NEJMoa1508502

The new engl and jour nal of medicine

Original Article

A Trial of Wound Irrigation in the Initial Management of Open Fracture Wounds

The FLOW Investigators*

ABSTR ACT

BACKGROUND

The management of open fractures requires wound irrigation and débridement to The members of the writing committee remove contaminants, but the effectiveness of various pressures and solutions for (Mohit Bhandari, M.D., Ph.D., Kyle J. Jeray, M.D., Brad A. Petrisor, M.D., P.J. Deverirrigation remains controversial. We investigated the effects of castile soap versus eaux, M.D., Ph.D., Diane Heels-Ansdell, normal saline irrigation delivered by means of high, low, or very low irrigation M.Sc., Emil H. Schemitsch, M.D., Jeff pressure. Anglen, M.D., Gregory J. Della Rocca, M.D., Ph.D., Clifford Jones, M.D., Hans Kreder,

METHODS M.D., M.P.H., Susan Liew, M.B., B.S., Paula McKay, B.Sc., Steven Papp, M.D., In this study with a 2-by-3 factorial design, conducted at 41 clinical centers, we Parag Sancheti, M.B., B.S., M.Ch., D.N.B., randomly assigned patients who had an open fracture of an extremity to undergo Sheila Sprague, Ph.D., Trevor B. Stone,

irrigation with one of three irrigation pressures (high pressure [>20 psi], low pres- M.D., Xin Sun, Ph.D., Stephanie L. Tanner, M.S., Paul Tornetta III, M.D., Ted Tufescu, sure [5 to 10 psi], or very low pressure [1 to 2 psi]) and one of two irrigation solu- M.D., Stephen Walter, Ph.D., and Gordon tions (castile soap or normal saline). The primary end point was reoperation H. Guyatt, M.D.) assume responsibility

within 12 months after the index surgery for promotion of wound or bone healing for the overall content and integrity of the article. The affiliations of the members or treatment of a wound infection. of the writing group are listed in the Ap pendix. Address reprint requests to Dr.

RESULTS Bhandari at the Division of Orthopaedic A total of 2551 patients underwent randomization, of whom 2447 were deemed Surgery, McMaster University, 293 Welling ton St. N., Suite 110, Hamilton, ON L8L eligible and included in the final analyses. Reoperation occurred in 109 of 826 pa- 8E7, Canada, or at bhandam@mcmaster.ca.

tients (13.2%) in the high-pressure group, 103 of 809 (12.7%) in the low-pressure

* A complete list of the Fluid Lavage of Open Wounds (FLOW) Investigators is group, and 111 of 812 (13.7%) in the very-low-pressure group. Hazard ratios for

the three pairwise comparisons were as follows: for low versus high pressure, 0.92 provided in the Supplementary Appen(95% confidence interval [CI], 0.70 to 1.20; P = 0.53), for high versus very low pres- dix, available at NEJM.org.

sure, 1.02 (95% CI, 0.78 to 1.33; P = 0.89), and for low versus very low pressure, This article was published on October 8, 0.93 (95% CI, 0.71 to 1.23; P = 0.62). Reoperation occurred in 182 of 1229 patients 2015, at NEJM.org.

(14.8%) in the soap group and in 141 of 1218 (11.6%) in the saline group (hazard N Engl J Med 2015;373:2629-41. ratio, 1.32, 95% CI, 1.06 to 1.66; P = 0.01). DOI: 10.1056/NEJMoa1508502

CONCLUSIONS

The rates of reoperation were similar regardless of irrigation pressure, a finding that indicates that very low pressure is an acceptable, low-cost alternative for the irrigation of open fractures. The reoperation rate was higher in the soap group than in the saline group. (Funded by the Canadian Institutes of Health Research and others; FLOW ClinicalTrials.gov number, NCT00788398.)

n engl j med 373;27 nejm.org December 31, 2015 2629

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

T promote wound and bone healing.2,4 Clinicians fractures requires thorough irrigation and débridement1-4 to prevent infection and The study was funded by the Canadian Institutes of Health Research, the U.S. Department of De- fense, and others. The Clinical Advances through he initial management of open Study Oversight

accomplish débridement by removing all visible Research and Information Translation (CLARITY) debris and necrotic tissue and by providing copi- Research Group at McMaster University coordi-ous irrigation of the wound. nated the trial and was responsible for the trial Controversy exists regarding the choice of ir- randomization, the maintenance, validation, and rigation pressure and solution.4-13 High pressure analysis of the data, and the study-center coor-may be more effective than low pressure in re- dination. The Greenville Health System assisted moving particulate matter and bacteria7-10 but at in the coordination of study sites in the United the expense of bone damage8,11 and a resultant States. Stryker donated Surgilav irrigators for delay in bone healing.12 Low pressure may avoid the trial for clinical sites in Asia. Zimmer pro-bone damage and delayed healing but at the pos- vided the Pulsavac irrigator at a reduced cost to sible cost of less effective removal of foreign selected clinical sites in North America. Triad matter and bacteria. Medical donated castile soap; castile soap from Regarding the preferred irrigation solution, Aplicare was purchased at full cost. No donor or there is a strong biologic rationale for the use of funder had a role in the design or conduct of the surfactants, such as soap.14-20 Because soap con- study, the collection or analyses of the data, or tains both nonpolar and polar molecules, it acts the preparation of the manuscript.

as an emulsifier, dispersing one liquid, or par- The steering committee (see the Supplemen-ticulate, into another immiscible liquid. As com- tary Appendix, available with the full text of this pared with other enhanced irrigation solutions article at NEJM.org), chaired by the principal (i.e., those that contain antiseptic or antibiotic investigators, designed the trial and prespecified agents), soap is less expensive,21 does not have the statistical analysis plan. The members of the a risk of antibiotic resistance,14 and is less steering committee vouch for the completeness toxic.1,7,17-19,22,23 and accuracy of the data and analyses reported To address these issues regarding irrigation and for the adherence of the trial to the proto-pressures and solutions, we conducted the Fluid col, available at NEJM.org. The first author, who Lavage of Open Wounds (FLOW) trial in patients was the chair of the writing committee, wrote requiring surgery for open fracture. We exam- the first draft of the manuscript; all the mem-ined the effect of alternative pressures and bers of the writing committee made revisions castile soap versus normal saline irrigation on and made the decision to submit the manuscript a composite of a number of different reasons for publication.

for reoperations within 12 months after the

index surgery. Patients

From June 2009 through September 2013, we

Methods recruited patients across 41 sites in the United States, Canada, Australia, Norway, and India.

