Incidence, Timing, and Risk Factors for Secondary

SCIENTIFIC ARTICLE

Incidence, Timing, and Risk Factors for Secondary Revision After Primary Revision of Traumatic Digit Amputations Andrew P. Harris, MD,* Avi D. Goodman, MD,* Joseph A. Gil, MD,* Andrew D. Sobel, MD,* Neill Y. Li, MD,* Jeremy Raducha, MD,* Grayson L. Baird, PhD,† Julia A. Katarincic, MD* Purpose Primary revision amputation is the most common treatment method for traumatic digit amputations in the United States. Few studies have reported secondary revision rates after primary revision amputation. The primary aim of our study was to identify risk factors for secondary revision within 1 year of the index procedure. Secondarily, we describe the incidence and timing of complications requiring secondary revision. Methods Our institution’s emergency department (ED) database was reviewed for traumatic digit amputations over a 6-year period. Patients were reviewed for demographic characteristics, comorbidities, site of treatment (ED versus operating room), and complications requiring secondary revision. Conditional Cox Proportional Hazard regression was used to model hazard of revision within 1 year of index procedure relative to site of initial management, mechanism of injury, injury characteristics, and patient demographics. Results Five hundred and thirty-seven patients with 677 digits were managed with primary revision amputation. Five hundred and eighty-six digits (86.6%) were revised in the ED, and 91 (13.4%) in the operating room. Ninety-one digits required secondary revision, including 83 within 1 year. No increased risk of secondary revision amputation within 1 year of the index procedure was observed for patients treated in the ED compared with the operating room. Relative to crush injuries, bite and sharp laceration amputations had 4.8 times and 2.6 times increased risk of secondary revision, respectively. The index finger had a 5.3-fold increased risk of revision with the thumb as the reference digit. Work-related injuries had a 1.9-fold increased risk of secondary revision compared with nonework-related injuries. Conclusions No evidence was found indicating that traumatic digit amputations primarily revised in the ED had an increased risk of secondary revision. Patients may be counseled on the risk of secondary procedures based on the mechanism of injury, injury characteristics and demographics, as well as the timing of complications. (J Hand Surg Am. 2018;-(-):1.e1-e11. Copyright Ó 2018 by the American Society for Surgery of the Hand. All rights reserved.) Type of study/level of evidence Therapeutic IV. Key words Amputation, digit, finger, reoperation, revision.

From the *Department of Orthopaedics, Alpert Medical School of Brown University; and †Lifespan Biostatistics Core, Rhode Island Hospital, Providence, RI. Received for publication June 24, 2017; accepted in revised form March 19, 2018. No benefits in any form have been received or will be received related directly or indirectly to the subject of this article.

Corresponding author: Andrew P. Harris, MD, Department of Orthopaedics, Alpert Medical School of Brown University, 593 Eddy Street, Providence, RI 02903; e-mail: aharri26@gmail. com. 0363-5023/18/---0001$36.00/0 https://doi.org/10.1016/j.jhsa.2018.03.028

Ó 2018 ASSH

r

Published by Elsevier, Inc. All rights reserved.

