ORIGINAL ARTICLE
Wound Complications Associated With Bone Morphogenetic Protein-2 in Orthopaedic Trauma Surgery Daniel S. Chan, MD,*† Joshua Garland, MD,† Anthony Infante, DO,* Roy W. Sanders, MD,* and H. Claude Sagi, MD*
Objective: To document the incidence of postoperative wound complications associated with the use of rhBMP-2 in a large series of patients for both acute traumatic and reconstructive extremity cases.
Design: Retrospective chart and radiographic review. Setting: Level I trauma center. Methods: A retrospective chart and x-ray review was performed on cases between 2002 and 2009, in which rhBMP-2 (Infuse) was used in acute trauma or posttraumatic reconstruction. The following data points were collected: age, surgical site, purpose (acute vs. reconstructive), associated wound factors (open fractures, soft tissue injury requiring coverage, or history of infection), signs of infection (seroma, erythema, prolonged drainage, abscess), reoperation rate secondary to wound complication, culture results, and union. These cases were then compared with a matched cohort without the use of bone morphogenetic protein-2 (matched for age, type of case, anatomic site, and open injury) for statistical analysis.
Conclusions: The use of rhBMP-2 in both acute traumatic and posttraumatic reconstructive extremity surgery may increase the incidence of prolonged postoperative serous wound drainage. However, this does not seem to correlate with an increased incidence of postoperative wound infection or the need for reoperation. The use of rhBMP-2 seems to have a beneficial effect in improving union rates for both acute trauma and posttraumatic reconstruction of the extremities (P = 0.002); however, this and the mechanism for prolonged serous drainage require further study before definitive recommendations can be made. Key Words: bone morphogenetic protein, wound complications, drainage, Infuse, nonunion, open fracture
Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence. (J Orthop Trauma 2014;28:599–604)
Results: Group 1 was comprised a total of 193 patients whose
INTRODUCTION
treatment included rhBMP-2 (155 reconstructive and 38 acute open fractures). Group 2 was comprised 181 patients treated without the use of rhBMP-2 (145 reconstructive and 36 acute open fractures). The incidences of documented wound complications were 31% (60/ 193) in group 1 and 18% (33/181) in group 2 (P = 0.004). Reoperation rates for wound complications were in 3.1% of group 1 and 8.3% of group 2 (P = 0.04). Age, sex, anatomic site, acute trauma, open fracture, and the need for soft tissue reconstruction did not correlate with the need to return to the operating room for presumed or actual wound infection. The rates of union between rhBMP-2 and control groups were 90% versus 74% (P , 0.001); for acute trauma cases, 94% versus 79% (P = 0.220); and for reconstructive cases, 89% versus 73% (P = 0.002).
Since the discovery of bone morphogenetic protein (BMP) by Urist in 1965,1 BMPs have further been identified to be a part of the transforming growth factor-b family,2 and a subset of these growth factor proteins has been recognized to induce bone formation and remodeling.3 Two recombinant BMPs are available for clinical use: rhBMP-2 (Infuse; Medtronic Sofamor Danek, Memphis, TN) and rhBMP-7 (OP-1; Olympus Biotech, Hopkinton, MA). In orthopaedic trauma, the use of Infuse is currently approved by the US Food and Drug Administration (FDA) for treating acute open tibial shaft fractures that have been stabilized with intramedullary nail fixation after appropriate wound management. OP-1 has received a humanitarian Investigational Device Exemption for the treatment of long bone nonunions. It has been reported that BMP is predominantly used in spinal fusions, and most of these procedures are for off-label uses.4 Indeed, reports in the literature of complications associated with the use of BMP are limited to its use in the spine. The inflammatory response and soft tissue edema secondary to BMP use in anterior cervical fusions and subsequent dysphagia, dysphonia, and respiratory distress has been well documented.5 The off-label use of BMP in orthopaedic trauma and the incidence of associated perioperative complications have not been previously reported. The purpose of this study was to document the incidence of union and postoperative wound complications that are thought to be directly attributable to the
Accepted for publication March 20, 2014. From the *Orthopaedic Trauma Service, Florida Orthopaedic Institute, Tampa, FL; †Naval Medical Center, Portsmouth, VA. D. S. Chan is a consultant for Biomet and Synthes. A. Infante is a consultant for OmniLife Sciences and Arthrosurface. H. C. Sagi is a consultant for Stryker. R. W. Sanders is a consultant for Smith & Nephew, Stryker, DePuy, and Linvatec. J. Garland reports no conflict of interest. Presented in part at the Annual Meeting of the Orthopaedic Trauma Association, October 15, 2011, San Antonio, TX. Reprints: Daniel S. Chan, MD, Orthopaedic Trauma Service, Florida Orthopaedic Institute, 5 Tampa General Circle, Suite 710, Tampa, FL 33606 (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
www.jorthotrauma.com |
599
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
use of rhBMP-2 in a large series of patients for both acute traumatic and posttraumatic reconstructive extremity cases.
MATERIALS AND METHODS After obtaining institutional review board approval, a retrospective chart review was performed on cases between 2002 and 2009 at a single level-1 regional referral trauma center. Group 1 inclusion criteria were skeletal maturity, and use of rhBMP-2 for acute upper or lower extremity trauma, or posttraumatic reconstruction for nonunion or posttraumatic arthritis of the upper or lower extremity. Exclusion criteria were skeletal immaturity, an active soft tissue infection, acute or chronic osteomyelitis, and any case with incomplete medical records. The following data were collected: patient age, surgical site, purpose (acute trauma vs. reconstructive), associated wound factors (open fractures, soft tissue injury requiring coverage, or history of infection), postoperative wound complications (seroma, erythema, prolonged drainage, dehiscence, abscess), antibiotic administration, reoperation rates, culture results, and union. Surgical site categories were broken up by anatomic location (Table 1).
Group 1 For the acute trauma cases, all were open fractures where rhBMP-2 was applied after the final debridement and immediately before definitive wound closure during the initial hospitalization. Reconstructive cases were comprised either diaphyseal nonunions or primary or revision posttraumatic foot and ankle fusions where rhBMP-2 was applied in an offlabel fashion to enhance nonunion repair or arthrodesis, respectively. Inpatient and outpatient medical records were closely reviewed for the documentation of any postoperative wound complications. If there was any mention of “seroma,” “abscess,” “erythema,” “drainage,” or “dehiscence” in the postoperative physical examination of any inpatient progress note or outpatient clinic note, a postoperative wound TABLE 1. Group Demographics Group 1—BMP (193 Patients) Mean age (range), y Traumatic cases Reconstructive cases Acute open fractures Required soft tissue reconstruction Upper extremity Proximal femur Femoral shaft Distal femur Proximal tibia Tibial shaft Distal tibia Foot and ankle
600
52 (16–90) 38 (20%) 155 (80%) 34 (18%) 18 (9%) (13 acute trauma) (5 reconstructive) 22 (11%) 6 (3%) 18 (9%) 18 (9%) 3 (2%) 36 (19%) 24 (13%) 66 (35%)
| www.jorthotrauma.com
Group 2—Control (181 Patients) 51 (19–85) 36 (20%) 145 (80%) 32 (15%) 29 (16%) (18 acute trauma) (11 reconstructive) 16 (9%) 4 (2%) 19 (10%) 14 (8%) 4 (2%) 36 (20%) 26 (14%) 62 (34%)
complication was recorded. Any administration of antibiotics given outside the standard 24-hour postoperative course, or 48 hours for open fractures, was considered prolonged antibiotic use. Reoperation rates and culture results from debridement procedures for postoperative wound complications were also recorded. Patient outcome end points were reoperation for wound complication, union, nonunion, or lost to follow-up. Union was determined through review of the patient clinic notes where it was distinctly described both radiographically (minimum 3 cortices of bridging bone on orthogonal x-rays) and clinically (no pain at fracture site with the use of the extremity). All x-rays were reviewed by a fellowship-trained orthopaedic trauma attending. The absence of healing, or any suggestion of a delay in union was considered to be a nonunion, and nonunion type was classified. If a patient failed to return for a scheduled clinic appointment toward ongoing treatment for more than 3 months, they were considered lost to follow-up and excluded in the analysis for union. However, these patients were included in the analysis for acute perioperative wound complications.