Study Design Eligible patients were 18 years of age or older Our study was an international, blinded, ran- with an open fracture of an extremity that re- quired operative fixation. Extremity was defined domized, controlled trial that used a 2-by-3

factorial design to evaluate the effects of high as arm, wrist, leg, ankle, foot, clavicle, or scap-versus low versus very low (gravity flow) irriga- ula. We excluded fractures of the pelvic ring and tion pressures and soap versus normal saline axial skeleton and fractures of the hand (meta- carpals and phalanges) and toes (phalanges). solutions on reoperation rates among patients

with an open fracture. The objectives and Detailed eligibility criteria are listed in the Sup-methods of the trial were published previous- plementary Appendix.

ly.24 The study was approved by the ethics com-

mittees at McMaster University, Greenville Procedures

Health System, and each participating center. Patients were stratified according to study center All the patients provided written informed and Gustilo–Anderson fracture grade (I or II vs. consent. III) (see the Supplementary Appendix). Patients

2630 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

underwent randomization in a 1:1:1:1:1:1 ratio and were assigned to one of six treatment groups: soap and very low pressure, soap and low pres-sure, soap and high pressure, saline and very low pressure, saline and low pressure, or saline and high pressure. Randomization was performed with the use of a central computerized system with variable block sizes, thus ensuring conceal-ment of the study-group assignments. Patients, end-point adjudicators, and data analysts were unaware of the study-group assignments.

During surgery, the initial management of the open fracture included irrigation that was deliv-ered by means of very low pressure (1 to 2 psi), low pressure (5 to 10 psi), or high pressure (>20 psi). In the operating room, surgeons used a sterile technique to prepare either a 0.45% solution of castile soap (Triad Medical and Aplicare) in nor-mal saline (see the Supplementary Appendix) or used sterile normal saline alone. We standardized the perioperative antibiotic regimens and the minimum amount of solution according to the severity of the open fracture wound, which was graded according to the Gustilo–Anderson clas-sification (3 liters for grade I fracture and 6 liters for grade II or III) (see the Supplementary Ap-pendix).2

Patients returned for follow-up assessments at 1, 2, and 6 weeks and 3, 6, 9, and 12 months after surgery. Details of the follow-up process are provided in the Supplementary Appendix.

Study End Points

The primary end point was reoperation, defined as surgery that occurred within 12 months after the initial procedure to treat an infection at the operative site or contiguous to it, manage a wound-healing problem, or promote bone heal-ing. The procedures included in this composite end point were the following: irrigation and dé-bridement for an infected wound; revision and closure for wound dehiscence; wound coverage for an infected or necrotic wound; drainage of a hematoma; reoperation for hardware failure that was probably related to an infection, wound-healing problem, or bone-healing problem (e.g., delayed union or nonunion); bone grafting or implant-exchange procedure for established non-union in patients with a postoperative fracture gap of less than 1 cm; intramedullary nail dynam-izations in the operating room (dynamization involves removal of locking screws from the intra-medullary nail to allow fracture ends to com-

press with weight bearing); fasciotomies for the compartment syndrome; and other events as determined by the adjudication committee. Full details are provided in the Supplementary Appen-dix. Secondary end points included nonopera-tively managed infection and wound-healing and bone-healing problems within 12 months after the index surgery.

A central adjudication committee, whose mem-bers were unaware of the study-group assign-ments, adjudicated all primary and key secondary end points. To minimize random error, the com-mittee blindly adjudicated trial eligibility on the basis of data available before or shortly after ran-domization (see the Supplementary Appendix).25

Statistical Analysis

We originally calculated that the sample size would have to be 2280 patients, with 1140 pa-tients per solution group and 760 patients per pressure group. This sample size was based on the size of the irrigation-pressure groups and was calculated to ensure that the study would have a power of 80% to identify differences among the three irrigation-pressure groups in effects of pair-wise comparisons at an adjusted alpha level of 0.0188, on the basis of a rate of reoperation within 12 months of 30% in a control group and a 25% lower relative risk with one irrigation pres-sure than with another. We estimated a similar control-group reoperation rate for normal sa-line,13,26,27 and the study therefore also had 98% power to detect a 25% lower relative risk with soap — a treatment effect that was endorsed by 80% of surgeons in our international survey as important enough to change practice.13

An interim analysis was performed in January 2013 after 2079 patients had been enrolled; 789 of these patients had 12-month outcomes avail-able. The external data and safety monitoring committee considered the O’Brien–Fleming stop-ping criterion that specifies a significance level that maintains the overall type I error rate of 0.05,28 and the committee recommended the re-cruitment of additional patients in the trial to account for a projected 10% loss to follow-up. We recruited a total of 2551 patients.

The analyses included all the patients in the groups to which they were randomly assigned. For patients for whom 12-month follow-up in-formation was unavailable, data were included to the date of their last documented follow-up and were censored at that time. The CLARITY

n engl j med 373;27 nejm.org December 31, 2015 2631

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

Research Group data analyst remained unaware We conducted an additional post hoc subgroup of the treatment-group assignments while con- analysis that evaluated the possible effect modi-ducting the primary analyses. fication according to time to surgery (<6 hours, Using Cox regression stratified according to 6 to 12 hours, or >12 hours after injury). We fracture grade (I or II vs. III) and study center, used multiple criteria to consider the credibility we first conducted a time-to-event analysis of of any possible subgroup effects.30 The Supple-the main effects with respect to solution and mentary Appendix provides details regarding pressure and the interaction between the two hypothesized subgroup effects.

with regard to the reoperation rate. If the inter- We first interpreted the results on the basis action was significant, we planned to explore of a blinded review of the results of our primary the nature of the effect modification. analysis.31 The randomization code was then Our primary analysis was a Cox regression broken, the correct interpretation chosen, and the stratified according to the severity of the open draft of the manuscript was written. The Supple-fracture2 and study center, with reoperation as mentary Appendix provides details regarding the end point in the time-to-event analysis. The specific analyses and our blinded interpretation.

Cox regression to investigate the effect of irriga-

Results tion pressure was also stratified according to

irrigation solution. Similarly, the irrigation-solu-

tion analysis was stratified according to irrigation Patients

pressure. We also performed analyses, using Cox From June 2009 through September 2013, we regression, that were adjusted for age, injury (up- randomly assigned 859 of 2551 enrolled patients per extremity vs. lower extremity), fracture gap to the high-pressure group, 846 to the low-(<1 cm vs. ≥1 cm), type of internal fixation (intra- pressure group, and 846 to the very-low-pressure medullary nail, external fixator, plate, other inter- group. A total of 1275 patients were assigned to nal fixation, other fixation, or none), and severity irrigation with soap and 1276 to irrigation with of wound contamination (mild vs. moderate vs. normal saline. Of 2551 patients enrolled, the severe). For both the primary and adjusted analyses, adjudication committee (whose members were we tested the proportional-hazards assumption. unaware of the treatment assignments) deter- We examined the three-category randomized mined that 104 patients were ineligible owing to pressure variable for statistical significance at no receipt of surgical treatment (47 patients), in-an alpha level of 0.05; if the results were signifi- correct fracture type (48), history of osteomyelitis cant, we planned to conduct pairwise compari- (1), retained hardware from a previous fracture sons using an alpha level of 0.0188. In instances in the same extremity (2), use of immunosup-of significant differences between results, we pressive medication (2), or age (4). The remain-conducted a sensitivity analysis that was based on ing 2447 patients were included in the final plausible differences in event rates among pa- analyses, with the patients’ data analyzed in the tients lost to follow-up versus those for whom treatment groups to which the patients had been follow-up was complete.29 In this case, we tested randomly assigned. We obtained 12-month follow-the effect of assuming that in the saline group, up data for 90% of the patients (Figs. S1 and S2 the event rate among patients who were lost to in the Supplementary Appendix).

