Antibiotic Use in Hand Surgery: Surgeon Decision Making and Adherence to Available Evidence

Abstract

Background: There are no clearly defined guidelines from hand surgical societies regarding preoperative antibiotic prophylaxis. Many hand surgeons continue to routinely use preoperative prophylaxis with limited supporting evidence. This study aimed to determine for which scenarios surgeons give antibiotics, the reasons for administration, and whether these decisions are evidence-based. Methods: An anonymous 25-question survey was e-mailed to the 921-member American Society for Surgery of the Hand listserv. We collected demographic information; participants were asked whether they would administer antibiotics in a number of surgical scenarios and for what reasons. Respondents were broken into 3 groups based on when they said they would administer antibiotics: Group 1 (40 respondents) would give antibiotics in the case of short cases, healthy patients, without hardware; group 2 (9 respondents) would not give antibiotics in any scenario; and group 3 (129 respondents) would give antibiotics situationally. The Fisher exact test compared demographic variables, frequency of use, and indications of antibiotic prophylaxis. Results: Of the 921 recipients, 178 (19%) responded. Demographic variables did not correlate with the antibiotic use group. Operative case time >60 minutes, medical comorbidity, and pinning each increased antibiotic use. Group 1 respondents were more likely to admit that their practice was not evidence-based (74.4%) and that they gave antibiotics for medical-legal concern (75%). Twenty-two percent of respondents reported seeing a complication from routine prophylaxis, including Clostridium difficile infection. Conclusions: Antibiotics are still given unnecessarily before hand surgery, most often for medical-legal concern. Clear guidelines for preoperative antibiotic use may help reduce excessive and potentially inappropriate treatment and provide medical-legal support.

Preoperative antibiotic prophylaxis regimens have been well described in hip,^1-3 knee,^4,5 and shoulder arthroplasty,^6,7 as well as in orthopedic trauma.^8,9 There are no specific guidelines from hand surgical societies regarding preoperative antibiotic prophylaxis.^9,10 This may be, in part, due to the low rate of surgical site infection (SSI), which is generally reported as less than 1%.^11-17 Even considering this overall low incidence, studies have shown no difference in rates of infection after hand surgery between antibiotic prophylaxis and no prophylaxis.^10,18-21 Unnecessary routine antibiotic use may be detrimental to both the patient and the society as this practice may induce Clostridium difficile infection, allergic reaction, and antibiotic resistance.^22,23

There is no evidence to strongly support using prophylaxis in hand surgery cases for repair of lacerations,^24-26 clean hand injuries,^19,21 small joint replacements,^17,27 and complex hand trauma^28-30; for patients with associated medical comorbidities^17,27,31,32; or for cases that last less than 2 hours.^19,21 Even pinning of fractures may not require antibiotic prophylaxis.³³

Furthermore, the Agency for Healthcare Research and Quality has recommended against using antimicrobial prophylaxis for clean hand procedures without implantation of foreign materials (strength of evidence against prophylaxis: C).³⁴ Despite the evidence against routine antibiotic prophylaxis in hand surgery, antibiotics are given preoperatively in 10% to 30% of clean, elective hand procedures.^10,35,36 The factors associated with unnecessary antibiotic treatment in hand surgery are not known. The primary aims of this study are to determine for which scenarios surgeons give antibiotics, the reasons for administration, and whether these decisions are evidence-based.

Materials and Methods

After obtaining institutional review board approval, a 25-question survey was sent via e-mail to the 921-member American Society for Surgery of the Hand (ASSH) listserv. The listserv is an ASSH member and candidate member e-mail chain used to interact, discuss challenging cases, and share clinical and surgical tips. The ASSH listserv membership includes attending surgeons, fellows, and residents. We created our survey with SurveyMonkey and posted it to the listserv twice over the course of 6 weeks. A link was provided in the e-mail that redirected the anonymous participant to the survey. To encourage participation, we announced that a US $500 prize would be awarded to a randomly chosen participant. No participant from our home institution was eligible to win the prize. Surgeons were only allowed to answer the survey once.