FLA 5.5.0 DTD YJHSU55469_proof 20 April 2018 2:23 am ce

r

1.e1

1.e2

H

RISK FOR SECONDARY REVISION AMPUTATION

account for approximately 4.8 million annual visits to emergency departments in the United States, and result from a variety of injury mechanisms including snow blowers, table saws, and lawnmowers.1,2 Traumatic digit amputations make up a considerable component of this injury burden.1,3 This is also a global problem, as other developed countries see approximately 10.2 digit amputations per 100,000 person-years.4 Treatment strategies include revision amputation or replantation, and management varies based on injury characteristics, patient factors, resources, and preference.5,6 Treatment strategies also have regional and cultural components: in the United States, revision amputation is the most common treatment, whereas in many Asian countries, replantation is the treatment of choice given strong beliefs in body preservation and cosmesis.7 Treatment also depends in part on the injury mechanism—a sharp laceration is more likely amenable to replantation than a digit mangled in a saw. Revision amputation can take place either in the operating room (OR) or in the emergency department (ED). Although performing a procedure in the OR has the theoretical advantages of improved visualization, instrument availability, patient relaxation, and treatment by an attending hand surgeon, the ED may offer a venue for more expedient, cost-effective care. Given the paucity of literature regarding complications and reoperation rates after primary revision amputation, our primary objective was to identify risk factors for, and complications requiring, secondary revision amputation.8,9 We hypothesized that amputations primarily revised in the ED have an increased risk of secondary revision compared with those treated in the OR. Our second objective was to describe the incidence and timing of complications, as well as to identify risk factors for secondary revision after primary revision amputation.

database includes records from 2 hospitals: a level 1 trauma center and level 2 trauma center/community hospital. Treatment was performed in either hospital’s ED, hospital-owned ambulatory surgery centers, or an outpatient surgery center not affiliated with the hospital. Codes from the International Classification for Diseases 9th and 10th editions constituted the initial search criteria (Appendix A, available on the Journal’s Web site at www.jhandsurg.org). Inclusion criteria was defined as any patient presenting with partial or complete amputations through Verdan flexor tendon zone I or II receiving definitive care. Call days were shared equally between the Orthopedics and Plastic & Reconstructive Surgery (PRS) services. The primary treatment method of revision amputation for both services is identical with shortening of the exposed bone, traction neurectomies, trimming of the germinal matrix, and primary closure of the soft tissue with absorbable suture. Fingertip amputations without bone loss, or those whose treatment did not require removal of bone resulting in an amputation equivalent, were excluded. To ensure that patients at all stages of care were captured, our departmental billing databases were reviewed for patients billed with Current Procedural Terminology codes (Appendix B, available on the Journal’s Web site at www.jhandsurg.org) pertinent to the treatment of traumatic amputations or secondary revision amputations, and these additional charts were reviewed to ensure that the treated injuries met the inclusion criteria. Each patient’s chart was reviewed for demographics, injury mechanism, treatment, and subsequent procedures. Demographics included age, sex, race, date of injury, occupation, insurance status, and hand dominance. Other data included setting of injury (eg, work), initial consult service (Orthopedics or PRS), smoking status, and medical comorbidities (diabetes, hypertension/hyperlipidemia, coronary artery disease/myocardial infarction/transient ischemic attack/peripheral vascular disease, and chronic kidney disease). Injury characteristics included mechanism (crush, laceration, avulsion, bite, blast, saw, snow blower, lawnmower), finger(s) involved, and Verdan flexor tendon zone of amputation (I and/or II). Treatment data consisted of date, location (ED or OR), and details of the initial definitive management. The primary outcomes included the need for secondary revision amputation, the reason for revision (neuroma, nail deformity, infection, soft tissue coverage or exposed bone, cosmesis, function), and the date of revision surgery. In addition to standard descriptive statistics, conditional Cox Proportional Hazard regression with

AND AND FINGER INJURIES

METHODS After institutional review board approval, our institution’s ED database was retrospectively used to identify patients seen from January 2010 to December 2015 and treated for traumatic digit amputations through Verdan flexor tendon zones I and II. The level of amputation was determined by review of the patient’s radiographs and reported physical examination, and the level of amputation recorded by the resident in the initial consult note. The Verdan classification was implemented in this study given that it is well known and reliably reported at the time of injury. The ED J Hand Surg Am.

r

Vol. -, - 2018



sandwich estimation was used to model hazard of secondary revision, where digits were nested within patients and observations were censored at 1 year from index procedure if no secondary revision was recorded or a replantation occurred. Given that the variable “time” was not normally distributed, estimates of time were calculated using the median. Significance was established at the .05 level with all interval estimates calculated at 95% confidence.