follow-up would be twice as high as the rate The majority of patients were men, were in among those successfully followed. All the analy- their 40s, were those with a lower-extremity frac-ses were performed with the use of SAS software, ture, and were those with no concomitant major version 9.2 (SAS Institute). trauma. The most common mechanism of injury Before unblinding and as described in our was motor vehicle accident. The characteristics statistical analysis plan, we prespecified 12 sub- were similar in the randomized study groups group analyses that explored a possible modifica- (Table 1). Typical patients underwent plate fixa-tion of the effect of alternative irrigation pressures tion, underwent their first irrigation within 10 and solutions in subgroups defined according to hours after their injury, and received antibiotic fracture severity, location of fracture (upper vs. prophylaxis; the treatments, including volumes lower extremity, tibial vs. nontibial fracture, and of irrigation solutions, were similar in the ran-intraarticular vs. extraarticular involvement), domized study groups (Table 1, and Tables S1 and aspects of the surgical wound débridement. and S2 in the Supplementary Appendix).

2632 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

Adherence to Assigned Intervention

Adherence by the surgeon to the initially as-signed irrigation pressure ranged from 96.5% to 98.8%. Adherence by the surgeon to the initially assigned irrigation solution was 97.9% in the soap group and 99.6% in the saline group (Tables S3 and S4 in the Supplementary Appendix).

Interaction between Irrigation Pressures and Solutions

Results showed no interaction between solution and pressure (P = 0.31). Therefore, we completed separate analyses for irrigation pressures and solutions.

in patients with a fracture gap of less than 1 cm was significantly higher in the soap group than in the saline group (hazard ratio, 1.59; 95% CI, 1.01 to 2.51; P = 0.046) (Table 3). Our sensitivity analysis showed that if we assumed that the patients who were lost to follow-up in the soap group had the same risk of the primary end point as those who had complete follow-up and that the patients who were lost to follow-up in the saline group had a risk of the primary end point that was twice as high as the risk among those with complete follow-up, then the study would lose statistical significance of the effect of soap versus saline (P = 0.16).

Primary End Point Secondary End Points

According to Irrigation Pressure

A primary study end-point event, reoperation within 12 months after the index procedure in order to treat an infection, manage a wound-healing problem, or promote bone healing, oc-curred in 323 of the 2447 patients (13.2%). The rate of the primary end point did not differ sig-nificantly according to type of irrigation pressure: 109 of 826 patients (13.2%) in the high-pressure group had a primary end-point event, as did 103 of 809 patients (12.7%) in the low-pressure group and 111 of 812 (13.7%) in the very-low-pressure group (P = 0.80 for the three-way comparison). Hazard ratios were as follows: for low pressure versus high pressure, 0.92 (95% confidence in-terval [CI], 0.70 to 1.20; P = 0.53); for high pres-sure versus very low pressure, 1.02 (95% CI, 0.78 to 1.33; P = 0.89); and for low pressure versus very low pressure, 0.93 (95% CI, 0.71 to 1.23; P = 0.62) (Table 2 and Fig. 1A). Adjusted analyses yielded similar results (Table S5 in the Supple-mentary Appendix).

We found no significant differences among the three irrigation pressures with respect to the sec-ondary end points of nonoperatively managed infection, wound-healing problem, and bone-healing problem (Table 2). Likewise, we found no significant differences between the two irri-gation solutions with respect to any of the sec-ondary end points (Table 3).

Subgroup Analyses

Subgroup analyses of the various irrigation pres-sures and solutions yielded results that were consistent with the primary treatment effects for each intervention. The exceptions were tibial ver-sus nontibial fracture, for which the results sug-gested a trend toward superiority of very low pressure over low or high pressure in patients with a tibial fracture, and a similarity in the soap group and the saline group when the duration of antibiotic use after surgery was 4 days or more (Fig. 2, and Tables S7 and S8 in the Supplemen-tary Appendix).

According to Irrigation Solution Discussion

The rate of the primary end point differed sig-

nificantly according to type of irrigation solution: We found no significant influence of irrigation 182 of 1229 patients (14.8%) in the soap group pressure on our composite primary end point of had a primary end-point event, as compared with

various forms of reoperation for treatment of 141 of 1218 (11.6%) in the saline group (hazard infection, wound-healing problem, or bone-heal-ratio in the soap group, 1.32; 95% CI, 1.06 to ing problem within 12 months after the initial 1.66; P = 0.01) (Table 3 and Fig. 1B). Adjusted surgery. The irrigation of open fracture wounds analyses yielded similar results for the effect of

with soap, as compared with saline solution, was solution (Table S6 in the Supplementary Appendix). associated with a significantly higher rate of re- The frequency of all components of the pri- operation within 12 months. The effects of the mary end point was higher in the soap group than irrigation pressures and solutions were consis-in the saline group. The frequency of implant-

tent across all components of the primary end exchange procedures for established nonunion point. No significant differences in the rates of

n engl j med 373;27 nejm.org December 31, 2015 2633

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

1/2428 (<0.1) 0/819 1/801 (0.1) 0/808 0/1209 1/1219 (0.1) Plane crash

1/2428 (<0.1) 0/819 0/801 1/808 (0.1) 1/1209 (0.1) 0/1219 Other

1680/2428 (69.2) 563/819 (68.7) 553/801 (69.0) 564/808 (69.8) 838/1209 (69.3) 842/1219 (69.1) Male

748/2428 (30.8) 256/819 (31.3) 248/801 (31.0) 244/808 (30.2) 371/1209 (30.7) 377/1219 (30.9) Female

613/2428 (25.2) 199/819 (24.3) 215/801 (26.8) 199/808 (24.6) 315/1209 (26.1) 298/1219 (24.4) Driver or passenger

331/2429 (13.6) 109/820 (13.3) 97/801 (12.1) 125/808 (15.5) 171/1209 (14.1) 160/1220 (13.1) Any of the above

721/2442 (29.5) 234/824 (28.4) 234/808 (29.0) 253/810 (31.2) 339/1214 (27.9) 382/1228 (31.1) I

304/2428 (12.5) 109/819 (13.3) 92/801 (11.5) 103/808 (12.7) 146/1209 (12.1) 158/1219 (13.0) Pedestrian

414/2428 (17.1) 136/819 (16.6) 138/801 (17.2) 140/808 (17.3) 203/1209 (16.8) 211/1219 (17.3) Motorcycle accident

261/2428 (10.7) 96/819 (11.7) 73/801 (9.1) 92/808 (11.4) 135/1209 (11.2) 126/1219 (10.3) Fall from standing

422/2428 (17.4) 144/819 (17.6) 144/801 (18.0) 134/808 (16.6) 201/1209 (16.6) 221/1219 (18.1) Fall from height

116/2428 (4.8) 44/819 (5.4) 39/801 (4.9) 33/808 (4.1) 58/1209 (4.8) 58/1219 (4.8) Crush injury

74/2428 (3.0) 13/819 (1.6) 28/801 (3.5) 33/808 (4.1) 36/1209 (3.0) 38/1219 (3.1) Penetrating direct trauma