The survey questions were developed by the primary and senior author and are provided in Figure 1. The survey questions sought to determine whether any demographic factors correlated with scenario response. The scenarios were modeled after the stratification of prior prospective studies¹⁸ and retrospective reviews.^10,11 We also asked whether the participant had already taken the survey, whether there was additional information they wanted to share, and for a contact address if they wanted to participate in the random drawing for the prize. Based on the survey responses, participants were divided into 3 groups. Group 1 chose to give antibiotics for question 8 scenario (in cases <30 minutes, a healthy patient, without pinning). Group 2 chose not to give antibiotics in any scenario. Group 3 chose to give antibiotics in some, but not all, scenarios.

Figure 1.

Participant survey.

Note. IV = intravenous.

Demographic variables, frequency of use by group, and indications of antibiotic prophylaxis were tabulated and compared using the Fisher exact test (Tables 1 and 2). For group 3, the varied use group, we reported percentages of antibiotic use for each scenario (Table 3). To determine which factors caused surgeons in group 3 to give antibiotics, we performed analysis using the McNemar test for comparison of marginal proportions of paired observations (Table 4).

Table 1.

Demographic Variables.

Variable	Total	Group 1 (n = 40)	Group 2 (n = 9)	Group 3 (n = 129)	P value
Level of experience (n = 178)					.40
Resident or fellow	6 (3.4)	1 (2.5)	0 (0)	5 (3.9)
In practice <5 years	38 (21.3)	8 (20)	0 (0)	30 (23.3)
In practice 5-20 years	60 (33.7)	15 (37.5)	2 (22.2)	43 (33.3)
In practice >20 years/retired	74 (41.6)	16 (40)	7 (77.8)	51 (39.5)
Completed fellowship (n = 172)					.82
No	1 (0.6)	0 (0)	0 (0)	1 (0.8)
Yes	171 (99.4)	39 (100)	9 (100)	123 (99.2)
In board collection period (n = 172)					.48
No	163 (94.8)	36 (92.3)	8 (88.9)	119 (96)
Yes	9 (5.2)	3 (7.7)	1 (11.1)	5 (4)
Type of residency (n =178)					.37
General surgery	6 (3.4)	0 (0)	1 (11.1)	5 (3.9)
Orthopedic surgery	153 (86)	36 (90)	8 (88.9)	109 (84.5)
Plastic surgery	19 (10.7)	4 (10)	0 (0)	15 (11.6)
Percentage of hand surgery (n = 178)					.68
<50%	5 (2.8)	1 (2.5)	0 (0)	4 (3.1)
50%-80%	35 (19.7)	7 (17.5)	1 (11.1)	27 (20.9)
80%-90%	29 (16.3)	10 (25)	1 (11.1)	18 (14)
>90%	109 (61.2)	22 (55)	7 (77.8)	80 (62)
Type of practice (n =178)					.25
Academic	17 (9.6)	7 (17.5)	0 (0)	10 (7.8)
Employed/limited academic	20 (11.2)	1 (2.5)	1 (11.1)	18 (14)
Military	4 (2.2)	0 (0)	0 (0)	4 (3.1)
Private	130 (73)	31 (77.5)	8 (88.9)	91 (70.5)
Training (resident/fellow)	7 (3.9)	1 (2.5)	0 (0)	6 (4.7)
Region (n = 177)					.82
Atlantic (NY, NJ, PA, MD, DE, DC, WV)	32 (18.1)	6 (15)	1 (11.1)	25 (19.5)
Great Lakes (IL, IN, MI, OH, WI, MN)	31 (17.5)	5 (12.5)	1 (11.1)	25 (19.5)
New England (CT, MA, ME, NH, RI, VT)	10 (5.6)	1 (2.5)	1 (11.1)	8 (6.2)
Pacific (AK, CA, HI, OR, WA)	28 (15.8)	9 (22.5)	1 (11.1)	18 (14.1)
South (FL, GA, NC, SC, VA, AR, LA) Al, KY MS, TN)	45 (25.4)	11 (27.5)	3 (33.3)	31 (24.2)
Southwest (TX, OK, NM, NV, UT, CO, AZ)	21 (11.9)	4 (10)	1 (11.1)	16 (12.5)
West (IA, KS, MO, ND, NE, SD, ID, MT, WY)	10 (5.6)	4 (10)	1 (11.1)	5 (3.9)

Note. The numbers in parentheses are the percentage distribution for each subgroup.