Group 2 The institutional orthopaedic trauma registry database was used to construct a matched control cohort of patients treated without the use of rhBMP-2 during the same time period at the same institution. These cases were matched for age, injury/surgical site, acute trauma/reconstructive surgical procedure, and history of open fracture or infection. There were 181 control patients matched to the rhBMP-2 patients. Statistical comparison confirmed no difference between the 2 groups for each matched characteristic and soft tissue reconstruction and nonunion type. The same data as in group 1 were collected. A power analysis was not performed because of the limitations of available patients with matched characteristics. Additionally, the expected difference in outcome for the off-label use of BMP in reconstruction did not enable a proper power analysis. As a retrospective chart review, the use of rhBMP-2 was not randomized. The decision to use rhBMP-2 was determined strictly by the treating surgeon. Incidence rates were reported as percentages. A statistical analysis was performed using the Fisher exact and x2 tests for categorical variables. Matched group characteristics between the groups and the compared subsets did not necessitate multivariate regression; univariate analysis was sufficient. Data analysis was processed using SPSS Statistics (SPSS Inc, Chicago, IL). Statistical significance was set for any P value less than 0.05. Analyzing for correlation between surgical sites and outcomes was assisted by grouping the 3 areas of a long bone (proximal, shaft, and distal) as 1 entity (upper extremity, femur, tibia, or foot/ankle) to provide enhanced power.
RESULTS Group 1: Cases (With rhBMP-2) A total of 193 cases in 193 patients were included in the analysis. The average age was 52 years (range, 16–90 years). Ó 2014 Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
Wound Complications in Orthopaedic Trauma Surgery
rhBMP-2 was used in 38 acute trauma patient and 155 reconstructive patients. Of the acute traumatic cases, 34 were open fractures, and 13 of these 34 patients (34%) required some form of soft tissue reconstruction (Table 2). The anatomic site breakdown is summarized in Table 1. Thirty-six patients of the 155 reconstructive cases (23%) had history of associated open fracture, more than half (19/36) were open tibia fractures. Twenty-eight patients (18%) had history of infection, in which 10 involved the tibia and 9 the foot and ankle. Seventeen patients (11%) had history of both an open fracture and infection, most (8/17) were open tibia fractures. The 5 patients (3%) who had history of soft tissue defects requiring reconstruction also had tibia fractures (3/5 secondary to open fractures). An anatomic breakdown of these associated factors of reconstructive cases is included in Table 2. All 155 reconstructive cases underwent radiographic review. There were 45 infected nonunion, 4 malunion, and 26 primary arthrodesis cases. The 80 remaining cases were comprised 27 atrophic, 49 oligotrophic, and 4 hypertrophic nonunions. Anatomic site breakdown of reconstructive cases is included in Table 3.
Group 2: Matched Controls (No rhBMP-2) A total of 181 cases (181 patients) without the use of rhBMP-2 matched for age, anatomic site, acute trauma/ reconstructive surgical procedure, and history of open fracture or infection were included in the analysis. The average age of the patients was 51 years (range, 19–85 years) (Table 1). Of the 36 acute traumatic cases, 32 were open fractures, and 18 required some form of soft tissue reconstruction. Thirty-eight patients of the 145 reconstructive cases had history of associated open fracture, more than half (20/38) were open tibia fractures (Table 2). Twenty-eight patients had history of infection, in which 12 involved the tibia and 7 the foot and ankle. Seventeen patients had history of both open fracture and infection, most (8/17) were open tibia fractures. The 12 patients who had history of soft tissue defects requiring reconstruction were predominantly tibia fractures (10/12 secondary to open fractures). Anatomic breakdown of these associated factors of the reconstructive cases is included in Table 2. All 145 reconstructive cases underwent radiographic review. There were 43 infected nonunion, 3 malunion, and 16 primary arthrodesis cases. The 83 remaining cases were
comprised 34 atrophic, 43 oligotrophic, and 6 hypertrophic nonunions. Anatomic site breakdown of the reconstructive cases is included in Table 3. The difference in frequencies of primary fusion, malunion, and nonunion type between acute and reconstructive patients in each group was not statistically significant (Table 3).
Group Comparison and Statistical Analysis Overall, 60 patients (31%) in the BMP group and 33 patients (18%) in the control group had documentation of at least 1 postoperative wound complication after the index procedure (P = 0.004) (Table 4). Among those patients with a wound complication, the most commonly documented was serous drainage [56/60 (93%) and 22/33 (67%), respectively]. The incidence of wound complications between acute and reconstructive patients in each group was not statistically significant (26% vs. 32% for BMP and 11% vs. 20% for controls). Prolonged antibiotic therapy secondary to a postoperative wound complication was documented in 17 patients (9%) in group 1 and 23 patients (13%) in group 2. In group 1, 1 acute trauma and 5 reconstructive patients had documentation of at least 1 reoperation for a wound complication. In group 2, 2 acute trauma and 13 reconstructive patients underwent reoperation for a wound complication (P = 0.04). Overall, the proportion of patients who returned to the operating room for debridement was 10% (6/60) in group 1 versus 46% (15/33) in group 2 (P , 0.001). Within the BMP group, 4 of 6 patients (2%) had a documented infection based on intraoperative cultures, whereas 2 patients (1%) had a sterile seroma/hematoma. Only 1 BMP patient in the acute trauma subgroup with an open tibia fracture and soft tissue reconstruction had a positive intraoperative culture. The other 3 patients (2 femoral and 1 subtalar nonunions) were in the reconstructive subset treated with BMP. Each of the 4 infections occurred with a different organism. In group 2, 9 of 15 cases had positive intraoperative cultures. Similar to group 1, one patient with an open tibia fracture that required soft tissue reconstruction had a positive intraoperative culture. The 8 in the reconstructive subgroup were comprised 6 tibial, 1 femoral, and 1 subtalar nonunion. No one organism was more common than another.
TABLE 2. Description of Reconstructive Cases Group 1—BMP Acute Open fracture Soft tissue reconstruction Reconstructive Open fractures Infection Open fracture and infection Soft tissue reconstruction
Group 2—Control
UE
F
T
F&A
Total
UE
F
T
F&A
Total
3 2 1 19 3 2 2 —
7 3 — 35 9 7 5 —
27 27 12 36 19 10 7 5
1 1 — 65 5 9 3 —
38 33 13 155 36 28 17 5
2 1 1 14 5 0 3 —
6 3 1 31 6 6 2 —
27 27 16 39 20 12 8 10
1 1 — 61 7 7 5 1
37 32 18 145 38 25 18 11
F, femur; F&A, foot and ankle; T, tibia; UE, upper extremity.