141/2428 (5.8) 43/819 (5.3) 47/801 (5.9) 51/808 (6.3) 73/1209 (6.0) 68/1219 (5.6) Blunt direct trauma

145/2429 (6.0) 45/820 (5.5) 43/801 (5.4) 57/808 (7.1) 75/1209 (6.2) 70/1220 (5.7) Head injury

192/2429 (7.9) 64/820 (7.8) 57/801 (7.1) 71/808 (8.8) 95/1209 (7.9) 97/1220 (8.0) Chest injury

95/2429 (3.9) 37/820 (4.5) 22/801 (2.7) 36/808 (4.5) 51/1209 (4.2) 44/1220 (3.6) Intraabdominal injury

47/2428 (1.9) 26/819 (3.2) 11/801 (1.4) 10/808 (1.2) 24/1209 (2.0) 23/1219 (1.9) All-terrain vehicle accident

26/2428 (1.1) 7/819 (0.9) 9/801 (1.1) 10/808 (1.2) 13/1209 (1.1) 13/1219 (1.1) Twist

2/2428 (0.1) 0/819 1/801 (0.1) 1/808 (0.1) 1/1209 (0.1) 1/1219 (0.1) Explosion

6/2428 (0.2) 2/819 (0.2) 3/801 (0.4) 1/808 (0.1) 3/1209 (0.2) 3/1219 (0.2) Bicycle accident

(N = 2447) Irrigation Pressure Irrigation Solution Characteristic

45.2±17.8 44.7±17.4 44.7±18.2 46.1±17.8 45.0±17.3 45.3±18.3 Age — yr

Total

(N = 826) (N = 809) (N = 812) (N = 1218) (N = 1229)

High Low Very Low Saline Soap

Table 1. Characteristics of the Patients and Surgical and Perioperative Treatment.*

Initial stratification according to Gustilo–

Mechanism of injury — no./total no. (%)

Anderson fracture grade

— no./total no. (%)

Major concomitant trauma

Motor vehicle accident

Sex — no./total no. (%)

2634 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society.

Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

Plus–minus values are means ±SD. Irrigation pressures were defined as follows: high pressure, more than 20 psi; low pressure, 5 to 10 psi, and very low pressure, 1 to 2 psi. There were

Total

(N = 2447) Irrigation Pressure Irrigation Solution Characteristic

High Low Very Low Saline Soap

(N = 826) (N = 809) (N = 812) (N = 1218) (N = 1229)

978/2442 (40.0) 333/824 (40.4) 340/808 (42.1) 305/810 (37.7) 522/1214 (43.0) 456/1228 (37.1) II

533/2442 (21.8) 189/824 (22.9) 175/808 (21.7) 169/810 (20.9) 249/1214 (20.5) 284/1228 (23.1) IIIA

210/2442 (8.6) 68/824 (8.3) 59/808 (7.3) 83/810 (10.2) 104/1214 (8.6) 106/1228 (8.6) IIIB

Location of fracture — no./total no. (%)

758/2428 (31.2) 255/819 (31.1) 242/801 (30.2) 261/808 (32.3) 359/1209 (29.7) 399/1219 (32.7) Upper extremity

1670/2428 (68.8) 564/819 (68.9) 559/801 (69.8) 547/808 (67.7) 850/1209 (70.3) 820/1219 (67.3) Lower extremity

9.8 (6.4–15.9) 10.0 (6.3–16.1) 9.6 (6.4–15.5) 9.5 (6.5–15.9) 9.6 (6.3–15.6) 9.9 (6.5–16.0) Median time from injury to first incision

(interquartile range) — hr

Surgical preparation solution

— no./total no. (%)

1246/2414 (51.6) 424/809 (52.4) 413/799 (51.7) 409/806 (50.7) 617/1200 (51.4) 629/1214 (51.8) Iodine

1065/2414 (44.1) 353/809 (43.6) 346/799 (43.3) 366/806 (45.4) 520/1200 (43.3) 545/1214 (44.9) Chlorhexidine

401/2414 (16.6) 134/809 (16.6) 131/799 (16.4) 136/806 (16.9) 200/1200 (16.7) 201/1214 (16.6) Alcohol

209/2414 (8.7) 66/809 (8.2) 69/799 (8.6) 74/806 (9.2) 106/1200 (8.8) 103/1214 (8.5) Other

2395/2421 (98.9) 804/814 (98.8) 792/800 (99.0) 799/807 (99.0) 1191/1204 (98.9) 1204/1217 (98.9) Adherence to antibiotic protocol

— no./total no. (%)

Volume of irrigation solution — liters

6.0±2.2 6.1±2.1 5.9±2.0 6.2±2.5 6.2±2.3 5.9±2.2 Total

According to postoperative fracture

grade

5.4±1.8 5.3±1.8 5.3±1.8 5.4±1.9 5.5±1.9 5.3±1.8 I or II

7.2±2.3 7.3±1.9 7.0±1.9 7.4±2.9 7.4±2.3 7.1±2.3 III

Definitive fixation — no./total no. (%)

821/2421 (33.9) 264/814 (32.4) 289/800 (36.1) 268/807 (33.2) 413/1204 (34.3) 408/1217 (33.5) Intramedullary nail

59/2421 (2.4) 22/814 (2.7) 20/800 (2.5) 17/807 (2.1) 25/1204 (2.1) 34/1217 (2.8) External fixator

1217/2421 (50.3) 397/814 (48.8) 391/800 (48.9) 429/807 (53.2) 604/1204 (50.2) 613/1217 (50.4) Plate

308/2421 (12.7) 125/814 (15.4) 97/800 (12.1) 86/807 (10.7) 156/1204 (13.0) 152/1217 (12.5) Other internal fixation

16/2421 (0.7) 6/814 (0.7) 3/800 (0.4) 7/807 (0.9) 6/1204 (0.5) 10/1217 (0.8) Other

patients in the soap group, for 15 in the saline group, for 5 in the very-low-pressure group, for 9 in the low-pressure group, and for 13 in the high-pressure group. Full details regarding

(P = 0.03) and the volume of irrigation solution (P = 0.01 for total; P = 0.02 for grade I or II, and P = 0.03 for grade III). Except for definitive fracture fixation, all surgical and perioperative

group, for 1 in the low-pressure group, and for 2 in the high-pressure group. Data on time from injury to the first incision were missing for 35 patients in the soap group, for 38 in the

treatment items listed refer to the initial irrigation and débridement. Data on age were missing for 1 patient in the soap group, for 4 in the saline group, for 2 in the very-low-pressure

saline group, for 14 in the very-low-pressure group, for 27 in the low-pressure group, and for 32 in the high-pressure group. Data on volume of irrigation solution were missing for 12

no significant between-group differences in the characteristics listed here, except for differences between the saline group and the soap group in the Gustilo–Anderson fracture grade

baseline characteristics and surgical approaches are available in Tables S1 and S2 in the Supplementary Appendix.

n engl j med 373;27 nejm.org December 31, 2015 2635

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

P Value†Hazard Ratio (95% CI) (N = 826) (N = 809) (N = 812) End Point

0.15 1.31 (0.79–2.19) 0.61 (0.38–1.00) 0.81 (0.52–1.27) 26 (3.1) 35 (4.3) 43 (5.3) Nonoperatively treated wound-healing problem