Table 2.

Surgeons in the Sometime User Group (Group 3, Total n = 129) Who Chose to Use Antibiotics in Each Case Scenario.

Question (case factor)	No. from group 3 (%)
Question 8 (<30 min)	0 (0.0)
Question 9 (>60 min)	64 (50)
Question 10 (<30 min + comorbidity)	57 (44)
Question 11 (>60 min + comorbidity)	101 (78)
Question 12 (<30 min + pinning)	102 (79)
Question 13 (>60 min + pinning)	116 (90)
Question 14 (<30 min + pinning + comorbidity)	111 (86)
Question 15 (>60 min + pinning + comorbidity)	121 (94)
Question 16 (minor open fracture with or without pinning)	116 (90)

Table 3.

Comparative Analysis of Sometime Users (Group 3; n = 129): Percentage of Surgeons Who Chose to Use Antibiotics by Each Pertinent Case Variable.

	Adding an additional factor		% change	P value
	< 30 min	> 60 min
Comorbidity + nonpinning	57 (44%)	101 (78%)	34	<.001
Comorbidity + pinning	111 (86%)	121 (94%)	8	.004
Noncomorbidity + pinning	102 (79%)	116 (90%)	11	<.001
Average			18
	Noncomorbidity	Comorbidity
>60 min + pinning	116 (90%)	121 (94%)	4	.07
>60 min + nonpinning	64 (50%)	101 (78%)	29	<.001
<30 min + pinning	102 (79%)	111 (86%)	7	.01
Average			13
	Nonpinning	Pinning
<30 min + comorbidity	57 (44%)	111 (86%)	42	<.001
>60 min + comorbidity	101 (78%)	121 (94%)	16	<.001
>60 minutes + noncomorbidity	64 (50%)	116 (90%)	40	<.001
Average			33

Note. For any given 2 questions with 2 factors in common, this table compares proportions of subjects who said yes to antibiotic use in a scenario between 2 levels of the third factor. The McNemar test for comparison of marginal proportions of paired observations was utilized.

Bolded items represent significant findings with P < 0.05.

Table 4.

Additional Questions on Frequency, Evidence, Indications, and Complications From Antibiotic Prophylaxis.

Question	Total responses (% of all responders)	Group 1	Group 2	Group 3
Is your decision on antibiotic prophylaxis based on scientific evidence?
Yes	89 (50)	10 (26)	6 (67)	72 (56)
No	88 (50)	29 (74)	3 (33)	57 (44)
How often do you give antibiotics despite thinking that it will not affect outcomes?
Often (once per day)		17 (45)	2 (22)	19 (15)
Sometimes (once per week)		13 (34)	0	44 (34)
Rarely (once per month)		5 (13)	4 (44)	51 (40)
Never		3 (8)	3 (33)	15 (12)
Why did you ultimately give the antibiotic unnecessarily?
Antibiotic given without my consent	41 (27)	5 (14)	1 (17)	35 (31)
Belief that current evidence is insufficient	40 (26)	14 (39)	0 (0)	26 (23)
Concern about board collection period	1 (0.6)	0 (0)	0 (0)	1 (0.9)
Hospital policy	42 (27)	16 (44)	2 (33)	24 (21)
Medical-legal concern	91 (59)	27 (75)	2 (33)	62 (54)
Other surgeons practice in this manner	26 (17)	7 (19)	1 (17)	18 (16)
Unaware of recommended indications	3 (2)	0 (0)	0 (0)	3 (2.6)
Other	24 (16)	4 (11)	0 (0)	20 (18)
Unsure	5 (3)	0 (0)	0 (0)	5 (4)
Witnessed a complication from antibiotic	39 (23)	7 (19)	2 (22)	30 (24)

Results

Of the 921 ASSH listserv recipients, 178 (19%) participated and were subsequently divided into 3 groups. The groups included group 1 (40 respondents) who gave antibiotics in the case of short cases, healthy patients, without hardware; group 2 (9 respondents) who never gave antibiotics; and group 3 (129 respondents) who sometimes gave antibiotics. Level of experience, fellowship training, type of residency training, volume of hand surgery within one’s practice, type of practice, and geographic region did not correlate with a particular group of antibiotic use (Table 1).

