Abstract
Open fractures of the distal phalanx commonly present to the Accident and Emergency Department. Controversy surrounds the use of prophylactic antibiotics in treating this injury. A double-blind, prospective, randomized placebo-controlled study was undertaken comparing the use of prophylactic flucloxacillin to placebo in addition to meticulous wound toilet. One hundred and ninety-three adult patients with an open fracture of the distal phalanx were studied. Seven patients developed superficial infections, an overall infection rate of 4%. No patient developed osteitis or a deep wound infection. There were three cases of infection in the 98 patients (3%) in the antibiotic group and four cases of infection in the 95 patients (4%) in the placebo group. A difference of proportion test confirmed no significant difference. It is concluded that the addition of prophylactic flucloxacillin to thorough wound toilet and careful soft-tissue repair of open fracture of the distal phalanx confers no benefit.
INTRODUCTION
Hand injuries commonly present to an Accident and Emergency Department, with 20% of all patients attending having injuries to the hand (Clark et al., 1985). 34% to 68% of all phalangeal fractures are open (Barton, 1979) and between 15% and 30% of all hand fractures are to the distal phalanx (Harrison and Hilliard, 1999).
The use of prophylactic antibiotics in the treatment of open fractures of the distal phalanx is acknowledged as controversial by a number of authors (Chow et al., 1991; Harrison and Hilliard, 1999; Ipsen et al., 1987). Despite this, many practitioners err on the side of prescribing antibiotics when treating this injury. Only two studies (Sloan et al., 1987; Suprock et al., 1990) have specifically looked at the use of prophylactic antibiotics in relation to this injury and these reported different conclusions. Neither appears to be a properly controlled, blinded trial and Sloan et al.’s (1987) study was abandoned, quoting an infection rate of 30% in the non-antibiotic group. Suprock et al.’s (1990) study included open fractures of all finger bones and quoted an overall infection rate of 8.4%, with no difference in the infection rates of the antibiotic and non-antibiotic groups. He placed particular emphasis on the need for aggressive surgical toilet and débridement.
Patients with hand injuries are often treated within the Accident and Emergency Department by relatively inexperienced junior staff. The dangers of inadequate wound toilet and poor attention to good surgical principles of wound care have been used to justify the routine use of prophylactic antibiotics (Davis and Stothard, 1990; Harrison and Hilliard, 1999). The commonest infecting organisms in hand wounds are staphylococci, probably from the patients own skin at the time of injury (Dellinger et al., 1988; Eaton and Butsch, 1970; Lewis, 1985; Linscheid and Dobyns, 1975; Stevenson and Anderson, 1993). Most injuries resulting in an open fracture of the distal phalanx of the digit occur in the community and almost all those treated in our department are managed on an out-patient basis. The risk of infection with a resistant hospital-based organism is therefore low.
This prospective, randomized, double-blind, placebo-controlled trial investigates the use of prophylactic flucloxacillin in the treatment of open fractures of the distal phalanx of the digit in an adult population presenting acutely to an Accident and Emergency Department. The null hypothesis was that prophylactic administration of flucloxacillin in addition to meticulous wound toilet and soft-tissue repair of an open fracture of the distal phalanx makes no difference to the rate of infection.
PATIENTS AND METHODS
The study was carried out in the Accident and Emergency Department of Crosshouse District General Hospital, Kilmarnock, Ayrshire, between 22 December 1998 and 10 October 2000. The hospital serves a mixed urban and rural population of approximately 220,000 people in North Ayrshire. The annual attendance to the Accident and Emergency department is approximately 53,000 new and 7,000 review patients. Ethical approval was obtained from the local research and ethics committee and informed written consent was obtained from each patient entered into the study.
Pilot study and literature review
A short retrospective pilot study was carried out to establish the incidence, identify variations in treatment and record the outcome for this injury within the Accident and Emergency Department. A computer word search of Accident and Emergency X-ray reports identified 303 X-rays of the digits in a 3 month period. Of these, 20 were open fractures of the distal phalanx. The age range was 9 to 83 years and the sex ratio was four men to one woman. It was noted that fractures could be graded using a simple grading system comprising of minor tuft fracture, linear fracture through the body of the distal phalanx, highly comminuted fracture and amputation. No attempt was made to classify the injuries in terms of the degree of soft-tissue injury. Scrutiny of case notes identified no cases of infection in any of these patients. All wounds healed to give a satisfactory functional result. Six of 20 patients had been treated with prophylactic flucloxacillin.
Power calculation
The incidence of wound infection following open fracture of the distal phalanx of the digit is not accurately known. The literature suggests that the incidence may vary from as low as 2% (Chow et al., 1991) to 30% (Sloan et al., 1987). For the purpose of our study it was assumed that the average infection rate for patients with an open fracture of the distal phalanx not treated with antibiotic was 15% and the average infection rate for those treated with an antibiotic was 4%. Assuming these infection rates for the treated and non-treated groups at a 90% level of significance a difference in infection rate of greater than 10% (which is deemed to be clinically significant) will be detected with a sample size of greater than 87 patients with a power of 80%.