ED, whereas 56 patients (10.4%) with 91 (48 zone I, 43 zone II) amputations were primarily revised in the OR. PRS managed 283 (49.5%) patients, compared with 289 (50.5%) by Orthopedics. Risk factors for secondary revision censored at 1 year Of the 537 patients with 677 primarily revised digit amputations, 81 (15.1% of patients) with 91 amputations required secondary revision (Table 2). Of these, 74 patients (13.8% of patients) with 83 amputations required secondary revision within 1 year of primary revision amputation. Calculating the increased risk of secondary revision within 1 year from the index procedure relative to crush injuries, amputations caused by bites were at 4.8 times increased risk of secondary revision (confidence interval [CI], 1.654e13.776) and those caused by lacerations were at 2.6 times increased risk (CI, 1.247e5.420). However, amputations caused by avulsion, lawnmower, saw, and snow blowers were not observed to be at higher risk for secondary revision. The index, middle, and small fingers had an increased secondary revision risk of 5.3-, 4.3-, and 4.5-fold (CI, 1.618e17.566, 1.292e14.049, 1.253e16.134), respectively, compared with the thumb. The thumb was chosen as the reference as it is the most important digit for hand function. The ring finger had an increased risk of 3-fold, though this only approached significance (CI, 0.919e10.938). No increased risk of secondary revision was observed for zone I compared with zone II injuries (CI, 0.532e3.808). Of the 74 patients with 83 primarily revised amputations that required secondary revision within 1 year of the index procedure, 68 patients were initially managed in the ED, whereas 6 patients were initially managed in the OR. No increased risk of secondary revision amputation within 1 year of primary revision amputation was observed for patients treated in the ED compared with the OR (hazard ratio, 0.723; CI, 0.145e1.148). Work-related injuries had a 1.9-fold increased risk of secondary revision relative to nonework-related injuries (CI, 1.06e3.422). Insured patients had a 1.6fold increased risk of secondary revision compared with the uninsured, although this only approached significance (CI, 0.903e2.672). PRS and Orthopedics each managed approximately half of all digit amputation patients during this time period. No increased risk of secondary revision was found comparing treatment groups (CI, 0.616e1.327).

RESULTS Demographics Five hundred and seventy-two patients (497 males, 86.9%), average age of 46.2 years (range, 1.5e98 y), incurred 719 traumatic digit amputations (Table 1). Most patients were Caucasian (n ¼ 442, 77.2%). Two hundred and eighty-four (49.7%) patients were insured, and 239 patients (41.8%) had work-related injuries. Injury characteristics The middle finger was the most frequently amputated digit (222, 30.8% of digits), followed by the index (185, 25.7%) and ring (153, 21.3%) fingers (Fig. 1). Verdan flexor tendon zone I amputations were more common than flexor tendon zone II amputations accounting for 609 (84.7%) and 110 (15.3%) of amputations, respectively. Amputations through zone I middle finger (193 digits) and zone I index finger (148 digits) were the most frequent locations of amputation. Crush (208, 36.4% of patients) and saw (147, 25.7% of patients) were the most common mechanisms of amputation (Fig. 2). Initial management Five hundred and thirty-two (93.0%) patients were initially managed with revision amputation alone. Fifteen patients with 17 complete digit amputations underwent replantation requiring revascularization. One of these patients underwent revision amputation of 1 digit in the ED and replantation of a separate digit in the OR (Fig. 3). A total of 12 patients with 12 incomplete amputations underwent repair not requiring revascularization. Four (0.7%) patients sustained multiple amputations and were initially managed with both primary revision amputation and replantation in the OR for different fingers. Thirteen (2.3%) patients were managed with secondary intention. Five hundred and thirty-seven (93.8%) patients received revision amputation of at least 1 digit (677 total digits) at initial treatment. Of these patients, 481 (89.6%) sustaining 586 (538 zone I, 48 zone II) amputations were primarily revised in the J Hand Surg Am.