Ó 2014 Lippincott Williams & Wilkins
www.jorthotrauma.com |
601
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
TABLE 3. Breakdown of Reconstructive Cases Group 1—BMP (155 Patients) Primary fusions Malunion Infected nonunion Atrophic nonunion Oligotrophic nonunion Hypertrophic nonunion Total
Group 2—Control (145 Patients)
UE
F
T
F&A
Total*
UE
F
T
F&A
Total*
— 2 4 6 7 — 19
— 1 12 5 14 3 35
— 1 17 6 11 1 36
26 — 12 10 17 — 65
26 4 45 27 49 4 155
— 1 3 8 2 — 14
— — 8 5 13 5 31
— 2 20 8 8 1 39
16 — 12 13 20 — 61
16 3 43 34 43 6 145
*There were no statistically significant differences in the subgroup totals between group 1 and group 2. F, femur; F&A, foot and ankle; T, tibia; UE, upper extremity.
Union Twenty of 193 patients (10%) were lost to follow-up in group 1 (7 acute and 13 reconstructive). Similarly, 26/181 patients (14.4%) were lost to follow-up in group 2 (12 acute and 14 reconstructive). For the 173 patients in group 1, union was achieved in 94% of acute and 89% of reconstructive cases. For the 155 patients in group 2, union occurred in 79% of acute and 73% of reconstructive cases. The differences in union rates between groups 1 and 2 for overall cases and for the reconstructive cases were statistically significant [90% vs. 74% (P , 0.001); 89% vs. 73% (P = 0.002), respectively] (Table 4). The union rate among acute cases was higher in group 1 than that in group 2 (94% vs. 79%), but this difference failed to reach statistical significance (P = 0.220). There were no other significant differences in union rates when comparing the 2 groups for anatomic site, need for soft tissue coverage, or history of infection. Among those patients with wound complications, union rates also differed significantly between the groups (group 1: 50/53, group 2: 18/31, P , 0.001).
TABLE 4. Summary of Results Group 1—BMP Group 2—Control (193 Patients) (181 Patients) Overall wound complications Wound concern—acute trauma Wound concern—reconstructive Overall prolonged antibiotic use Prolonged antibiotic use— acute trauma Prolonged antibiotic use— reconstructive Overall reoperation for wound concern Reoperation for wound concern— acute trauma Reoperation for wound concern— reconstructive Overall union rate Union rate for acute trauma Union rate for reconstructive
602
| www.jorthotrauma.com
60 (31%) 10 (5%) 50 (26%) 17 (9%) 2 (1%)
33 (18%), P = 0.004 4 (2%) 29 (16%) 23 (13%) 3 (2%)
15 (8%)
20 (11%)
6 (3%)
15 (8%), P = 0.04
1 (0.5%)
2 (1%)
5 (3%)
13 (7%), P = 0.05
90% 94% 89%
74%, P , 0.001 79%, P = 0.220 73%, P = 0.002
Finally, age, sex, acute trauma, reconstructive surgery, history of an open fracture or infection, or the need for a soft tissue reconstruction did not correlate with the development of a wound complication or a return to the operating room.
DISCUSSION Persistent postoperative wound drainage associated with the use of rhBMP-2 has been reported mostly in the spine literature.5 In our series, there was not an increased incidence of postoperative wound infection with the use of rhBMP-2; however, rhBMP-2 does seem to have a beneficial effect in improving union rates overall. A BMP-induced immune-modulated inflammatory response is the proposed mechanism for the observed wound drainage and dehiscence.6,7 Although 31% of our patients in group 1 had a wound complication, most commonly serous drainage, observation alone was successful 72% (43/60) of the time. It has been shown that the concentration of rhBMP2, and length of time at the implant site, directly affects the amount of bone formed.8 Conversely, if local or implant site retention is low and rhBMP-2 leaks into the surrounding tissue, an inflammatory response might be more likely. Similarly, it is possible that antibodies to rhBMP-2 or to the bovine collagen carrier are produced by the “host” and induce a physiologic inflammatory reaction. An immune-modulated inflammatory response to the protein moiety and or carrier for BMP-2 is, at present, the best explanation for these observations; however, the exact mechanism and the clinical relevance are unknown.9 The rhBMP-2 Evaluation in Surgery for Tibial Trauma (BESTT) Study Group10 conducted a large European multinational randomized control trial, which showed that open tibial fractures treated with intramedullary nails and 1.50 mg/mL of rhBMP-2 at final closure had fewer implant failures, faster wound healing, and fewer infections in Gustilo– Anderson type-III injuries. The authors stated that the adverse events observed (eg, edema, pain, infection) “were consistent with those normally observed in the trauma setting.” Anecdotally, the wound complications most often associated with the use of BMP’s are characterized by drainage, erythema, and edema; nonspecific signs that are also indicative of inflammation commonly seen after trauma and Ó 2014 Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
prolonged reconstructive procedures. This observation was reported by Jones et al,11 where 33% (5/15) of the diaphyseal tibia fractures treated with freeze-dried cancellous allograft and rhBMP-2 developed postoperative erythema compared with 0% (also of 15) in the group treated with autograft only. This incidence reported by Jones et al is in agreement with our data. Similar to our findings, these wound complications were typically self-limited and resolved without treatment. Although we did observe an increased incidence of prolonged serous drainage, we also noted a decreased reoperation rate for wound complications in group 1. Although the presence or absence of a specific wound complication can easily be documented, the quality or severity of these concerns is more difficult to accurately quantify. This study relies on an accurate documentation as outlined in the Methods. We observed that the progress notes were appropriately detailed, including the pertinent positive and negatives, which suggests that a thorough assessment and documentation of a potential postoperative wound complication was performed. A reporting bias and a clinical decisionmaking bias may however exist; because serous drainage related to rhBMP-2 became a more familiar entity, its presence was more frequently observed and documented, and its management with observation alone was more frequently practiced. However, if this were in fact true, we believe that there would also be a higher number of reoperations with negative cultures in the control group, which was not the case [4/6 (67%) rhBMP-2 reoperations with proven infection vs. 9/15 (60%) in the control group]. Data such as reoperation, culture results, and antibiotic administration are concrete end points; based on the equal proportion of reoperations for proven infection, the decisionmaking algorithm for operative intervention for a wound complication were the same for both groups. Although the incidence of postoperative wound drainage occurred more frequently in the rhBMP-2 group, the need for reoperation (and the incidence of proven infection) was more frequent in the control group. Reasons for this are unclear; however, the findings are in agreement with the results of the BESST study,10 where the need for reoperation and infection were also decreased in the group of patients treated with rhBMP-2. Postoperative infection rates have been reported extensively, especially in the antibiotic prophylaxis literature that conclusively shows the efficacy of preoperative antibiotics. Boxma et al12 reported a postoperative infection rate of 3.6% for more than 1000 closed fracture cases that received preoperative antibiotics. Similar results were found by Patzakis et al13 in their study of open fractures treated with preoperative cephalothin with an infection rate of 2.3%. More recently, Williams et al14 reported a similar wound infection rate (2.4%) in an analysis of 55,862 cases of spinal fusion; this rate was the same for both cases with and without BMP. In our study, we observed a 2% infection rate (based on positive intraoperative cultures in cases in which rhBMP-2 was used), and this is consistent with the above reported studies. With regard to union in acute trauma, the findings of this study are in agreement with those of BESST.8 The Ó 2014 Lippincott Williams & Wilkins