0.63 0.82 (0.47–1.44) 0.94 (0.55–1.59) 0.77 (0.44–1.33) 29 (3.5) 23 (2.8) 30 (3.7) Nonoperatively treated nonunion or delayed

0.80 0.92 (0.70–1.20) 1.02 (0.78–1.33) 0.93 (0.71–1.23) 109 (13.2) 103 (12.7) 111 (13.7) Primary end point: reoperation within 12 mo for

0.87 0.95 (0.66–1.38) 1.11 (0.76–1.62) 1.06 (0.72–1.54) 56 (6.8) 56 (6.9) 54 (6.7) Irrigation and débridement for wound infection

0.57 1.24 (0.70–2.17) 0.74 (0.43–1.30) 0.92 (0.55–1.55) 22 (2.7) 28 (3.5) 30 (3.7) Wound-coverage procedure for infected or

0.44 0.69 (0.39–1.22) 1.21 (0.69–2.11) 0.83 (0.45–1.52) 28 (3.4) 20 (2.5) 24 (3.0) Bone graft for established nonunion in patient

0.41 0.79 (0.45–1.40) 0.87 (0.51–1.47) 0.69 (0.40–1.19) 26 (3.1) 22 (2.7) 31 (3.8) Implant-exchange procedure for established non

0.10 1.08 (0.73–1.60) 0.69 (0.48–1.00) 0.75 (0.52–1.07) 49 (5.9) 53 (6.6) 69 (8.5) Nonoperatively treated infection at wound site

————3 (0.4) 2 (0.2) 3 (0.4) Revision and closure for wound dehiscence

————0 1 (0.1) 0 Drainage of a hematoma

————8 (1.0) 2 (0.2) 11 (1.4) Reoperation for hardware failure probably related

————7 (0.8) 3 (0.4) 6 (0.7) Intramedullary nail dynamizations in the

————5 (0.6) 4 (0.5) 3 (0.4) Fasciotomies for the compartment syndrome

————18 (2.2) 15 (1.9) 16 (2.0) Other reoperation

The Cox regressions were stratified according to randomly assigned irrigation solution, Gustilo–Anderson fracture grade (I or II vs. III), and study center. Cox regressions were per

Low vs. High High vs. Very Low Low vs. Very Low

Pressure Pressure Pressure

High Low Very Low

number of patients (percent)

Table 2. Study End Points for the Comparison of Irrigation Pressures.*

formed only when there were more than 50 end-point events.

† Overall P values are for the three-way comparison.

infection, wound healing, or bone healing

to infection, wound-healing problem, or

union in patient with fracture gap <1 cm

Primary end-point components

with fracture gap <1 cm

bone-healing problem

operating room

necrotic wound

Secondary end points

union

2636 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society.

Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

secondary end points (nonoperatively managed infection, wound-healing problem, and bone-healing problem) were observed between the two irrigation solutions or among the three ir-rigation pressures.

A possible effect modification was observed in two subgroups: subgroup analyses suggested that very low pressure was superior to low or high pressure in patients with a tibial fracture but inferior in patients with other fractures (P = 0.05 for interaction) and that saline solution was superior to soap when antibiotics were given for less than 4 days after surgery, but saline and soap were similar when antibiotics were given for 4 or more days (P = 0.03 for interaction). Be-cause many subgroup analyses were performed, the positive results in these two subgroups have relatively low credibility.30

Our study had several strengths. We included safeguards against potential bias (concealed randomization and concealment of study-group assignments from patients, end-point adjudica-tors, and data analysts) and safeguards against interpretation bias.31 The study also had broad inclusion criteria with a large number of centers in countries with diverse health care systems, as well as a focus on an end point (i.e., reoperation) that is of unequivocal importance to both pa-tients and the health care system.

Our study has certain limitations. Although the surgeons had high adherence to the initial irrigation protocol during the index surgery, the rate of surgeons’ adherence to the originally as-signed pressure and solution for the 615 patients who required a secondary operative irrigation and débridement decreased to 75.9% for irrigation pressure and 79.3% for irrigation solution. This level of adherence is still relatively high, and we found a detrimental effect of soap that any inter-vention crossover would tend to obscure.

Our estimates of the treatment effect of high versus low pressure, high versus very low pres-sure, and low versus very low pressure, although close to 1.0, exclude large differences but do not exclude small but potentially important differ-ences. For instance, on the basis of the 95% confidence intervals, our results are consistent with high-pressure irrigation resulting in either 22% fewer operations or 33% more operations as compared with very low pressure.

A total of 10% of the patients who underwent randomization were not followed to 12 months; the survival analysis included available data for

A Freedom from Event, According to Irrigation Pressure

1.00

Probablity of Freedom from Event Probablity of Freedom from Event

0.90

1.00

0.80

0.95

0.70

0.90

0.60

0.85

0.50

0.40 0.80

Very low pressure

0.30 0.75

Low pressure

0.20 0.70 High pressure

0.10 0.00

0 3 6 9 12

0.00

0 3 6 9 12

Months

No. at Risk

Very low 812 704 661 623 608

Low 809 696 667 633 615

High 826 715 678 641 616

B Freedom from Event, According to Irrigation Solution

1.00

0.90

1.00

0.80

0.95

0.70

0.90

0.60

0.85

0.50

0.40 0.80

0.30 0.75 Saline

0.20 0.70 Soap

0.10 0.00

0 3 6 9 12

0.00

0 3 6 9 12

Months

No. at Risk

Soap 1229 1055 995 936 905

Saline 1218 1060 1011 961 934

Figure 1. Kaplan–Meier Estimates of Freedom from the Primary End Point.

Panel A shows the Kaplan–Meier estimates of the probability of freedom

from the primary end point (reoperation within 12 months after the index

surgery) according to irrigation pressure (very low pressure, 1 to 2 psi; low

pressure, 5 to 10 psi; or high pressure, >20 psi). Tick marks indicate censored

data. The number at risk reported at 12 months includes patients whose

12-month visit was completed between 11 months and 12 months. Clinical

sites were permitted to complete the 12-month visit between 11 months and

12 months, provided that the patient’s wound and fracture had previously

been deemed to be healed and that the patient would not be returning for any

further clinical assessment. Panel B shows the Kaplan–Meier estimates of the

probability of freedom from the primary end point according to irrigation so

lution. In each panel, the inset shows the same data on an enlarged y axis.

these patients. The finding of the superiority of

saline to soap was not robust to a sensitivity

analysis that assumed that patients in the saline

group who were lost to follow-up had a risk of

event that was twice as high as the risk among

those with complete follow-up. This analysis

reduces our strength of inference that soap is

n engl j med 373;27 nejm.org December 31, 2015 2637

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

Table 3. Study End Points for the Comparison of Irrigation Solutions.*

Soap Saline Hazard Ratio

End Point (N = 1229) (N = 1218) (95% CI) P Value

number of patients (percent)