1.e3

r

Vol. -, - 2018


1.e4


Incidence and timing of complications requiring secondary revision Seventy-four (91.5%) patients (Table 3) with 83 (91.2%) (Fig. 4) amputations required secondary revision within 1 year of primary revision amputation. The patient and digit incidence of complications occurring within 1 year of primary revision was 13.8% and 12.3%, respectively. Of these complications occurring within 1 year, isolated soft tissue coverage/necrosis, nail deformity, and neuroma were the most common reasons for revision with a patient incidence of 5.2%, 3.5%, and 1.3%, and returned to the OR for secondary revision at a median of 15.5, 132, and 147 days, respectively. Beyond the censored follow-up period of 1 year, 7 patients with 8 digits required secondary revision, for a total of 81 patients with 91 digits (84 zone I, 7 zone II). Graph representation of all complications by mechanism of injury is provided in Figure 5.

TABLE 1.

No. of Patients Gender Male Female

497 75

Ethnicity Caucasian African American Asian

442 37 4

Hispanic

70

Other

15

Not provided

3

Insurance status Insured

284

Uninsured

288

Hand dominance Right

414

Left

41

Unknown

DISCUSSION In our study, we reported complications resulting in secondary revision amputation as well as analyzed the risk of secondary revision amputation within 1 year of the index procedure for 537 patients with 677 digit amputations treated with primary revision amputation. The incidence of secondary revision within 1 year after primary revision amputation was 12.3% for digits and 13.8% for patients; these rates were similar for patients treated in the ED and the OR. Two studies have published reoperation rates after primary revision amputation.10,11 The study by Wilkens et al,10 focusing on primary revision of combined injury to 39 index finger amputations, reported a reoperation rate of 21% (8 digits), although the timing was not provided. In our study, we found that the overall index finger secondary revision rate after primary revision amputation was 17.1% (30 of 175), with 90% of these occurring within 1 year of index procedure. The high index finger secondary revision rate is likely related to complications giving discomfort with pinch, which may otherwise be tolerated by other digits. The second study that reported a secondary revision rate focused solely on symptomatic neuromas.11 In their study of 1,083 patients treated with primary revision amputation, Vlot et al11 found that 71 (6.6%) developed a symptomatic neuroma, in which 41 (4.3% of all patients) required a secondary revision for symptomatic neuroma. Our overall revision rate for neuroma was similar with 2.8% of patients and 2.4% of digits requiring secondary revision. J Hand Surg Am.

Patient Cohort Demographics

117

Work related Yes

239

No

267

Unknown

66

Tobacco use Yes

193

No

337

Unknown

42

Comorbidities Diabetes Cardiovascular disease HTN/HLD Chronic kidney disease

33 45 165 7

Profession Skilled laborer Unskilled laborer Service/retail Knowledge

71 112 59 8

Retired

71

Unemployed

17

Student

9

Infant/toddler

2

Unknown

223

HLD, hyperlipidemia; HTN, hypertension.

Dy et al12 found a 1.63-fold increased risk of reoperation after flexor tendon repair in patients with workers’ compensation compared with other forms of r

Vol. -, - 2018


1.e5

print & web 4C=FPO


FIGURE 2: Pie chart representation of the number of patients experiencing amputation by each mechanism (n ¼ 572).

