Influence of Level 1 Evidence on Management of Clavicle and Distal Humerus Fractures: A Nationwide Comparative Study of Records From 2005 to 2014

Abstract

The purpose of this study was to examine alterations in national trends managing midshaft clavicle fractures (MCF) and intra-articular distal humerus fractures (DHF) surrounding recent level 1 publications. A retrospective review of the PearlDiver supercomputer for DHF and MCF was performed. Using age limits defined in the original level 1 studies, total use and annual use rates were examined. Nonoperative management and open reduction and internal fixation (ORIF) were reviewed for MCF. ORIF and total elbow arthroplasty (TEA) were reviewed for DHF. A query yielded 4929 MCF and 106,535 DHF patients. A significant increase in ORIF use for MCF following the publication of the level 1 study (p = .002) and a strong, positive correlation (p = .007) were evident. Annual TEA (p = .515) use for DHF was not observed.

Level 1 therapeutic studies are high-quality randomized control trials (RCT) that tend to have the most impact on clinical practice. While these studies often provide sufficient evidence on the outcomes of various treatment modalities for a given pathology, their impact on treatment trends is not well understood. Popularity of a treatment selection for a given pathology is likely multifactorial. Nevertheless, trends on surgical management for various orthopaedic fractures are likely to be influenced by these studies.

One of the most notable shifts in fracture treatment in recent years was that of displaced midshaft clavicle fractures (MCF). Historically, these fractures were, for the most part, treated nonoperatively. In 2007, a level 1 prospective, multicenter, randomized clinical trial was performed by the Canadian Orthopaedic Trauma Society (COTS), comparing nonoperative to operative management of these fractures (1). This study created a greater appreciation for MCF that may be better managed with surgical treatment. It has been shown that use of open reduction and internal fixation (ORIF) has since increased, as observed in the American Board of Orthopaedic Surgery database (2), adolescents (3), and the Finnish registry (4). In 2009, McKee et al. similarly performed a prospective, randomized, double-blind, controlled trial comparing ORIF and total elbow arthroplasty (TEA) for displaced intra-articular distal humerus fractures (DHF) in the elderly (5). Despite demonstrating superiority of TEA compared with ORIF, it remains unclear if the use of TEA has changed. TEA is often considered a more technically demanding operation than clavicle ORIF and thus is performed only by select sub-specialists comfortable performing the operation.

With the strong emphasis on evidence-based medicine, there exists a need to gain a greater understanding of the influence of high-level evidence studies on upper extremity fracture management. The purpose of this study was to examine alteration in national trends in management of MCF and intra-articular DHF surrounding recent level 1 publications. The authors hypothesized that management of displaced MCF was significantly influenced by level 1 evidence, while the management of DHF was not.

Materials and Methods

The entire Medicare and Humana database was retrospectively reviewed through the use of the Pearl-Diver supercomputer (Warsaw, Indiana) for intra-articular DHF and MCF, respectively. This software is a publicly available Health Insurance Portability and Accountability Act–compliant database that compiles the entire Medicare and Humana files into a server. Data compiled within the Humana database are derived from a combination of Medicare/ Medicare Advantage and private payers yielding 18,620,198 patients from 2007 to 2014. The Medicare database is derived from Medicare parts A and B to include 49,550,651 from 2005 to 2012. PearlDiver software provides access to the Medicare and Humana registries with the use of Current Procedural Terminology (CPT; American Medical Association, Chicago, Illinois) codes, as well as International Classification of Disease, Ninth Edition (ICD-9) codes.

A query of the Humana database was performed for open and closed MCF using the CPT codes 810.12 and 810.02, respectively. Patients treated with ORIF were identified among this group through a query using CPT code 23515 and ICD-9 78.51. Nonoperatively treated MCF were selected using CPT 23500 and CPT 23505 (Table 1). An age restriction of 15 to 59 years was set for MCF, in accordance with the methodology described in the level 1 study performed by the Canadian Orthopaedic Trauma Society in 2007 (1).

TABLE 1.