Among surgeons in group 3, antibiotic use for the different scenarios ranged from never (cases <30 min, without comorbidity or pinning) to 94% (cases >60 minutes, with comorbidity and pinning) (Table 2). The impact of each factor (length of case, medical comorbidity, and pinning) was reviewed from all scenarios and compared for influence on frequency of antibiotic use. Although each of the 3 factors affected the likelihood of antibiotic use in group 3, pinning was the single most impactful factor. Every factor significantly increased antibiotic use in every scenario (P < .05) except for the addition of a medical comorbidity to a case greater than 60 minutes in which pins were used (94% vs 90%; P = .07) (Table 3).

Although 50% of all participants reported their decision on prophylaxis as evidence-based, the majority of group 1 respondents (74%) admitted that their practice was not evidence-based (P = .002). The most frequently reported reason for unnecessary prophylaxis was medical-legal concern (59%), which was most notable in group 1 (75%) (P = .03). Other common reasons for unnecessary prophylaxis included antibiotics given without surgeon’s consent (27%), hospital policy (27%), belief that the current evidence to not give antibiotics is insufficient (26%), and that other surgeons practice in this manner (17%). Other factors for use included a patient’s prior history of infection, belief that benefits outweigh the risks, patients with implants like a total joint arthroplasty, and patient request (Table 4).

Nearly one-quarter (23%) of responding surgeons have witnessed a complication that they attributed to antibiotic use in hand surgery (Table 2). These included Clostridium difficile infection (16 cases, 9%), diarrhea (9 cases, 5%), rash (8 cases, 4%), anaphylaxis (5 cases), “Redman” Syndrome (3 cases), superimposed infection (2 cases), and death (1 case).

Discussion

This survey study assessing the frequency and indications of preoperative antibiotic prophylaxis by hand surgeons has 3 key findings. First, 23% of respondents chose to give antibiotics in clean hand surgery cases of less than 30 minutes, healthy patients, that did not involve hardware (group 1). This group was significantly more likely to admit their practice as not evidence-based (74%) and to have their prophylaxis decisions precipitated by medical-legal concern (75%). Second, the majority of respondents (73%) chose to give antibiotics only during some of the scenarios, and the frequency of prophylaxis in this cohort (group 3) was increased by longer operative time, medical comorbidity, and pinning. Third, 23% of respondents reported witnessing a complication from antibiotic prophylaxis, including 9% who had patients develop a Clostridium difficile infection.

Of the 23% of hand surgeons who reported they would administer preoperative antibiotics in short, clean cases—healthy patients—without hardware, 79% reported doing so at least once per week. This rate is comparable to the 31% of patients who received unnecessary prophylaxis in a retrospective review from a large academic hand center.¹⁰ In that review of SSI following clean elective hand cases, administering antibiotics did not change the rate of SSI, even in the presence of medical comorbidity and longer operative time. Furthermore, in a prospective trial including 1340 patients undergoing a variety of hand surgery procedures (clean and dirty, elective and emergent, with and without hardware), patients were randomized either to preoperative intravenous cephalosporin or placebo.¹⁸ No difference was found in the infection rate between preoperative antibiotic use and placebo in any scenario. The authors concluded that their findings do not support prophylaxis except in particular conditions under which a specific patient may have a high baseline infection risk. A well-conducted meta-analysis¹¹ and a large database review from an American insurance repository³⁶ concluded that preoperative antibiotic prophylaxis does not lower the risk of SSI while controlling for patient and surgical risk factors. Despite these findings, one recent review found an increase in antibiotic use in hand surgery by 73% between 2009 and 2015.³⁵