Primary end point: reoperation within 12 mo for 182 (14.8) 141 (11.6) 1.32 (1.06–1.66) 0.01

infection, wound healing, or bone healing

Primary end-point components

Irrigation and débridement for wound infection 87 (7.1) 79 (6.5) 1.11 (0.81–1.51) 0.53

Revision and closure for wound dehiscence 8 (0.7) 0 — —

Wound-coverage procedure for infected or necrotic 45 (3.7) 35 (2.9) 1.33 (0.85–2.08) 0.22

wound

Drainage of hematoma 1 (0.1) 0 — —

Reoperation for hardware failure probably related 11 (0.9) 10 (0.8) — —

to infection, wound-healing problem, or bone-

healing problem

Bone graft for established nonunion in patient with 39 (3.2) 33 (2.7) 1.25 (0.78–2.01) 0.35

fracture gap <1 cm

Implant-exchange procedure for established non48 (3.9) 31 (2.5) 1.59 (1.01–2.51) 0.046

union in patient with fracture gap <1 cm

Intramedullary nail dynamizations in the operating room 8 (0.7) 8 (0.7) — —

Fasciotomies for the compartment syndrome 7 (0.6) 5 (0.4) ——

Other 25 (2.0) 24 (2.0) ——

Secondary end points

Nonoperatively treated infection at wound site 82 (6.7) 89 (7.3) 0.97 (0.71–1.31) 0.82

Nonoperatively treated wound-healing problem 51 (4.1) 53 (4.4) 0.97 (0.65–1.43) 0.86

Nonoperatively treated nonunion or delayed union 41 (3.3) 41 (3.4) 1.02 (0.65–1.58) 0.94

The Cox regressions were stratified according to randomly assigned irrigation pressure, Gustilo–Anderson fracture grade

(I or II vs. III), and study center. Cox regressions were performed only when there were more than 50 end-point events.

inferior, but it does not undermine the conclu-sion that soap is no better than saline. The use of a single concentration of soap solution limit-ed our ability to explore a potentially efficacious dose. It is plausible that the soap-solution con-centration (0.45%) was too high and that a lower concentration might have been effective. Our choice of castile soap and dosing was, however, based on a large body of experimental evi-dence,7,14-19 a recent clinical trial that used this formulation21 without adverse effects, and our pilot study, which suggested its safety.27

Our trial defined the highest pressure catego-ry as 20 psi or higher, whereas prior experimen-tal studies have used pressures of more than 50 psi. Our cutoff points for pressure were based on a prior survey of surgeons and on the American College of Surgeons definition of high pressure as 15 to 35 psi and low pressure as 1 to 15 psi. We further subcategorized the low-pressure cate-gory to low (5 to 10 psi) and very low (1 to 2 psi), given the available settings on the handheld, battery-operated irrigators in this trial.

A prior randomized trial addressed the relative effect of irrigation pressures on patient-impor-tant outcomes.32 That trial, which involved 335 patients who presented to the emergency depart-ment with open wounds within 24 hours after injury, compared pressures of 13 psi (intermedi-ate between our low and high pressures) with very low pressure (1 to 2 psi) administered with a bulb syringe. The authors found a significantly lower rate of wound infection with the higher pressure (1.3% vs. 6.9%, P = 0.02). That study did not conceal randomization assignments, did not blind the assessment of infections, and had a high loss to follow-up (19% of patients). The methodologic differences between that study and ours may explain the differences in outcomes.

Our results challenge the results of prior studies, guidelines, and a large body of experi-mental evidence that have favored higher pres-sures (typically >20 psi) for the effective removal of contaminants. A number of nursing guide-lines recommend high-pressure irrigation.33-35 Al-though surgical and orthopedic organizations

2638 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

Subgroup Hazard Ratio (95% CI)

Overall 1.32 (1.06–1.66)

Gustilo–Anderson fracture grade

I or II 1.20 (0.89–1.61)

III 1.47 (1.05–2.06)

Extremity

Upper 0.92 (0.49–1.76)

Lower 1.42 (1.12–1.81)

Fracture gap

<1 cm 1.27 (0.99–1.62)

≥1 cm 1.60 (0.84–3.07)

Tibial injury

No 1.16 (0.83–1.61)

Yes 1.56 (1.14–2.13)

Definitive fixation

Intramedullary nail 1.72 (1.16–2.55)

External fixator 1.39 (0.89–2.17)

Plate 1.00 (0.65–1.54)

Articular involvement

Intraarticular 1.20 (0.84–1.70)

Extraarticular 1.42 (1.05–1.92)

Surgical preparation in emergency department

Yes 1.18 (0.72–1.95)

No 1.36 (1.06–1.76)

Surgical preparation solution

Chlorhexidine 1.29 (0.92–1.80)

Iodine 1.34 (0.94–1.91)

Alcohol or other 1.55 (0.80–3.00)

Adequate volume of fluid

Grade III fracture

No 1.92 (0.16–22.53)

Yes 1.36 (1.00–1.85)

Grade I or II fracture

No —

Yes 1.21 (0.86–1.70)

Wound contamination

Mild 1.32 (1.00–1.76)

Moderate 1.47 (0.90–2.39)

Severe 1.48 (0.74–2.95)

Duration of antibiotic use within

first wk after surgery

<4 days 1.87 (1.28–2.74)

≥4 days 1.08 (0.81–1.44)

Time from injury to surgery

<6 hr 0.93 (0.58–1.48)

6–12 hr 1.59 (1.12–2.26)

>12 hr 1.31 (0.87–1.98)

0.5 0.7 1.0 2.0 3.0 4.0

Soap Better Saline Better

Figure 2. Subgroup Analyses of the Primary End Point, According to Irrigation Solution.

The primary end point was reoperation within 12 months after the index surgery. Arrows indicate that the upper

and lower limits of the 95% confidence interval are not shown.

do not provide guidance on irrigation pressures, several expert authors suggest the use of irriga-tion pressures between 8 and 12 psi.5,6

Experimental evidence has suggested that irri-gation pressures of less than 10 psi are ineffec-tive in removing soil contaminants from con-

taminated open wounds.5 Experimental studies have also shown that high-pressure irrigation is more effective than low-pressure irrigation in the removal of bacteria, especially when the time to irrigation was delayed beyond 6 hours.36 In con-trast, some experimental studies have shown com-

n engl j med 373;27 nejm.org December 31, 2015 2639

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

The new engl and jour nal of medicine

plications from high-pressure irrigation, including increased damage to fractured bone,37 bacterial propagation into soft tissues and the intramedul-lary canal of the fractured bone,11,37 promotion of stem-cell differentiation from bone-forming cells (osteoblasts) toward the adipocyte cell type,38 and impairment of in vivo fracture healing.12 Our results suggest that findings from experimental studies do not always translate into differences in patient-important outcomes in clinical practice.