Wound Complications in Orthopaedic Trauma Surgery
present analysis demonstrated an improved union rate (94% vs. 79%) for the acute trauma cases whose treatment included rhBMP-2. This is an important finding because it is contradictory to a recent report that found no difference with rhBMP-2 use in open tibia fractures.15 Furthermore, this effect (not previously noted in the literature) was also found in the reconstructive cases (nonunion and posttraumatic arthritis), whereby those patients whose treatment included rhBMP-2 demonstrated an increased incidence of clinical and radiographic union (89% vs. 73%, P = 0.002). Additional studies are needed to more closely evaluate and report the benefit of rhBMP-2 in arthrodesis cases. One of the weaknesses of this analysis is the inclusion of multiple surgical sites in each group, inducing type 1 error and not allowing for complete homogeneity for comparison. Stratifying patients into smaller groups for subgroup analysis inherently diminished the power for a meaningful comparison. We believe this to be the case with regard to the significant differences observed in the infection rates between femoral and foot/ankle cases, as well as the observed differences in prolonged antibiotic administration and postoperative complications between tibial and foot/ankle cases. However, despite the heterogeneity, the comparison with a 1:1 matched cohort (age, injury site, acute trauma vs. reconstructive surgery, history of open fracture or infection, and soft tissue coverage) does address the pooling of subgroups. This means that there were equivalent number of upper and lower extremity injuries, acute fractures, nonunions, and arthrodeses, and sterile versus previously infected cases. Similarly, treatment methods or approaches are also likely to be comparable between cohorts (ie, nails vs. plates or early total care vs. staged treatment). Therefore, we believe that the associated differences with each of these subsets with regard to treatment, healing, or complications are the same when comparing a matched cohort. In this study, the surgeon chose to use rhBMP-2 independently. The decision-making process cannot accurately be determined from a retrospective chart review; however, comparing 2 groups of open fractures matched to age, fracture site, and need for soft tissue reconstruction does enable us to compare similar severity of injuries (ie, grade of open fracture, role for rhBMP-2). Similarly, the decision to use rhBMP-2 to repair a nonunion cannot consistently be determined by chart review; however, the characteristics of the 2 subsets (history of open fracture, infection, or need for soft tissue reconstruction) are matched as mentioned above, suggesting that the state of the “biology” of the nonunion also were equivalent. Similar nonunion type categorization supports the equivalence of these matched groups, suggests proper use of BMP in group 1 (95% oligotrophic or atrophic nonunions), and enhances the union results and conclusions. The strengths of this study include the large volume of cases performed at a single institution over a short period of time, providing adequate power to show differences between 2 similar groups matched for case type, anatomic region, and patient demographics. The choice of concrete end points, such as reoperation and union, allows for accurate and reliable results. Data collected from descriptions of postoperative surgical wounds within progress notes were also accurate www.jorthotrauma.com |
603
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
because of the fact that prolonged drainage is a significant event, it lengthens hospital stays, and/or is the reason for prolonged or additional inpatient antibiotic administration.
CONCLUSIONS
This is the first study, to the best of our knowledge, specifically investigating the incidence of wound complications associated with the use of rhBMP-2 in orthopaedic trauma. We found that in both acute and reconstructive extremity trauma, rhBMP-2 does seem to result in an increased incidence of postoperative serous wound drainage. However, this does not seem to correlate with an increased incidence of postoperative wound infection or the need for reoperation. Looking forward, the mechanism for prolonged serous drainage can be better delineated and definitive recommendations for its management. Equally important, and also not previously reported, is our finding that the use of rhBMP-2 improves union rates for both acute trauma and posttraumatic reconstruction of the extremities. Further investigation is needed in rhBMP-2 cases with minimal heterogeneity and specific focus on its use in posttraumatic arthrodesis.
5. 6. 7. 8. 9. 10.
11.
12.
REFERENCES 1. Urist MR. Bone: formation by autoinduction. Science. 1965;150: 893–899. 2. Kingsley DM. The TGF-beta superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 1994;8:133–146. 3. Cheng H, Jiang W, Phillips FM, et al. Osteogenic activity of the fourteen types of human bone morphogenetic proteins (BMPs). J Bone Joint Surg Am. 2003;85-A:1544–1552. 4. Ong KL, Lau E, Villarraga ML, et al. Evaluating on-label and off-label use of BMP in the U.S. using administrative data. Scientific Exhibit No.
604
| www.jorthotrauma.com
13. 14.
15.
71, 77th Annual Meeting of the American Academy of Orthopaedic Surgeons, New Orleans, LA, March 9–13, 2010. Owens J, Glassman SD, Howard JM, et al. Perioperative complications with rhBMP-2 in transforaminal lumbar interbody fusion. Eur Spine J. 2011;20:612–617. Vaidya R, Carp J, Sethi A, et al. Complications of anterior cervical discectomy and fusion using recombinant human bone morphogenetic protein-2. Eur Spine J. 2007;16:1257–1265. Wang EA, Rosen V, D’Alessandro JS, et al. Recombinant human bone morphogenetic protein induces bone formation. Proc Natl Acad Sci U S A. 1990;87:2220–2224. Blum B, Moseley J, Miller L, et al. Measurement of bone morphogenetic proteins and other growth factors in demineralized bone matrix. Orthopedics. 2004;27(suppl):S161–S165. Szpalski M, Gunzburg R. Recombinant human bone morphogenetic protein-2: a novel osteoinductive alternative to autogenous bone graft? Acta Orthop Belg. 2005;71:133–148. Govender S, Csimma C, Genant HK, et al. Recombinant human bone morphogenetic protein-2 for treatment of open tibial fractures: a prospective, controlled, randomized study of four hundred and fifty patients. J Bone Joint Surg Am. 2002;84:2123–2134. Jones AL, Bucholz RW, Bosse MJ, et al; BMP-2 Evaluation in Surgery for Tibial Trauma-Allgraft (BESTT-ALL) Study Group. Recombinant human BMP-2 and allograft compared with autogenous bone graft for reconstruction of diaphyseal tibial fractures with cortical defects. A randomized, controlled trial. J Bone Joint Surg Am. 2006;88: 1431–1441. Boxma H, Broekhuizen T, Patka P, et al. Randomised controlled trial of single-dose antibiotic prophylaxis in surgical treatment of closed fractures: the Dutch Trauma Trial. Lancet. 1996;347:1133–1137. Patzakis MJ, Harvey JP Jr, Ivler D. The role of antibiotics in the management of open fractures. J Bone Joint Surg Am. 1974;56:532–541. Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? A comparison of 55,862 cases of spinal fusion with and without bone morphogenetic protein. Spine 2011;36:1685–1691. Aro HT, Govender S, Patel AD, et al. Recombinant human bone morphogenetic protein-2: a randomized trial in open tibial fractures treated with reamed nail fixation. J Bone Joint Surg Am. 2011;93: 801–808.