Codes used for the query of the Medicare and Humana database

Procedure	Code
Clavicle fracture
Closed fracture of the shaft of clavicle	ICD-9: 810.02
Open fracture of the shaft of clavicle	ICD-9: 810.12
ORIF of clavicle
Open treatment of clavicular fracture with or without internal or external fixation	CPT 23515
Internal fixation of bone without fracture reduction clavicle	ICD-9: 78.51
Nonoperative management of clavicle fractures
Closed treatment of clavicle fracture without manipulation	CPT 23500
Closed treatment of clavicle fracture with manipulation	CPT 23505
Distal humerus fracture
Closed fracture of unspecified part of lower end of humerus	ICD-9: 812.40
Closed supracondylar fracture of humerus	ICD-9: 812.41
Closed fracture of lateral condyle of humerus	ICD-9: 812.42
Closed fracture of medial condyle of humerus	ICD-9: 812.43
Closed fracture of unspecified condyle&Obracket;s&Cbracket; of humerus	ICD-9: 812.44
Other closed fracture of lower end of humerus	ICD-9: 812.49
Open fracture of unspecified part of lower end of humerus	ICD-9: 812.50
Open supracondylar fracture of humerus	ICD-9: 812.51
Open fracture of lateral condyle of humerus	ICD-9: 812.52
Open fracture of medial condyle of humerus	ICD-9: 812.53
Open fracture of unspecified condyle&Obracket;s&Cbracket; of humerus	ICD-9: 812.54
Other open fracture of lower end of humerus	ICD-9: 812.59
Total elbow arthroplasty
Arthroplasty elbow with distal humerus and proximal ulnar prosthetic replacement	CPT 24363
Open treatment of periarticular fracture of the elbow with implant arthroplasty	CPT 24587
Total elbow arthroplasty	ICD-9: 81.84
Open reduction internal fixation of humerus
Open reduction with internal fixation, humerus	ICD-9: 79.31
Open treatment of humeral supracondylar or transcondylar fracture, includes internal fixation	CPT 24545
Open treatment of humeral supracondylar or transcondylar fracture, with intercondylar extension	CPT 24546
Open treatment of condylar fracture, medial or lateral	CPT 24579
Open treatment of periarticular fracture of the elbow	CPT 24586

Similarly, the Medicare database was queried for TEA using CPT 24363, CPT 24587, and ICD-9 81.84. ORIF for DHF was identified with CPT codes 24545, 24546, 24579, and 24586 (Table 1) and ICD-9 79.31 (Table 1). The age evaluated for intra-articular DHF was >65 years, in accordance with the methodology described in the level 1 study performed by McKee et al. in 2009 (5).

Total use and utilization rates were examined. In addition, the Humana population was analyzed to evaluate any change in the percentage of ORIF used and trends in the incidence of displaced MCF among the same population. Similarly, the Medicare population was also analyzed to evaluate any change in the percentage of TEA used and trends in the incidence of DHF among the same population.

Statistical analysis was mainly descriptive to define the incidence of each procedure throughout the study period. Analysis of linear regressions, chi-square trend analysis, correlation studies using the Pearson method, and Student t tests were used where appropriate. A p value of <.05 was selected to be significant. Statistical analysis was performed with Prism 6 software (GraphPad Software, La Jolla, California).

Results

The query returned a total of 106,535 visits coded for DHF in patients age >65 years within the Medicare population between 2005 and 2012 (Table 2). There were no significant trend changes in annual visits coded for DHF in this cohort between 2005 and 2012 (p = .614) (Fig. 1). Between 2005 and 2012, a total of 4849 TEAs were performed for DHF in patients >65 years of age in the Medicare population. There was no significant change in the annual use of TEA for DHF in this cohort between 2005 and 2012 (p = .515) (Fig. 1) as well as no significant change in the percentage of TEA used as the definitive management choice (p = .625) (Table 2). Moreover, there were no significant trend differences in the annual use of TEA for DHF in the Medicare population between 2005 and 2009, the time frame before publication of the level 1 study when compared to the period between 2010 and 2012, the time after publication (p = .602).

TABLE 2.

Incidence of DHF and TEA use

Year	Incidence of DHF in Cohort	TEA Used as Definitive Treatment	Percentage Use of TEA Definitive Management
Before level 1 publication
2005	13,155	537	4.08%
2006	13,289	635	4.78%
2007	13,772	589	4.28%
2008	12,834	591	4.60%
2009	13,306	647	4.86%
After level 1 publication
2010	13,158	663	5.04%
2011	13,578	624	4.60%
2012	13,443	563	4.19%
Totals	106,535	4849
Annual trend pvalue	.614	.515	.625

FIGURE 1.

DHF visits and treatment in Medicare population.

There were 4929 MCF coded in the Humana database between 2007 and 2014 for patients aged 15 to 59 years (Table 3). There was no significant change in annual volume of MCF coded (p = .078) (Fig. 2). During this time period, 1558 of these MCF were treated with ORIF (Table 3). Among this cohort of clavicle fracture patients, there was a significant trend increase in annual volume (p = .002) and percentage of ORIF used for definitive treatment since publication of the level 1 study in 2007 (p = .003) (Fig. 2). Moreover, a strong positive correlation exists between total annual MCF and the use of ORIF as definitive treatment (r = .853, p = .007).

TABLE 3.