With regard to irrigation solutions, our find-ings contrast with those of prior experimental studies1-7,14-21,27 in laboratory and animal models that showed soap solution to be more effective than normal saline in removing bacteria and particulate matter from wounds and bone,7,14,16,17 without toxic effects to soft tissues and bone.7 One trial involving 400 patients showed that, at a mean follow-up of 1.3 years, a 0.45% soap solution was associated with a lower risk of in-fection than the risk with an antibiotic solution (100,000 U of bacitracin per 3 liters of normal saline) (13% vs. 18%; relative risk, 0.74; 95% CI, 0.45 to 1.26),21 and a lower risk of wound-healing complications (4% vs. 10%, P = 0.03). That trial, however, had unblinded outcome adjudication and appeared to have unconcealed randomiza-tion; there was also bias in that soap was com-pared with normal saline that contained antibiotics. The point estimate from our randomized FLOW pilot trial, which involved 111 patients, also favored the soap solution over normal saline (hazard ratio, 0.77; 95% CI, 0.35 to 1.69).27

Some experimental data support the results of our clinical trial. In an established animal model of a contaminated complex musculoskeletal wound, the initial reduction in pseudomonas bac-terial counts was greater when wound irrigation was performed with castile soap than when it was performed with normal saline (with counts reduced to 13% vs. 29% of the pretreatment level),39 but at 48 hours, bacterial counts in the soap group increased to 120% of the pretreat-ment levels, whereas the bacterial counts with normal saline solution were 68% of the pretreat-ment levels. Similarly, investigators using a Staphy-lococcus aureus–contaminated rat-femur model have suggested that host-tissue toxicity and necrosis from antibacterial solutions allow bacteria to

thrive and bacterial levels to rebound to pretreat-ment levels.40

Our study may have implications for the care of patients with open fractures worldwide and may inform protocols for the management of wound irrigation for paramedics, nurses, emer-gency physicians, and surgeons caring for pa-tients with open fractures. Our findings may be particularly relevant for low-income and middle-income countries, in which 90% of the road traffic fatalities globally, and probably a similar percentage of open fractures, occur.41 In such contexts, the knowledge that there is no benefit to the use of irrigation-pressure devices can guide the allocation of limited resources — a result that is also very important for the man-agement of open fractures in combat settings.

In conclusion, our results suggest similar re-operation rates regardless of irrigation pressure and establish very low pressure as an acceptable, low-cost alternative in the irrigation of open fractures. Our findings indicate that saline was superior to castile soap solution for the routine irrigation of acute open fractures.

The views expressed in this article are those of the authors and are not necessarily endorsed by the Department of Defense.

Supported by research grants from the Canadian Institutes of Health Research (MCT-93173), the U.S. Army Institute of Surgical Research Orthopedic Trauma Research Program (W81X-WH-08-1-0473), U.S. Army Institute of Surgical Research Peer Reviewed Orthopedic Research Program (W81XWH-12-1-0530), and Association Internationale pour l’Ostéosynthèse Dynamique.

Dr. Bhandari reports receiving consulting fees from Stryker, Smith & Nephew, Amgen, Eli Lilly, DJO Global, Zimmer, and Ferring Pharmaceuticals, and grant support from Stryker, Amgen, DePuy Synthes, Eli Lilly, and DJO Global; Dr. Jeray, receiving consulting fees from Zimmer and Eli Lilly and lecture fees from Zimmer and AO North America; Dr. Devereaux, receiving grant support from Abbott Diagnostics, Boehringer Ingelheim, Covidien, Octapharma, Roche Diagnostics, and Stryker; Dr. Schemitsch, receiving consulting fees from Stryker, Smith & Nephew, Amgen, Sanofi Aventis, Bioventus, Acumed, and Celgene, royalties from Stryker related to commercial products (clavicle and elbow plates), grant support from the Association for the Study of Inter-nal Fixation and OMeGA Medical Grants Association, and non-specific research and fellowship support through his institution from Stryker, Smith & Nephew, Zimmer, and DePuy Synthes; Dr. Della Rocca, receiving consulting fees from Bioventus and lecture fees and grant support from DePuy Synthes, and holding stock in Mergenet Medical, Amedica, and the Orthopaedic Implant Com-pany; Dr. Sprague, being an employee of Global Research Solu-tions; and Dr. Tornetta, receiving royalties from a patent (U.S. patent no. 88888824 B2) related to a system and methods for us-ing polyaxial plates that is licensed to Smith & Nephew. No other potential conflict of interest relevant to this article was reported.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

Appendix

The affiliations of the members of the writing group are as follows: Division of Orthopaedic Surgery, the Department of Surgery (M.B., B.A.P., S.S.), the Departments of Clinical Epidemiology and Biostatistics (M.B., P.J.D., D.H.-A., P.M., S.S., S.W., G.H.G.) and Medicine (P.J.D., G.H.G.), and the Population Health Research Institute (P.J.D.), McMaster University, Hamilton, ON, the Division of Orthopaedic Surgery (H.K., E.H.S.) and Health Policy Management and Evaluation (H.K.), University of Toronto, and Sunnybrook Health Sciences Centre (H.K.),

2640 n engl j med 373;27 nejm.org December 31, 2015

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on March 16, 2025. For personal use only.

Irrigation in Open Fracture Wounds

Toronto, the Division of Orthopaedic Surgery, University of Ottawa, Ottawa, ON (S.P.), the Department of Orthopaedics, Faculty of Medicine,

University of British Columbia, Vancouver (T.B.S.), and the Department of Surgery, University of Manitoba, Winnipeg (T.T.) — all in Canada;

the Department of Orthopedic Surgery, Greenville Health System, Greenville, SC (K.J.J., S.L.T.); the Department of Orthopedic Surgery, In-

diana University, Indianapolis (J.A.); the Department of Orthopedic Surgery, University of Missouri School of Medicine, Columbia (G.J.D.R.);

Center for Orthopedic Research and Education Institute, Phoenix, AZ (C.J.); the Department of Surgery, Monash University, Melbourne, VIC,

Australia (S.L.); Sancheti Institute for Orthopedics and Rehabilitation, Pune, India (P.S.); Chinese Evidence-based Medicine Center, Sichuan

University, Chengdu, China (X.S); and the Department of Orthopedic Surgery, Boston University School of Medicine, Boston (P.T.).

References

1. Anglen JO. Wound irrigation in muscu- 16. Burd T, Christensen GD, Anglen JO, 28. Pocock SJ. Clinical trials: a practical

loskeletal injury. J Am Acad Orthop Surg Gainor BJ, Conroy BP, Simpson WA. approach. Toronto: John Wiley, 1984.

2001;9:219-26. Sequential irrigation with common de- 29. Akl EA, Briel M, You JJ, et al. Potential

2. Gustilo RB, Merkow RL, Templeman D. tergents: a promising new method for impact on estimated treatment effects of

The management of open fractures. J Bone decontaminating orthopedic wounds. information lost to follow-up in random

Joint Surg Am 1990;72:299-304. Am J Orthop (Belle Mead NJ) 1999;28: ised controlled trials (LOST-IT): system-

3. Foote CJ, Guyatt GH, Vignesh KN, et al. 156-60. atic review. BMJ 2012;344:e2809.

Which surgical treatment for open tibial 17. Conroy BP, Anglen JO, Simpson WA, 30. Sun X, Briel M, Walter SD, Guyatt GH.

shaft fractures results in the fewest reop- et al. Comparison of castile soap, benzal- Is a subgroup effect believable? Updating

erations? A network meta-analysis. Clin konium chloride, and bacitracin as irriga- criteria to evaluate the credibility of sub-

Orthop Relat Res 2015;473:2179-92. tion solutions for complex contaminated group analyses. BMJ 2010;340:c117.