Ó 2014 Lippincott Williams & Wilkins
Wound Complications Associated With Bone Morphogenetic Protein-2 in Orthopaedic Trauma Surgery Daniel S. Chan, MD,*† Joshua Garland, MD,† Anthony Infante, DO,* Roy W. Sanders, MD,* and H. Claude Sagi, MD*
Objective: To document the incidence of postoperative wound complications associated with the use of rhBMP-2 in a large series of patients for both acute traumatic and reconstructive extremity cases.
Design: Retrospective chart and radiographic review. Setting: Level I trauma center. Methods: A retrospective chart and x-ray review was performed on cases between 2002 and 2009, in which rhBMP-2 (Infuse) was used in acute trauma or posttraumatic reconstruction. The following data points were collected: age, surgical site, purpose (acute vs. reconstructive), associated wound factors (open fractures, soft tissue injury requiring coverage, or history of infection), signs of infection (seroma, erythema, prolonged drainage, abscess), reoperation rate secondary to wound complication, culture results, and union. These cases were then compared with a matched cohort without the use of bone morphogenetic protein-2 (matched for age, type of case, anatomic site, and open injury) for statistical analysis.
Conclusions: The use of rhBMP-2 in both acute traumatic and posttraumatic reconstructive extremity surgery may increase the incidence of prolonged postoperative serous wound drainage. However, this does not seem to correlate with an increased incidence of postoperative wound infection or the need for reoperation. The use of rhBMP-2 seems to have a beneficial effect in improving union rates for both acute trauma and posttraumatic reconstruction of the extremities (P = 0.002); however, this and the mechanism for prolonged serous drainage require further study before definitive recommendations can be made. Key Words: bone morphogenetic protein, wound complications, drainage, Infuse, nonunion, open fracture
Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence. (J Orthop Trauma 2014;28:599–604)
Results: Group 1 was comprised a total of 193 patients whose
INTRODUCTION
treatment included rhBMP-2 (155 reconstructive and 38 acute open fractures). Group 2 was comprised 181 patients treated without the use of rhBMP-2 (145 reconstructive and 36 acute open fractures). The incidences of documented wound complications were 31% (60/ 193) in group 1 and 18% (33/181) in group 2 (P = 0.004). Reoperation rates for wound complications were in 3.1% of group 1 and 8.3% of group 2 (P = 0.04). Age, sex, anatomic site, acute trauma, open fracture, and the need for soft tissue reconstruction did not correlate with the need to return to the operating room for presumed or actual wound infection. The rates of union between rhBMP-2 and control groups were 90% versus 74% (P , 0.001); for acute trauma cases, 94% versus 79% (P = 0.220); and for reconstructive cases, 89% versus 73% (P = 0.002).
Since the discovery of bone morphogenetic protein (BMP) by Urist in 1965,1 BMPs have further been identified to be a part of the transforming growth factor-b family,2 and a subset of these growth factor proteins has been recognized to induce bone formation and remodeling.3 Two recombinant BMPs are available for clinical use: rhBMP-2 (Infuse; Medtronic Sofamor Danek, Memphis, TN) and rhBMP-7 (OP-1; Olympus Biotech, Hopkinton, MA). In orthopaedic trauma, the use of Infuse is currently approved by the US Food and Drug Administration (FDA) for treating acute open tibial shaft fractures that have been stabilized with intramedullary nail fixation after appropriate wound management. OP-1 has received a humanitarian Investigational Device Exemption for the treatment of long bone nonunions. It has been reported that BMP is predominantly used in spinal fusions, and most of these procedures are for off-label uses.4 Indeed, reports in the literature of complications associated with the use of BMP are limited to its use in the spine. The inflammatory response and soft tissue edema secondary to BMP use in anterior cervical fusions and subsequent dysphagia, dysphonia, and respiratory distress has been well documented.5 The off-label use of BMP in orthopaedic trauma and the incidence of associated perioperative complications have not been previously reported. The purpose of this study was to document the incidence of union and postoperative wound complications that are thought to be directly attributable to the
Accepted for publication March 20, 2014. From the *Orthopaedic Trauma Service, Florida Orthopaedic Institute, Tampa, FL; †Naval Medical Center, Portsmouth, VA. D. S. Chan is a consultant for Biomet and Synthes. A. Infante is a consultant for OmniLife Sciences and Arthrosurface. H. C. Sagi is a consultant for Stryker. R. W. Sanders is a consultant for Smith & Nephew, Stryker, DePuy, and Linvatec. J. Garland reports no conflict of interest. Presented in part at the Annual Meeting of the Orthopaedic Trauma Association, October 15, 2011, San Antonio, TX. Reprints: Daniel S. Chan, MD, Orthopaedic Trauma Service, Florida Orthopaedic Institute, 5 Tampa General Circle, Suite 710, Tampa, FL 33606 (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
www.jorthotrauma.com |
599
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
use of rhBMP-2 in a large series of patients for both acute traumatic and posttraumatic reconstructive extremity cases.
MATERIALS AND METHODS After obtaining institutional review board approval, a retrospective chart review was performed on cases between 2002 and 2009 at a single level-1 regional referral trauma center. Group 1 inclusion criteria were skeletal maturity, and use of rhBMP-2 for acute upper or lower extremity trauma, or posttraumatic reconstruction for nonunion or posttraumatic arthritis of the upper or lower extremity. Exclusion criteria were skeletal immaturity, an active soft tissue infection, acute or chronic osteomyelitis, and any case with incomplete medical records. The following data were collected: patient age, surgical site, purpose (acute trauma vs. reconstructive), associated wound factors (open fractures, soft tissue injury requiring coverage, or history of infection), postoperative wound complications (seroma, erythema, prolonged drainage, dehiscence, abscess), antibiotic administration, reoperation rates, culture results, and union. Surgical site categories were broken up by anatomic location (Table 1).
Group 1 For the acute trauma cases, all were open fractures where rhBMP-2 was applied after the final debridement and immediately before definitive wound closure during the initial hospitalization. Reconstructive cases were comprised either diaphyseal nonunions or primary or revision posttraumatic foot and ankle fusions where rhBMP-2 was applied in an offlabel fashion to enhance nonunion repair or arthrodesis, respectively. Inpatient and outpatient medical records were closely reviewed for the documentation of any postoperative wound complications. If there was any mention of “seroma,” “abscess,” “erythema,” “drainage,” or “dehiscence” in the postoperative physical examination of any inpatient progress note or outpatient clinic note, a postoperative wound TABLE 1. Group Demographics Group 1—BMP (193 Patients) Mean age (range), y Traumatic cases Reconstructive cases Acute open fractures Required soft tissue reconstruction Upper extremity Proximal femur Femoral shaft Distal femur Proximal tibia Tibial shaft Distal tibia Foot and ankle
600
52 (16–90) 38 (20%) 155 (80%) 34 (18%) 18 (9%) (13 acute trauma) (5 reconstructive) 22 (11%) 6 (3%) 18 (9%) 18 (9%) 3 (2%) 36 (19%) 24 (13%) 66 (35%)
| www.jorthotrauma.com
Group 2—Control (181 Patients) 51 (19–85) 36 (20%) 145 (80%) 32 (15%) 29 (16%) (18 acute trauma) (11 reconstructive) 16 (9%) 4 (2%) 19 (10%) 14 (8%) 4 (2%) 36 (20%) 26 (14%) 62 (34%)
complication was recorded. Any administration of antibiotics given outside the standard 24-hour postoperative course, or 48 hours for open fractures, was considered prolonged antibiotic use. Reoperation rates and culture results from debridement procedures for postoperative wound complications were also recorded. Patient outcome end points were reoperation for wound complication, union, nonunion, or lost to follow-up. Union was determined through review of the patient clinic notes where it was distinctly described both radiographically (minimum 3 cortices of bridging bone on orthogonal x-rays) and clinically (no pain at fracture site with the use of the extremity). All x-rays were reviewed by a fellowship-trained orthopaedic trauma attending. The absence of healing, or any suggestion of a delay in union was considered to be a nonunion, and nonunion type was classified. If a patient failed to return for a scheduled clinic appointment toward ongoing treatment for more than 3 months, they were considered lost to follow-up and excluded in the analysis for union. However, these patients were included in the analysis for acute perioperative wound complications.