Incidence of midshaft clavicle fracture and ORIF use

Year	Incidence of MCF	ORIF Volume as Treatment	Percentage Use of ORIF as Definitive Management
2007	528	100	18.94%
2008	608	149	24.51%
2009	632	201	31.80%
2010	548	172	31.39%
2011	582	206	35.40%
2012	649	232	35.75%
2013	588	208	35.37%
2014	794	290	36.52%
Totals	4929	1558
Annual trend pvalue	.078	.002	.003

FIGURE 2.

Midshaft clavicle fractures for ages 15 to 59.

Discussion

Over the past few decades, there has been a strong emphasis on level 1 studies in the orthopaedic literature. Evidence-based treatment is considered to be best guided by these studies. Unfortunately, for upper extremity fractures, there remains a paucity of level 1 data available. Furthermore, the influence of level 1 studies on practice trends remains unclear. Results of this study clearly show a significant increase in use of ORIF in the treatment of MCF following the publication of supporting level 1 evidence. However, for intra-articular DHF, there was no change in the use of TEA despite level 1 evidence demonstrating superiority of TEA over ORIF for intraarticular DHF in the elderly. These findings suggest that level 1 evidence may not influence treatment patterns for particular fractures.

Clavicle

MCF remain one of the more common upper extremity fractures (6) and represent 80% of clavicle fractures (7). These fractures commonly occur in younger, active patients. Historically, these fractures have been treated nonoperatively because early low-level evidence suggested nonunions were rare (<1%) and malunions of little clinical importance (7, 8). However, several series have since shown the incidence of nonunions to be more prevalent than previously thought (1, 9–14) and malunions to have clinical significance (1, 15–18). In 2007, the Canadian Orthopaedic Trauma Society published a level 1 prospective, multicenter, randomized clinical trial contrasting management of displaced MCF (1). In their review of 132 patients between the ages of 16 and 60 years, they compared functional outcomes, complications, and radiographic findings between nonoperative management and open reduction with plate fixation for this indication. They found significant improvements in the Constant and the Disabilities of the Arm, Shoulder, and Hand (DASH) scores in the operative group at all time intervals. Furthermore, they demonstrated a significantly faster time to union (16.4 vs. 24.8 weeks), lower incidence of nonunions (two vs. four), and less symptomatic malunions (zero vs. nine) in the operative group. Subjective findings also favored operative fixation as 84% were satisfied with the appearance of the shoulder, contrasted to 53% in the non-operative group. These findings demonstrated superiority with open reduction and plate fixation compared to non-operative management of displaced MCF. These findings supported several previous studies (13, 19, 20) and have since been validated by other prospective randomized trials (21). A meta-analysis of the original and subsequent level 1 studies comparing ORIF to nonoperative treatment of displaced fractures found a more rapid return of function, lower complication rate, and less early residual disability following ORIF (22).

Despite the findings by COTS, several considerations may influence treatment selection in favor of nonoperative management of displaced MCF. The age of the patient may influence this decision, as these reported findings may not be anticipated in younger patients (6, 23). Additionally, the cost of open reduction and plate fixation has been shown to be significantly greater (21), and the risks of infection, hardware-related problems, and neurovascular injury may influence management as well (19, 24–29). Moreover, a reoperation rate as high as 24.6% has been reported after clavicle ORIF (30). With these considerations, as well as several studies demonstrating that correction of clavicular malunions may result in predictable satisfactory outcomes (15, 17), a “wait and see” approach has been favored by some, reserving surgical management if symptomatic malunion ensues (24). Also, it has been shown that no difference in DASH score, strength, or patient satisfaction was observed among those who underwent early surgical intervention compared with those treated surgically following symptomatic malunions and nonunions (31). Conversely, it has been shown that early surgical intervention results in higher Constant scores and greater shoulder flexion than delayed surgical management in the setting of symptomatic malunions and nonunions (31), and studies have shown low patient satisfaction and high complication rates following nonsurgical treatment of clavicle fractures (11, 32, 33). A relative risk reduction of 86% of nonunion after primary plate fixation compared with nonoperative management has since been demonstrated (13). The conflicting considerations of these influences, and that of the level 1 evidence provided by COTS, merited the importance of the current trend-monitoring study for this indication.

