Complications in type III open tibial shaft fractures treated with open reduction and internal fixation

The primary objective of this study was to estimate the short- and mid-term rate of FRI and nonunions, in a commercially insured population of patients treated with ORIF for all GA Type III open TSF. The secondary objectives of this study were to identify risk factors associated with FRI, and to describe the economic burden on the healthcare system from the perspective of a payer.

There were 580 patients included in the study (Figure 1) of which 21.03% (n = 122) were female. The majority of patients (63%) were covered under preferred provider organization (PPO) plans. On average patients were enrolled for 638 days (SD: 677; median 419). Eighty seven percent of patients had their index ORIF procedure during an inpatient admission and 8.45% patients had index ORIF procedure in the emergency care, which would most often lead to an inpatient admission (Table 1). The median number of days to first FRI recorded after index was 58 days among patients with FRI. In the first 30 days of follow-up, 16.55 % of patients had a FRI diagnosis (Table 2). Over longer follow-up durations, the rates of FRI steadily increased up to 44.44% at month 24. For nonunion, there was an increase in the rate from 36.94% at day 365 to 41.98% at month 24 (Table 2).

The readmission rate for the group of patients with FRI was significantly higher throughout the different follow-up time periods. At the 30-day period, the rate of readmissions was more than three-times higher for the FRI group as compared the no FRI group (44.79 vs 13.22%; p < 0.0001). The rate of readmission consistently stayed higher for the group with FRI as the follow-up duration increased (Table 3). The rates for emergency room visits were also higher for patients with FRI throughout the different follow-up time periods, however they were not always statistically significant. Consistently, throughout the different follow-up time periods the group of patients with FRI had greater mean total healthcare costs compared with those without (at day 30: $37,597 vs $9886; p < 0.0001; at 24 months: $140,227 vs 50,259; p < 0.0001; Table 4).

At 365 days follow-up, the adjusted results showed that the rates of readmission (79.08 vs 30.49%; p < 0.0001) and emergency room visits (44.79 vs 26.00%; p = 0.003) were statistically significantly higher for patients with FRI as compared with patients without FRI (Figure 2). Similarly, patients with FRI had higher mean adjusted total costs compared with patients without FRI ($ 91,885 vs $31,186; p < 0.0001) (Figure 3). Logistic regression of risk factors showed that region and year for index ORIF were significantly associated with the risk of FRI for 365 days follow-up (Supplementary Table 2). North central region had 57% lower odds of FRI as compared with the West region. As compared with the year 2016, all prior years for index ORIF surgery had statistically significantly higher odds of FRI.

Based on intention to treat approach (n = 580), we found that rates of FRI and nonunion were high as well. For FRI, the rates were 34.48 and 35.69% at 365 days and 24 months follow-up, respectively. For nonunion, the rates were 24.48% and 25.86% at 365 days and 24 months follow-up, respectively.

To our knowledge, this is the first US-based study utilizing a national database of insured patients to estimate risk of both FRI and nonunions specific to high-risk TSF patients. The results of our study show a persistent risk of FRI and nonunion, especially in this high-risk fracture population, substantially complicates patient recovery and escalates healthcare cost. The literature reports rates of FRIs following ORIF procedures on the tibia ranging from 2 to 11% [13,14] without specifically considering open/high-risk fractures. Previous studies have reported ranges from 16 to 36% of patients developing FRI after Type III open TSFs [3,15]. In line with published literature, our nationally representative study of patients with GA Type III open TSF, found that 17% and 36% of patients developed FRI within 30 days and 365 days, respectively, post surgery. Our study also found that nonunions developed in 41.98% of cases at 24 months similar to the rates reported in open fractures (17–48%) [6–9]. Study by Salem reported that the mean time to union for Type III open tibial fractures was 44 weeks and 40% patients had nonunion, this study showed a similar nonunion rate of 41% at 52 weeks [16].