Group 2 The institutional orthopaedic trauma registry database was used to construct a matched control cohort of patients treated without the use of rhBMP-2 during the same time period at the same institution. These cases were matched for age, injury/surgical site, acute trauma/reconstructive surgical procedure, and history of open fracture or infection. There were 181 control patients matched to the rhBMP-2 patients. Statistical comparison confirmed no difference between the 2 groups for each matched characteristic and soft tissue reconstruction and nonunion type. The same data as in group 1 were collected. A power analysis was not performed because of the limitations of available patients with matched characteristics. Additionally, the expected difference in outcome for the off-label use of BMP in reconstruction did not enable a proper power analysis. As a retrospective chart review, the use of rhBMP-2 was not randomized. The decision to use rhBMP-2 was determined strictly by the treating surgeon. Incidence rates were reported as percentages. A statistical analysis was performed using the Fisher exact and x2 tests for categorical variables. Matched group characteristics between the groups and the compared subsets did not necessitate multivariate regression; univariate analysis was sufficient. Data analysis was processed using SPSS Statistics (SPSS Inc, Chicago, IL). Statistical significance was set for any P value less than 0.05. Analyzing for correlation between surgical sites and outcomes was assisted by grouping the 3 areas of a long bone (proximal, shaft, and distal) as 1 entity (upper extremity, femur, tibia, or foot/ankle) to provide enhanced power.
RESULTS Group 1: Cases (With rhBMP-2) A total of 193 cases in 193 patients were included in the analysis. The average age was 52 years (range, 16–90 years). Ó 2014 Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
Wound Complications in Orthopaedic Trauma Surgery
rhBMP-2 was used in 38 acute trauma patient and 155 reconstructive patients. Of the acute traumatic cases, 34 were open fractures, and 13 of these 34 patients (34%) required some form of soft tissue reconstruction (Table 2). The anatomic site breakdown is summarized in Table 1. Thirty-six patients of the 155 reconstructive cases (23%) had history of associated open fracture, more than half (19/36) were open tibia fractures. Twenty-eight patients (18%) had history of infection, in which 10 involved the tibia and 9 the foot and ankle. Seventeen patients (11%) had history of both an open fracture and infection, most (8/17) were open tibia fractures. The 5 patients (3%) who had history of soft tissue defects requiring reconstruction also had tibia fractures (3/5 secondary to open fractures). An anatomic breakdown of these associated factors of reconstructive cases is included in Table 2. All 155 reconstructive cases underwent radiographic review. There were 45 infected nonunion, 4 malunion, and 26 primary arthrodesis cases. The 80 remaining cases were comprised 27 atrophic, 49 oligotrophic, and 4 hypertrophic nonunions. Anatomic site breakdown of reconstructive cases is included in Table 3.
Group 2: Matched Controls (No rhBMP-2) A total of 181 cases (181 patients) without the use of rhBMP-2 matched for age, anatomic site, acute trauma/ reconstructive surgical procedure, and history of open fracture or infection were included in the analysis. The average age of the patients was 51 years (range, 19–85 years) (Table 1). Of the 36 acute traumatic cases, 32 were open fractures, and 18 required some form of soft tissue reconstruction. Thirty-eight patients of the 145 reconstructive cases had history of associated open fracture, more than half (20/38) were open tibia fractures (Table 2). Twenty-eight patients had history of infection, in which 12 involved the tibia and 7 the foot and ankle. Seventeen patients had history of both open fracture and infection, most (8/17) were open tibia fractures. The 12 patients who had history of soft tissue defects requiring reconstruction were predominantly tibia fractures (10/12 secondary to open fractures). Anatomic breakdown of these associated factors of the reconstructive cases is included in Table 2. All 145 reconstructive cases underwent radiographic review. There were 43 infected nonunion, 3 malunion, and 16 primary arthrodesis cases. The 83 remaining cases were
comprised 34 atrophic, 43 oligotrophic, and 6 hypertrophic nonunions. Anatomic site breakdown of the reconstructive cases is included in Table 3. The difference in frequencies of primary fusion, malunion, and nonunion type between acute and reconstructive patients in each group was not statistically significant (Table 3).
Group Comparison and Statistical Analysis Overall, 60 patients (31%) in the BMP group and 33 patients (18%) in the control group had documentation of at least 1 postoperative wound complication after the index procedure (P = 0.004) (Table 4). Among those patients with a wound complication, the most commonly documented was serous drainage [56/60 (93%) and 22/33 (67%), respectively]. The incidence of wound complications between acute and reconstructive patients in each group was not statistically significant (26% vs. 32% for BMP and 11% vs. 20% for controls). Prolonged antibiotic therapy secondary to a postoperative wound complication was documented in 17 patients (9%) in group 1 and 23 patients (13%) in group 2. In group 1, 1 acute trauma and 5 reconstructive patients had documentation of at least 1 reoperation for a wound complication. In group 2, 2 acute trauma and 13 reconstructive patients underwent reoperation for a wound complication (P = 0.04). Overall, the proportion of patients who returned to the operating room for debridement was 10% (6/60) in group 1 versus 46% (15/33) in group 2 (P , 0.001). Within the BMP group, 4 of 6 patients (2%) had a documented infection based on intraoperative cultures, whereas 2 patients (1%) had a sterile seroma/hematoma. Only 1 BMP patient in the acute trauma subgroup with an open tibia fracture and soft tissue reconstruction had a positive intraoperative culture. The other 3 patients (2 femoral and 1 subtalar nonunions) were in the reconstructive subset treated with BMP. Each of the 4 infections occurred with a different organism. In group 2, 9 of 15 cases had positive intraoperative cultures. Similar to group 1, one patient with an open tibia fracture that required soft tissue reconstruction had a positive intraoperative culture. The 8 in the reconstructive subgroup were comprised 6 tibial, 1 femoral, and 1 subtalar nonunion. No one organism was more common than another.
TABLE 2. Description of Reconstructive Cases Group 1—BMP Acute Open fracture Soft tissue reconstruction Reconstructive Open fractures Infection Open fracture and infection Soft tissue reconstruction
Group 2—Control
UE
F
T
F&A
Total
UE
F
T
F&A
Total
3 2 1 19 3 2 2 —
7 3 — 35 9 7 5 —
27 27 12 36 19 10 7 5
1 1 — 65 5 9 3 —
38 33 13 155 36 28 17 5
2 1 1 14 5 0 3 —
6 3 1 31 6 6 2 —
27 27 16 39 20 12 8 10
1 1 — 61 7 7 5 1
37 32 18 145 38 25 18 11
F, femur; F&A, foot and ankle; T, tibia; UE, upper extremity.