4. Mundi R, Chaudhry H, Niroopan G, orthopaedic wounds. J Orthop Trauma 31. Järvinen TL, Sihvonen R, Bhandari M,

Petrisor B, Bhandari M. Open tibial frac- 1999;13:332-7. et al. Blinded interpretation of study results

tures: updated guidelines for management. 18. Gainor BJ, Hockman DE, Anglen JO, can feasibly and effectively diminish inter-

JBJS Rev 2015;3(2):e1. Christensen G, Simpson WA. Benzalkoni- pretation bias. J Clin Epidemiol 2014;67:

5. Barnes S, Spencer M, Graham D, John- um chloride: a potential disinfecting irri- 769-72.

son HB. Surgical wound irrigation: a call gation solution. J Orthop Trauma 1997;11: 32. Longmire AW, Broom LA, Burch J.

for evidence-based standardization of prac- 121-5. Wound infection following high-pressure

tice. Am J Infect Control 2014;42:525-9. 19. Tarbox BB, Conroy BP, Malicky ES, et al. syringe and needle irrigation. Am J Emerg

6. Crowley DJ, Kanakaris NK, Giannoudis Benzalkonium chloride: a potential disin- Med 1987;5:179-81.

PV. Irrigation of the wounds in open frac- fecting irrigation solution for orthopaedic 33. Storer A, Lindauer C, Proehl J, et al.

tures. J Bone Joint Surg Br 2007;89:580-5. wounds. Clin Orthop Relat Res 1998; Clinical practice guideline: wound prepara-

7. Bhandari M, Adili A, Schemitsch EH. (346):255-61. tion, 2011 (https://www.ena.org/practice

The efficacy of low-pressure lavage with 20. Huyette DR, Simpson WA, Walsh R, -research/research/CPG/Documents/

different irrigating solutions to remove et al. Eradication by surfactant irrigation WoundPreparationCPG.pdf).

adherent bacteria from bone. J Bone Joint of Staphylococcus aureus from infected 34. Ackley BJ, Ladwig GB, Swan BA, Tucker

Surg Am 2001;83-A:412-9. complex wounds. Clin Orthop Relat Res SJ. Evidence-based nursing care guidelines:

8. Bhandari M, Schemitsch EH, Adili A, 2004;427:28-36. medical-surgical interventions. St. Louis:

Lachowski RJ, Shaughnessy SG. High and 21. Anglen JO. Comparison of soap and Mosby Elsevier, 2007:954.

low pressure pulsatile lavage of contami- antibiotic solutions for irrigation of lower- 35. Fernandez R, Griffiths R, Ussia C. Ef-

nated tibial fractures: an in vitro study of limb open fracture wounds: a prospective, fectiveness of solutions, techniques and

bacterial adherence and bone damage. randomized study. J Bone Joint Surg Am pressure in wound cleansing. JBI Reports

J Orthop Trauma 1999;13:526-33. 2005;87:1415-22. 2004;2:231-70.

9. Gross A, Bhaskar SN, Cutright DE, 22. Kaysinger KK, Nicholson NC, Ramp 36. Bhandari M, Thompson K, Adili A,

Beasley JD III, Perez B. The effect of pul- WK, Kellam JF. Toxic effects of wound irri- Shaughnessy SG. High and low pressure

sating water jet lavage on experimental gation solutions on cultured tibiae and os- irrigation in contaminated wounds with

contaminated wounds. J Oral Surg 1971; teoblasts. J Orthop Trauma 1995;9:303-11. exposed bone. Int J Surg Investig 2000;2:

29:187-90. 23. Sprung J, Schedewie HK, Kampine JP. 179-82.

10. Bhhaskar SN, Cutright DE, Runsuck Intraoperative anaphylactic shock after 37. Bhandari M, Adili A, Lachowski RJ.

EE, Gross A. Pulsating water jet devices in bacitracin irrigation. Anesth Analg 1990; High pressure pulsatile lavage of contam-

debridement of combat wounds. Mil Med 71:430-3. inated human tibiae: an in vitro study. J

1971;136:264-6. 24. FLOW Investigators. Fluid Lavage of Orthop Trauma 1998;12:479-84.

11. Hassinger SM, Harding G, Wong- Open Wounds (FLOW): design and rationale 38. Bhandari M, Schemitsch EH. High-

worawat MD. High-pressure pulsatile la- for a large, multicenter collaborative 2 x 3 pressure irrigation increases adipocyte-

vage propagates bacteria into soft tissue. factorial trial of irrigating pressures and like cells at the expense of osteoblasts in

Clin Orthop Relat Res 2005;439:27-31. solutions in patients with open fractures. vitro. J Bone Joint Surg Br 2002;84:1054-

12. Polzin B, Ellis T, Dirschl DR. Effects BMC Musculoskelet Disord 2010;11:85. 61.

of varying pulsatile lavage pressure on 25. Fergusson D, Aaron S, Guyatt GH, 39. Owens BD, White DW, Wenke JC.

cancellous bone structure and fracture Hebert P. Post-randomization exclusion Comparison of irrigation solutions and

healing. J Orthop Trauma 2006;20:261-6. of patients enrolled in clinical trials: the devices in a contaminated musculoskele-

13. Petrisor B, Jeray K, Schemitsch E, et al. intention-to-treat principle does not ne- tal wound survival model. J Bone Joint

Fluid lavage in patients with open fracture cessitate that all patients be analyzed. Surg Am 2009;91:92-8.

wounds (FLOW): an international survey BMJ 2002;325:652-4. 40. Penn-Barwell JG, Murray CK, Wenke

of 984 surgeons. BMC Musculoskelet Dis- 26. Bhandari M, Guyatt G, Tornetta P III, JC. Comparison of the antimicrobial ef-

ord 2008;9:7. et al. Randomized trial of reamed and un- fect of chlorhexidine and saline for irri-

14. Anglen JO, Apostoles S, Christensen reamed intramedullary nailing of tibial gating a contaminated open fracture

G, Gainor B. The efficacy of various irri- shaft fractures. J Bone Joint Surg Am model. J Orthop Trauma 2012;26:728-32.

gation solutions in removing slime-pro- 2008;90:2567-78. 41. Global status report on road safety:

ducing staphylococcus. J Orthop Trauma 27. Petrisor B, Sun X, Bhandari M, et al. time for action. Geneva: World Health Or-

1994;8:390-6. Fluid Lavage of Open Wounds (FLOW): ganization, 2009 (http://www.who.int/

15. Anglen JO, Gainor BJ, Simpson WA, a multicenter, blinded, factorial pilot trial violence_injury_prevention/road_safety_

Christensen G. The use of detergent irri- comparing alternative irrigating solutions status/2009).

Orthop 2003;27:40-6. tures. J Trauma 2011;71:596-606.

n engl j med 373;27 nejm.org December 31, 2015 2641

The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society.

Downloaded from nejm.org on March 16, 2025. For personal use only.