print & web 4C=FPO

revisions for infection in this group, possibly due to prompt antibiotic treatment and debridement at presentation. The increased risk of secondary revision after sharp lacerations is due to the large number of zone I injuries through the nail bed, germinal matrix, and distal phalanx yielding nail deformities and soft tissue coverage issues. In contrast, the blast mechanism did not have any secondary revisions, likely because hand surgeons expect evolving tissue demarcation,15 and plan for return to the OR for adequate debridement. Low numbers of blast amputations may have also resulted in underdetection of the burden of secondary revisions in this group. Men comprised the majority (86.9%) of digit amputations, which is slightly higher than the 76.3% found by Conn et al.1 Their study analyzed only nonework-related injuries and may account for this difference. We found that the index (25.7%) and middle (30.8%) fingers were the most frequently amputated digits, similar to the results of Conn et al.1 Because their study focused on characterizing amputations and did not report on success rates of surgical interventions, they did not compare success rates of revision surgery based on the digit involved. Our higher risk of secondary revision of the index and middle fingers in reference to the thumb may be explained by our exclusion of amputations treated by replantation, and it is possible that many severe thumb injuries treated by replantation would have yielded higher rates of secondary revision had they been elected for primary revision amputation. Thumb amputation, because of its importance to prehensile activities, is the most commonly replanted digit,16 and it is likely that this was part of the discussion with patients before intervention at our institution, therefore potentially limiting the thumbs that could

FIGURE 1: Bar graph representation of the total number and percentage of each digit amputated.

insurance. Our results corroborated this, as patients with work-related amputations had a 1.9-fold increased risk of secondary revision compared with nonework-related amputations. Injury characteristics often dictate treatment. Crush was the most common mechanism of amputation in our cohort, causing amputation in 208 (36.4%) of all patients. This is consistent with the finding that crush is one of the top mechanisms of digit injuries.13 However, it was not the mechanism that resulted in the highest increased risk of secondary revision amputation, as bites and sharp lacerations were at an increased risk of 4.8 times and 2.6 times to require secondary revisions in reference to crush injury, respectively. This may be due to an underestimation of the need for secondary procedures with these mechanisms. Human and animal bites are known to cause infections including infectious flexor tenosynovitis, joint infections, and deep and superficial abscesses,14 so it would be expected that infections would be the indication for secondary revision after bite mechanism. Despite this, there were no secondary J Hand Surg Am.

r

Vol. -, - 2018


1.e6


FIGURE 3: Flowchart of amputation treatment by digit.

have progressed to secondary revision after primary revision amputation. No significantly increased risk for secondary revision of zone I compared with zone II amputations was observed. A potential for nail deformity to increase the risk of secondary revision of zone I injuries would be expected; however, we were likely underpowered to detect a difference given the low secondary revision rate. Other studies have reported complications after primary revision amputation, but have not assessed the incidence of complications requiring secondary revision.17,18 In our cohort, 74 (91.5%) of 81 patients and 83 (91.2%) of 91 digits with complications requiring secondary revision occurred within 1 year of the index procedure. Of the complications requiring secondary revision within 1 year, the most common complication was soft tissue coverage or necrosis reported as the single indication, which occurred in 28 patients (32 digits, 4.7% per-digit incidence). These patients returned to the OR at a median of 15.5 days after the primary procedure. Isolated nail deformity (19 patients, 19 digits, 2.8% per-digit incidence) and neuroma (7 patients, 8 digits, 1.2% digit incidence) were the second and third most common presenting complications, respectively, and often presented relatively late, returning to the OR at a median of 132 and 147 days, respectively. The timing of these complications is intuitive, though the incidence is similar to, or lower than, those previously reported.11,19,20 Vlot et al11 reported a similar patient incidence of 4.3% and median time to presentation of 3.4 months for symptomatic neuroma. Van der Avoort et al19 reported an incidence of neuroma at 7.8% after revision amputation of 177 digits, though Fisher and Boswick21 described an J Hand Surg Am.

incidence of 2.8% in their series of 144 digits treated with revision amputation and traction neurectomies. Complications presenting after 1 year were censored in the Cox Hazard regression analysis, as extended (over 1 y) follow-up was not available for all patients. Seven patients required secondary revision after 1 year, and it is uncertain why they presented later, but it is likely relative to the patient’s ability to tolerate the complication. Cosmesis and function were the most common complications presenting after 1 year, with 4 (57%) of these patients requiring secondary procedures for these complications. A systematic review by Yuan et al17 reported an incidence of infection of 11%, much higher than in our cohort (