Ó 2014 Lippincott Williams & Wilkins
www.jorthotrauma.com |
601
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
TABLE 3. Breakdown of Reconstructive Cases Group 1—BMP (155 Patients) Primary fusions Malunion Infected nonunion Atrophic nonunion Oligotrophic nonunion Hypertrophic nonunion Total
Group 2—Control (145 Patients)
UE
F
T
F&A
Total*
UE
F
T
F&A
Total*
— 2 4 6 7 — 19
— 1 12 5 14 3 35
— 1 17 6 11 1 36
26 — 12 10 17 — 65
26 4 45 27 49 4 155
— 1 3 8 2 — 14
— — 8 5 13 5 31
— 2 20 8 8 1 39
16 — 12 13 20 — 61
16 3 43 34 43 6 145
*There were no statistically significant differences in the subgroup totals between group 1 and group 2. F, femur; F&A, foot and ankle; T, tibia; UE, upper extremity.
Union Twenty of 193 patients (10%) were lost to follow-up in group 1 (7 acute and 13 reconstructive). Similarly, 26/181 patients (14.4%) were lost to follow-up in group 2 (12 acute and 14 reconstructive). For the 173 patients in group 1, union was achieved in 94% of acute and 89% of reconstructive cases. For the 155 patients in group 2, union occurred in 79% of acute and 73% of reconstructive cases. The differences in union rates between groups 1 and 2 for overall cases and for the reconstructive cases were statistically significant [90% vs. 74% (P , 0.001); 89% vs. 73% (P = 0.002), respectively] (Table 4). The union rate among acute cases was higher in group 1 than that in group 2 (94% vs. 79%), but this difference failed to reach statistical significance (P = 0.220). There were no other significant differences in union rates when comparing the 2 groups for anatomic site, need for soft tissue coverage, or history of infection. Among those patients with wound complications, union rates also differed significantly between the groups (group 1: 50/53, group 2: 18/31, P , 0.001).
TABLE 4. Summary of Results Group 1—BMP Group 2—Control (193 Patients) (181 Patients) Overall wound complications Wound concern—acute trauma Wound concern—reconstructive Overall prolonged antibiotic use Prolonged antibiotic use— acute trauma Prolonged antibiotic use— reconstructive Overall reoperation for wound concern Reoperation for wound concern— acute trauma Reoperation for wound concern— reconstructive Overall union rate Union rate for acute trauma Union rate for reconstructive
602
| www.jorthotrauma.com
60 (31%) 10 (5%) 50 (26%) 17 (9%) 2 (1%)
33 (18%), P = 0.004 4 (2%) 29 (16%) 23 (13%) 3 (2%)
15 (8%)
20 (11%)
6 (3%)
15 (8%), P = 0.04
1 (0.5%)
2 (1%)
5 (3%)
13 (7%), P = 0.05
90% 94% 89%
74%, P , 0.001 79%, P = 0.220 73%, P = 0.002
Finally, age, sex, acute trauma, reconstructive surgery, history of an open fracture or infection, or the need for a soft tissue reconstruction did not correlate with the development of a wound complication or a return to the operating room.
DISCUSSION Persistent postoperative wound drainage associated with the use of rhBMP-2 has been reported mostly in the spine literature.5 In our series, there was not an increased incidence of postoperative wound infection with the use of rhBMP-2; however, rhBMP-2 does seem to have a beneficial effect in improving union rates overall. A BMP-induced immune-modulated inflammatory response is the proposed mechanism for the observed wound drainage and dehiscence.6,7 Although 31% of our patients in group 1 had a wound complication, most commonly serous drainage, observation alone was successful 72% (43/60) of the time. It has been shown that the concentration of rhBMP2, and length of time at the implant site, directly affects the amount of bone formed.8 Conversely, if local or implant site retention is low and rhBMP-2 leaks into the surrounding tissue, an inflammatory response might be more likely. Similarly, it is possible that antibodies to rhBMP-2 or to the bovine collagen carrier are produced by the “host” and induce a physiologic inflammatory reaction. An immune-modulated inflammatory response to the protein moiety and or carrier for BMP-2 is, at present, the best explanation for these observations; however, the exact mechanism and the clinical relevance are unknown.9 The rhBMP-2 Evaluation in Surgery for Tibial Trauma (BESTT) Study Group10 conducted a large European multinational randomized control trial, which showed that open tibial fractures treated with intramedullary nails and 1.50 mg/mL of rhBMP-2 at final closure had fewer implant failures, faster wound healing, and fewer infections in Gustilo– Anderson type-III injuries. The authors stated that the adverse events observed (eg, edema, pain, infection) “were consistent with those normally observed in the trauma setting.” Anecdotally, the wound complications most often associated with the use of BMP’s are characterized by drainage, erythema, and edema; nonspecific signs that are also indicative of inflammation commonly seen after trauma and Ó 2014 Lippincott Williams & Wilkins
J Orthop Trauma Volume 28, Number 10, October 2014
prolonged reconstructive procedures. This observation was reported by Jones et al,11 where 33% (5/15) of the diaphyseal tibia fractures treated with freeze-dried cancellous allograft and rhBMP-2 developed postoperative erythema compared with 0% (also of 15) in the group treated with autograft only. This incidence reported by Jones et al is in agreement with our data. Similar to our findings, these wound complications were typically self-limited and resolved without treatment. Although we did observe an increased incidence of prolonged serous drainage, we also noted a decreased reoperation rate for wound complications in group 1. Although the presence or absence of a specific wound complication can easily be documented, the quality or severity of these concerns is more difficult to accurately quantify. This study relies on an accurate documentation as outlined in the Methods. We observed that the progress notes were appropriately detailed, including the pertinent positive and negatives, which suggests that a thorough assessment and documentation of a potential postoperative wound complication was performed. A reporting bias and a clinical decisionmaking bias may however exist; because serous drainage related to rhBMP-2 became a more familiar entity, its presence was more frequently observed and documented, and its management with observation alone was more frequently practiced. However, if this were in fact true, we believe that there would also be a higher number of reoperations with negative cultures in the control group, which was not the case [4/6 (67%) rhBMP-2 reoperations with proven infection vs. 9/15 (60%) in the control group]. Data such as reoperation, culture results, and antibiotic administration are concrete end points; based on the equal proportion of reoperations for proven infection, the decisionmaking algorithm for operative intervention for a wound complication were the same for both groups. Although the incidence of postoperative wound drainage occurred more frequently in the rhBMP-2 group, the need for reoperation (and the incidence of proven infection) was more frequent in the control group. Reasons for this are unclear; however, the findings are in agreement with the results of the BESST study,10 where the need for reoperation and infection were also decreased in the group of patients treated with rhBMP-2. Postoperative infection rates have been reported extensively, especially in the antibiotic prophylaxis literature that conclusively shows the efficacy of preoperative antibiotics. Boxma et al12 reported a postoperative infection rate of 3.6% for more than 1000 closed fracture cases that received preoperative antibiotics. Similar results were found by Patzakis et al13 in their study of open fractures treated with preoperative cephalothin with an infection rate of 2.3%. More recently, Williams et al14 reported a similar wound infection rate (2.4%) in an analysis of 55,862 cases of spinal fusion; this rate was the same for both cases with and without BMP. In our study, we observed a 2% infection rate (based on positive intraoperative cultures in cases in which rhBMP-2 was used), and this is consistent with the above reported studies. With regard to union in acute trauma, the findings of this study are in agreement with those of BESST.8 The Ó 2014 Lippincott Williams & Wilkins