Our survey suggests that medical-legal concerns are driving preoperative antibiotic decisions for those hand surgeons who are not following the current best evidence available. Pressure from medical litigation has changed the way hand surgeons practice,^37,38 and considering that SSI is a common cause of litigation,³⁹ medical-legal concern may be seen as a reasonable cause for surgeons to give antibiotics. Although it is understandable for surgeons to tailor their practice in an attempt to reduce the likelihood of litigation, it is relieving to note that claims against hand surgeons in general are rare and most often unsuccessful. In a review of 23 legal cases relating to hand trauma from a single institution in New York City over an 8-year period, the authors found that only 1 patient won a settlement.⁴⁰ Specific guidelines from professional societies on the standard of care for antibiotic use in hand surgery may help clarify the current evidence against routine prophylaxis and lessen medical-legal concerns among surgeons.

Not all participants in this survey gave antibiotics readily, as most respondents (72%) gave antibiotics only during some of the 9 scenarios. Gaps in the evidence may precipitate a wide spectrum of practice regarding antibiotic prophylaxis in hand surgery. In the previously detailed prospective randomized controlled trial by Aydin et al,¹⁸ patients with comorbid medical conditions were excluded and it is unclear whether the study was powered appropriately to ascertain small differences in infection rates in the hardware group. In addition, the retrospective review of antibiotic use by Bykowski and colleagues excluded all procedures with hardware.¹⁰ These notable omissions limit our ability to delineate appropriate comprehensive evidence-based guidelines.

Hand surgeons are correct to scrutinize their practice of antibiotic use because routine prophylaxis involves a certain degree of risk. Twenty-three percent of the respondents in our study have witnessed a complication that they attributed to antibiotic use. In one retrospective analysis of general, cardiac, and vascular surgery in Quebec, the incidence of Clostridium difficile infection was 1.5% following perioperative antibiotic prophylaxis.²² For this reason, the authors recommended against prophylaxis in cases where infection is infrequent and benign. Recent prospective randomized controlled trials have also suggested that routine prophylaxis is not necessary in certain abdominal⁴¹ and urological cases.⁴² Reducing unnecessary prophylaxis in hand surgery would serve to reduce the development of complications, as well as subsequent superinfection and antibiotic resistance.

Our study has limitations. The methods of the survey collection are imperfect. The ASSH listserv represents less than a third of the members of the ASSH. There is a barrier to entry onto the listserv, as to receive the e-mails one must sign up on the ASSH website and be approved. Thus, it is possible that the listserv members are not representative of all ASSH members or of all surgeons who perform hand surgery. In addition, the survey response rate was low (19%), which further potentiates the sampling bias. It is possible that the most conscientious of hand surgeons subscribe to the listserv and the most dedicated of these will view and ultimately respond to the survey requests. Although our response rate is low compared with traditional methods,⁴³ it is comparable to other studies with similar methods. Even after sending up to 6 e-mails, the online response rate in other studies with analogous methods has ranged from 11% to 26%.^44-46 We chose the online survey for its ease of use, low cost, and efficiency, which may have limited our findings and the transferability of our results. Furthermore, the surgeon responses are subject to the Hawthorne effect and may not reflect actual choices made in patient care.⁴⁷ This is the effect of feeling pressured to choose the “correct” answer instead of choosing what one may actually do when presented with the same scenario in actual practice. For this reason, it is possible that antibiotic use was underestimated in the present survey. In addition, all respondents are subject to recall bias with regard to complications following prophylaxis.

Despite these limitations, the present analysis provides a snapshot of current antibiotic prophylaxis in hand surgery. Nearly a quarter of surgeons give prophylaxis when it is likely not indicated. The majority (74%) are aware of the lack of evidence to support this practice but make these choices based on medical-legal concern (75%). In addition, many report antibiotics being given without their consent (27%), state that their hospital policy dictates use (27%), believe that the current evidence is insufficient (26%), or simply do so because they believe other surgeons also use unnecessary prophylaxis (17%). To limit inappropriate antibiotic prophylaxis, the hand surgery community must widen the inclusion criteria on studies of antibiotic prophylaxis, improve the education of practicing surgeons and trainees regarding antibiotic prophylaxis, and look to develop clear and specific evidence-based guidelines that could be easily and universally adopted.