This study was performed with the approval of the Human Subjects Review Boards at the two institutions. Between January 2000 and August 2005, eighty-one patients with unilateral tibial plafond fractures (AO/OTA type B-2, B-3, C-1, C-2, or C-3) were treated and entered into a database. Patients with bilateral fractures, ipsilateral calcaneus or talus fractures, those with type III open wounds, and those patients with an Injury Severity Score of 18 or greater were not included. Patients with head injuries, severe osteopenia, or patients who had previous attempts to surgically reduce the articular surface were also excluded. Pregnancy, previous ankle fractures, contralateral ankle abnormalities, and intervening ankle trauma between the index injury and the two-year follow-ups were additional exclusion criteria.
The protocol was for patient clinic visits to obtain outcome measures at specific time points after injury - 1, 2, 3, 6, 12, and 24 months. Since this study was designed to describe a specific time-course to recovery, only patients with reasonably complete data were included. To be included, patients had to have a 24 month follow-up and not miss more than two of the previous follow-up appointments. These inclusion criteria were chosen to balance keeping sufficient patient numbers in the study against the importance of having complete follow up at all time points. Of the 81 patients entered into the database, 43 patients (mean age, 41.8 years; range, 20 - 60 years) met these criteria and were included in this study. These 43 patients had a 24 month follow-up and were seen for at least three additional visits prior to 24 months. Two study patients passed away during follow-up, one patient moved, and thirty-five patients were not included due to lack of adequate follow-up. To assess the effect of the high number of excluded patients the forty three included patients were compared to the thirty-eight excluded patients for factors such as age, gender, fracture classification, articular comminution, education, associated fibula fracture, income, associated injuries, other medical problems, employment status, plans to return to work, involvement in legal action due to injury, and compensation status. The only significant difference between the two groups was in education level with a larger proportion of patients that completed college in the group that were included in the study (p-value of 0.05).
There were 28 males (65.2%) and 15 females (34.8%). The AO/OTA classification of the fractures was 31 types C (4 C1, 9 C2, and 18 C3) and 12 type B (5 B2, 7 B3).12
The mechanisms of injury were falls from a height (range, 4 - 25 feet) in 60%, motor vehicle accidents in 20%, and the other 20% were caused by other mechanisms including crush, bicycle accidents, and motorcycle accidents. Each fracture was treated with a spanning articulated external fixator and limited internal fixation of the articular surface. The approach to reducing and internally fixing the articular surface was variable based on the case and the treating surgeon and ranged from percutaneous techniques only to fairly extensive open reductions. This technique has been previously published.12
At each clinic visit patients were assessed with a general health status measure, Short-Form 36 (SF-36), and a joint specific outcome measure, Ankle Osteoarthritis Scale (AOS). The SF-36 was recorded at all time points. The AOS was recorded at 6, 12, and 24 months after injury.
The SF-36 MOS version 2.0 is a thirty-six question general health survey designed to assess physical and emotional health and health-related quality of life.13
It produces an eight-scale profile of scores as well as physical (PCS) and mental (MCS) component summary scores based on twenty-one and fifteen questions, respectively.13
These scores were scaled and normalized to 1998 population norms. For both the PCS and the MCS a score of 50 is equivalent to age matched norms with one and two standard deviations above and below the normal scores being plus and minus 10 and 20 points, respectively.
The AOS was used to measure pain and disability related to the injured ankle. Both the pain and disability scores are nine question visual analogue scales measured from 0 - 100, with higher scores indicating greater pain and disability. The AOS has previously been found to be a reliable and valid assessment instrument. Normal subjects report little ankle pain or disability and have been found to have scores close to one. 14, 15
For each clinic visit, the scores for the MCS, PCS, AOS pain and AOS disability for all patients were averaged and the averages were compared to the averages at the previous and subsequent clinic visits. The time-course to recovery was assessed by comparing averages of the scores for the four outcome measures across adjacent time points.
To assess whether there were factors that affected the speed of recovery, the patients’ PCS at 12 and 24 months were compared. Patients were divided into three categories; those that showed no significant change between the 12 and 24 month time points, those that deteriorated (< 1 STD, PCS − 10), and those that showed continued improvement between the 12 and 24 month visits (≥ 1 STD, PCS + 10). Factors such as age, gender, fracture classification, articular comminution, associated fibula fracture, and education were analyzed for significant differences between these three groups of patients. Articular comminution was assessed on plain x-ray and classified by the investigators using the AO/OTA classification of fractures of the distal tibia. Fractures classified as C3 and B3 were scored as articularly comminuted and those classified as B1, B2, C1, and C2 were scored as not comminuted. Levels of education were classified according to four categories; did not complete high school, completed high school, some college or vocational school, completed college. Articular comminution and level of education were classified as dichotomous variables.
In addition to the speed of recovery, the effect of a broader set of factors on the outcome at 24 months after injury was assessed. The risk factors assessed were age, gender, fracture classification, articular comminution, education, income, associated injuries, other medical problems, employment status, plans to return to work, involvement in a legal action due to injury, and compensation status. The outcomes assessed were the AOS and SF-36 scores at the 24 month follow-up.
A mixed model with statistical contrasts was used to compare outcomes at adjacent time-points because the data included repeated measures and contained correlation within each group with some empty cells. Several tests with different correlation (covariance) structures were performed, and based on the model fitting criteria, the SP(POW) covariance structure was chosen. Predictive factors in the speed of recovery were modeled using logistic regression with a forward variable selection strategy. A univariate model was also used due to the limited power for this analysis. At two years after injury, a general linear model was used to assess the relationships between response variables PCS/MCS and several demographic variables.