Wound Complications in Orthopaedic Trauma Surgery
present analysis demonstrated an improved union rate (94% vs. 79%) for the acute trauma cases whose treatment included rhBMP-2. This is an important finding because it is contradictory to a recent report that found no difference with rhBMP-2 use in open tibia fractures.15 Furthermore, this effect (not previously noted in the literature) was also found in the reconstructive cases (nonunion and posttraumatic arthritis), whereby those patients whose treatment included rhBMP-2 demonstrated an increased incidence of clinical and radiographic union (89% vs. 73%, P = 0.002). Additional studies are needed to more closely evaluate and report the benefit of rhBMP-2 in arthrodesis cases. One of the weaknesses of this analysis is the inclusion of multiple surgical sites in each group, inducing type 1 error and not allowing for complete homogeneity for comparison. Stratifying patients into smaller groups for subgroup analysis inherently diminished the power for a meaningful comparison. We believe this to be the case with regard to the significant differences observed in the infection rates between femoral and foot/ankle cases, as well as the observed differences in prolonged antibiotic administration and postoperative complications between tibial and foot/ankle cases. However, despite the heterogeneity, the comparison with a 1:1 matched cohort (age, injury site, acute trauma vs. reconstructive surgery, history of open fracture or infection, and soft tissue coverage) does address the pooling of subgroups. This means that there were equivalent number of upper and lower extremity injuries, acute fractures, nonunions, and arthrodeses, and sterile versus previously infected cases. Similarly, treatment methods or approaches are also likely to be comparable between cohorts (ie, nails vs. plates or early total care vs. staged treatment). Therefore, we believe that the associated differences with each of these subsets with regard to treatment, healing, or complications are the same when comparing a matched cohort. In this study, the surgeon chose to use rhBMP-2 independently. The decision-making process cannot accurately be determined from a retrospective chart review; however, comparing 2 groups of open fractures matched to age, fracture site, and need for soft tissue reconstruction does enable us to compare similar severity of injuries (ie, grade of open fracture, role for rhBMP-2). Similarly, the decision to use rhBMP-2 to repair a nonunion cannot consistently be determined by chart review; however, the characteristics of the 2 subsets (history of open fracture, infection, or need for soft tissue reconstruction) are matched as mentioned above, suggesting that the state of the “biology” of the nonunion also were equivalent. Similar nonunion type categorization supports the equivalence of these matched groups, suggests proper use of BMP in group 1 (95% oligotrophic or atrophic nonunions), and enhances the union results and conclusions. The strengths of this study include the large volume of cases performed at a single institution over a short period of time, providing adequate power to show differences between 2 similar groups matched for case type, anatomic region, and patient demographics. The choice of concrete end points, such as reoperation and union, allows for accurate and reliable results. Data collected from descriptions of postoperative surgical wounds within progress notes were also accurate www.jorthotrauma.com |
603
J Orthop Trauma Volume 28, Number 10, October 2014
Chan et al
because of the fact that prolonged drainage is a significant event, it lengthens hospital stays, and/or is the reason for prolonged or additional inpatient antibiotic administration.
CONCLUSIONS
This is the first study, to the best of our knowledge, specifically investigating the incidence of wound complications associated with the use of rhBMP-2 in orthopaedic trauma. We found that in both acute and reconstructive extremity trauma, rhBMP-2 does seem to result in an increased incidence of postoperative serous wound drainage. However, this does not seem to correlate with an increased incidence of postoperative wound infection or the need for reoperation. Looking forward, the mechanism for prolonged serous drainage can be better delineated and definitive recommendations for its management. Equally important, and also not previously reported, is our finding that the use of rhBMP-2 improves union rates for both acute trauma and posttraumatic reconstruction of the extremities. Further investigation is needed in rhBMP-2 cases with minimal heterogeneity and specific focus on its use in posttraumatic arthrodesis.
5. 6. 7. 8. 9. 10.
11.
12.
REFERENCES 1. Urist MR. Bone: formation by autoinduction. Science. 1965;150: 893–899. 2. Kingsley DM. The TGF-beta superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 1994;8:133–146. 3. Cheng H, Jiang W, Phillips FM, et al. Osteogenic activity of the fourteen types of human bone morphogenetic proteins (BMPs). J Bone Joint Surg Am. 2003;85-A:1544–1552. 4. Ong KL, Lau E, Villarraga ML, et al. Evaluating on-label and off-label use of BMP in the U.S. using administrative data. Scientific Exhibit No.
604
| www.jorthotrauma.com
13. 14.
15.
71, 77th Annual Meeting of the American Academy of Orthopaedic Surgeons, New Orleans, LA, March 9–13, 2010. Owens J, Glassman SD, Howard JM, et al. Perioperative complications with rhBMP-2 in transforaminal lumbar interbody fusion. Eur Spine J. 2011;20:612–617. Vaidya R, Carp J, Sethi A, et al. Complications of anterior cervical discectomy and fusion using recombinant human bone morphogenetic protein-2. Eur Spine J. 2007;16:1257–1265. Wang EA, Rosen V, D’Alessandro JS, et al. Recombinant human bone morphogenetic protein induces bone formation. Proc Natl Acad Sci U S A. 1990;87:2220–2224. Blum B, Moseley J, Miller L, et al. Measurement of bone morphogenetic proteins and other growth factors in demineralized bone matrix. Orthopedics. 2004;27(suppl):S161–S165. Szpalski M, Gunzburg R. Recombinant human bone morphogenetic protein-2: a novel osteoinductive alternative to autogenous bone graft? Acta Orthop Belg. 2005;71:133–148. Govender S, Csimma C, Genant HK, et al. Recombinant human bone morphogenetic protein-2 for treatment of open tibial fractures: a prospective, controlled, randomized study of four hundred and fifty patients. J Bone Joint Surg Am. 2002;84:2123–2134. Jones AL, Bucholz RW, Bosse MJ, et al; BMP-2 Evaluation in Surgery for Tibial Trauma-Allgraft (BESTT-ALL) Study Group. Recombinant human BMP-2 and allograft compared with autogenous bone graft for reconstruction of diaphyseal tibial fractures with cortical defects. A randomized, controlled trial. J Bone Joint Surg Am. 2006;88: 1431–1441. Boxma H, Broekhuizen T, Patka P, et al. Randomised controlled trial of single-dose antibiotic prophylaxis in surgical treatment of closed fractures: the Dutch Trauma Trial. Lancet. 1996;347:1133–1137. Patzakis MJ, Harvey JP Jr, Ivler D. The role of antibiotics in the management of open fractures. J Bone Joint Surg Am. 1974;56:532–541. Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? A comparison of 55,862 cases of spinal fusion with and without bone morphogenetic protein. Spine 2011;36:1685–1691. Aro HT, Govender S, Patel AD, et al. Recombinant human bone morphogenetic protein-2: a randomized trial in open tibial fractures treated with reamed nail fixation. J Bone Joint Surg Am. 2011;93: 801–808.
Ó 2014 Lippincott Williams & Wilkins