All-cause readmissions and total healthcare cost were consistently greater in the FRI group at all follow-up durations. In the first month after ORIF procedures, patients with FRI recorded average healthcare cost of $37,597 (SD: 59,689), and patients without FRI had an average cost of $9886 (SD: 23,539). At the end of 24 months, the average cost of the group with FRI rose to $140,227 (SD: 117,317) which was still almost three-times greater than the average cost for the group without FRI ($50,259 [SD: 60,807]). More effective clinical management during surgical intervention may reduce the risk of FRI and minimize the ongoing complication risk and costs. The focus must be on strategies that not only mitigate the immediate risk of FRI, but also reduce the risk during postoperative follow-up. Beaver et al. calculated the costs among patients with and without FRI and found that in case of absence of FRI, the total costs were very close to the amounts that we have found, $9407 per patient in Beaver et al. study and $9886 in our study [17]. However, the costs calculated in case of FRI were very different $13,739 as shown in Beaver et al. study and $37,597 in our study. The possible reasons for difference can be different periods for which the costs were measured and small sample size in case of the study of Beaver et al. but there could be other contributing factors. Unadjusted analysis by Hoekstra et al. reported a fivefold increase in the total healthcare costs among patients with a deep infection versus those without infection in open tibial fractures [18]. Study by Hoekstra et al. was conducted in Belgium with maximum follow-up time of 2 years for all patients [18]. Our US-based study, at 1 years follow-up showed about a threefold increase in costs among patients with FRI as compared with patients without FRI. The main difference in the results may be because Hoekstra et al. compared patients with deep infections, which may be severe cases than the cohort in our study that included patients with any FRI.

Underlying reasons for the statistical significance of procedure year as a predictor of FRI risk in our study cannot be ascertained without further investigation. Lower rates of FRI in 2016 may have resulted from undetected temporal trends in patient characteristics, surgical techniques, technologies and/or provider characteristics. Unknown or unmeasured procedural characteristics may also have influenced FRI rates over time (e.g., nature of fractures type IIIA, IIIB and IIIC, quality of reduction, infection control practices, etc.)

These results should be viewed in light of several limitations. Our study looked only at the direct costs. Other studies, like the one by Hak et al. looked also on the indirect costs, caused by the loss of productivity [10]. This can be a subject for further studies. Another limitation is that all measures and factors included were only those available in the claims, from recorded diagnoses for those who have insurance coverage. Therefore, other factors not captured by administrative claims data which contribute to complication-related costs may have been under reported. Most serious complications, however, are reliably documented in the claim file. Our findings may not generalize to high-risk patients without insurance coverage. However, this database consists of healthcare information of over 40 million enrollees. Additionally, since it includes major US health plans across private insurances, it is representative of the vast majority of patients aged 18–64 years. Another challenge with claims data is underestimation of outcome variables when patients do not seek care (or are denied access). To that end, these data reflect estimated complication risk, resource use and costs of those who sought care and were covered by one of the many plans in the database. Also, the FRIs were identified based on ICD-9, ICD-10 and CPT codes and may be subject to coding errors associated with any claims database. Similarly, identification of nonunion was solely based on the ICD diagnosis codes without information about the union of bone through radiographs. Future large database studies utilizing electronic medical records with information on radiograph to report nonunion are warranted to confirm the study findings. Finally, we were not able to differentiate between GA Type IIIA, IIIB and IIIC fractures due lack of specificity of the codes.

This study found that the short and mid-term rates of FRI and nonunion are high, reaching about 44 and 42%, respectively, over the 2-year period in patients with an ORIF procedure for GA Type III open TSFs. Within this high-risk population, FRI significantly increased need for hospital readmission and total healthcare costs.

To view the supplementary data that accompany this paper please visit the journal website at: Supplementary Material

This study was funded by Johnson and Johnson. AS Chitnis, M Vanderkarr, C Sparks, J McGlohorn and CE Holy are employees and stock holders of Johnson & Johnson. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing assistance for the initial draft was provided by eMAX Health, White Plains, NY, which received funding from Johnson and Johnson.

IBM MarketScan database is in compliance with the Health Information Portability and Accountability Act of 1996 (HIPAA). Only de-identified patient records from IBM MarketScan database were used in this study; hence Institutional Review Board approval was not required.