysis. For each model, a multivariate analysis of variance was performed to assess significance, which indicates a linear relationship between at least one of the independent variables and the dependent variable (i.e., the DASH score). Backwards stepwise multiple linear regression analysis was used to analyze the ability of the independent variables to account for variation in the dependent variables, thereby accounting for confounding between the independent variables. This multiple linear regression model produced the adjusted R-squared (R2), reflecting the percentage of the overall variability in the dependent variable explained or accounted for by the independent variables included in the multiple linear regression model.
Source of Funding
There was no external funding source.
Long-Term Functional Outcome (Tables I and II)
Outcome Measures
The average DASH score of 8 points (and standard deviation of ±12 points) was near the mean DASH score (10.1) in healthy individuals in the United States30. There were no significant differences in DASH, CES-D, PCS, grip strength, and radiographic measures of mean location of radial bow and mean radial bow between skeletally mature and immature patients. The mean CES-D and PCS were comparable with Dutch community norms26-28. There were small but significant differences in the average wrist and forearm motion compared with the opposite arm. There were small but significant differences in elbow flexion-extension, wrist flexion-extension, pronation-supination, and radioulnar deviation arc favoring patients who were skeletally immature when they sustained the injury (Tables I and II). There was no significant difference in average pain score at the time of long-term follow-up between patients with and without retained plates and screws (mean 1.41 vs. mean 1.38; p = 0.88). The result, according to the system of Anderson et al.1, was excellent or satisfactory in sixty-nine patients (97%) and unsatisfactory in two (3%). Because all fractures had healed by the time of follow-up, there were no failures according to the rating of Anderson et al.
Bivariate and Multivariate Analysis (Tables III and IV)
With bivariate analysis, there was a significant association between DASH scores and the arc of forearm rotation, the arc of wrist flexion and extension, the arc of radioulnar deviation, the arc of elbow flexion and extension, CES-D, PCS, age at injury, pain, and associated injuries. There was no significant association between DASH scores with radial bow, number of surgical procedures, follow-up time, patients that were skeletally mature at the time of injury, open injury, limb dominance, sex, occupation, and AO fracture type.
Backward stepwise multiple linear regression was executed, starting with pronation-supination arc, grip strength, age at injury, pain, ipsilateral injury, CES-D, and PCS. This type of regression analysis initially includes all variables and then removes the least important and statistically nonsignificant variables one by one until the model retains only the strongest, significant explanatory variables. The final model included grip strength, pain, and PCS and accounted for 56% of the variability in DASH scores (R2 = 0.559, p < 0.001). A model with pain alone explained 40% of the variability in DASH scores (R2 = 0.40, p < 0.001).