Since the COTS’s level 1 publication (1), several studies have observed a recent increase in use of ORIF in clavicle fracture managements (2–4, 24, 34). Navarro et al. recently published a review of the American Board of Orthopaedic Surgery database between 1999 and 2010 and showed a significant increase in ORIF use for clavicle fractures over this time (2). They suggested that the increase in open treatment of clavicle fractures might be influenced by clinical trials suggesting superior outcomes with ORIF. In a review of the Finnish National Hospital Discharge Register for adults greater than 18 years of age, Huttunen et al. reported a ninefold increase in surgical management of clavicle fractures between 1987 and 2010, from 1.3 to 10.8 per 100,000 persons per year (4). Yang et al. recently performed a PearlDiver review of nonoperative and operative trends in management of middle-third clavicle fractures in U.S. adolescents between 2007 and 2011 (3). They showed a significant increase in operative management among patients aged 10 to19 years during this period; 15- to 19-year-olds were more likely to be managed operatively than 10- to 14-year-olds (p < .001). The current study expands the support of these findings to a generalized U.S. population by using the age criteria described in the COTS study (16 to 60 years) and an analysis of a large managed care database. The findings support the trend of a significantly increased use of ORIF (p = .002) since the first publication of level 1 evidence in support of ORIF of displaced MCF.

Distal Humerus

In 2009, McKee at al. performed a level 1, prospective, randomized, double-blind, controlled trial of ORIF versus TEA for displaced intra-articular DHF in the elderly (>65 years) (5). They compared functional outcomes, complications, and reoperation rates between groups. The TEA group demonstrated superiority in the DASH and Mayo Elbow Performance Scores (MEPS), as well as a 32-minute reduction in operating time. Compared with the ORIF group, the TEA group also showed a nonsignificant improvement in the flexion–extension arc of motion and a lower reoperation rate. The study concluded that for intra-articular comminuted fractures of the distal humerus in the elderly, TEA had a more predictable outcome with improved 2-year functional scores compared with ORIF. This study supported the earlier retrospective findings of Frankle et al., which showed that all outcomes in the TEA group were either excellent (92%) or good (8%), compared with the ORIF group of 33% and 33%, respectively (35).

Despite level 1 evidence demonstrating superiority of TEA over ORIF in managing intra-articular DHF in the elderly, several considerations may influence decision making in favor of ORIF. First, TEA is often considered a technically demanding operation performed confidently by only a select group of subspecialists. Thus not all surgeons are comfortable performing this operation. Reported complications following TEA include infection, aseptic loosening, wear, periprosthetic fractures, mechanical failure, extensor mechanism dysfunction, and ulnar neuropathy (36–39), and an association between complications and surgeon experience has been demonstrated (40–42). Surgeons may feel more uncomfortable managing the complications associated with TEA than managing complications associated with ORIF. In a systematic review of the literature from 1993 to 2009, Voloshin et al. showed that the complication rate related to TEA for acute DHF was 22%(39). Additionally, selection of TEA is greatly influenced by the age of the patient, as younger patients typically have better bone quality, higher functional demands, and a greater likelihood of requiring revision TEA (43, 44). ORIF is thus typically selected for younger patients with higher functional demands (5, 45). Given the increasing functional activities demanded by patients over 65, selection of ORIF over TEA may be more influenced by functional expectations than age alone. Previous studies have shown ORIF to be an effective treatment modality even in the elderly osteoporotic population with comminuted, intra-articular DHF (46). Nevertheless, ORIF in the elderly tends to have less predictable outcomes secondary to the prevalence of osteoporosis, metaphyseal comminution, and poor soft tissue conditions often seen in this subset of patients (47–52). TEA remains a well-established alternative for these patients (5, 53–61), even in the setting of an open fracture (5, 62). The lack of change in use of TEA for DHF reported in these findings is not surprising given the conflicting evidence supporting one treatment modality over another for these fractures.

No study to date has reported on TEA trends following DHF for this patient population in the United States. Previous studies have shown an increase in overall TEA prevalence in recent years (42, 63), with TEA trends increasing for associated trauma (40, 42). However, no study has examined the TEA trends in a large subgroup of patients such as the Medicare population. This review of the Medicare database for patients greater than 65 years of age found no significant shift in annual TEA use for DHF (p = .515). Moreover, there were no significant trend differences in the annual use of TEA for DHF in the Medicare population as a whole between 2005 and 2009 compared with 2010 to 2012 (p = .602). This finding illustrates that no shift in surgical management occurred following the 2009 level 1 publication that demonstrated improved outcomes with TEA (5).

This study is not without limitations. First, retrospective bias may exist despite the data being collected prospectively. Additionally, the accuracy of the data is contingent on proper coding practices. Also, selection bias may have been introduced when patients entered or exited the Medicare and Humana databases during the study period. Because of the retrospective nature of this article, the authors can only demonstrate association, not causation, and speculate whether any shift in trends before or following the publication of level 1 studies are influential or coincidental. Last, because of the Humana data in the PearlDiver data set, data before 2007 are unavailable. Thus this study is unable to evaluate trends in the treatment of MCF before the COTS 2007 level 1 publication (1). The major strength of this study relates to the completeness of the data. The PearlDiver database represents complete data for the entire Medicare and Humana population.