Data collection
All patients with an injury likely to cause an open fracture of the distal phalanx of a digit were identified prospectively at initial presentation to the Accident and Emergency Department and a standard proforma was completed to record demographic data, mechanism and site of injury. If a fracture was present this was categorized using the simple grading system devised in the pilot study. The presence or absence of the nail and the presence or absence of injury to the nailbed were noted. All proformas were entered onto the main study database. A daily search of the total Accident and Emergency and radiology patient databases was carried out to identify any potentially missed cases.
Inclusion and exclusion criteria
All patients for whom a proforma was completed were considered for inclusion in the study and an X-ray was taken of the injured digit. A fracture was defined as a breach in the cortex of the distal phalanx visible on X-ray or subsequently identified at the time of wound exploration. An open fracture was defined as a fracture of the distal phalanx with an overlying wound. Fractures associated with closed subungal haematomas were not included. However, fractures associated with subungal haematomas that were bleeding externally or were trephined as part of the treatment process were included in the study. All patients entered into the study gave informed written consent. Patients were excluded from the study if they refused to participate, if they were less than 16 years of age, if the wound was more than 12 hours old, if they had a history of diabetes or symptomatic peripheral vascular disease, if they were taking oral steroids, if the fracture was caused by a bite or if they were already coincidentally taking an antibiotic or had a known allergy to penicillin.
Intervention
A local anaesthetic digital or wrist block was performed using 4% Prilocaine administered via a 27 gauge dental needle. After exsanguination of the limb and application of a pneumatic arm tourniquet all open wounds were inspected and meticulously cleaned with normal saline. Removal of residual nail, loose bone fragments, amputation of devitalized tissue and terminalization were carried out if required. Soft tissues were carefully repaired using 5-O Polyglactin 910 (Vicryl Rapide® Ethicon Inc, New Jersey) for the nail bed and 4-O Polyamide 6 Monofilament (Ethilon® Ethicon Inc, New Jersey) for skin. All wounds were dressed with paraffin tulle gras, gauze and a crepe bandage and elevated in a high arm sling.
Randomization
Two hundred identical bottles of study medication were prepared in the hospital pharmacy department. The bottles were randomized into blocks of ten using a random number table and sequentially labelled. Each group of ten bottles consisted of five of placebo and five of flucloxacillin. Each bottle held 40 identical capsules of either flucloxacillin 250 mg or lactose placebo. The sealed code identifying each bottle was kept in the pharmacy department and not opened until completion of the study. Each patient entered in the study was given the next sequential bottle containing the study medication and instructed to take two capsules four times daily for five days.
Review
All patients were reviewed by a limited number of observers as part of the study protocol at 4 or 5 days following the injury, with further reviews at approximately 14 days and 8 weeks unless the wounds had healed completely and the patient was asymptomatic. Both the patient and the observer were blind to which limb of the study the patient had been entered. Additional review appointments were undertaken as dictated by the patient’s clinical condition.
On the first visit, the mechanism of injury and the clinical and radiological findings at presentation were verified. Compliance with study medication was verified by questioning the patient and counting capsules remaining within the bottle. Any reasons for non-compliance were recorded. On this and subsequent visits, the wound was cleaned with sterile saline and further paraffin tulle gras and gauze dressings were applied if necessary.
At each review the wound was inspected for signs of infection. Infection was defined using clinical parameters of erythema, pain, swelling, wound discharge, presence of pus or cellulitis. Any sign of local infection was treated by wound toilet and débridement and a swab was taken for bacteriological culture. If there was evidence of spreading cellulitis the patient was withdrawn from the study and was treated initially with oral flucloxacillin until antibiotic sensitivities were available, at which point the antibiotic was changed if required. Hand therapy was carried out by the department of occupational therapy for any signs of joint stiffness or loss of function. On later reviews an assessment was made of the range of movement, cosmetic appearance and degree of function of the injured digit. Any residual disability was recorded.
RESULTS
During the study period 104,069 patients presented to the Accident and Emergency Department. 1,345 of these patients were identified at initial presentation to have an injury potentially causing a fracture of the distal phalanx of a digit or digits. The recruitment process for these patients is listed in Fig 1. Twenty-one patients were inadvertently omitted during this recruitment process, the reasons are outlined in Table 1. Six patients were referred to the orthopaedic service, and did not differ in any way from the study group: they were referred only as a result of failure to follow study protocol and not for reasons of clinical complexity.
Two hundred patients with 209 finger injuries were identified for inclusion in the study. Seven were excluded from further analysis as they failed to meet the entry criteria or were lost to follow-up (Table 2). One hundred and ninety-three patients with 202 open fractures of the distal phalanx were analysed in the study. The age range was 16 to 88 years, and the male to female ratio, 4.7:1. The right:left hand dominance ratio was 9.2:1, and the right:left injured hand ratio was 1.2:1. The most commonly injured digit was the index finger. The mechanisms of injury are shown in Table 3 and the numbers of different fracture types are detailed in the totals column of Table 5.
Analysis
The 193 patients included in the study were analysed on an intention-to-treat basis. Analysis of individual injuries was completed using the number of injured fingers, whereas analysis comparing treatment and placebo groups was carried out using individual patients. A patient with more than one finger injury clearly cannot be entered into more than one limb of the study.
Randomization
Table 4 demonstrates that randomization of patients to the two groups was effective. This is confirmed by the χ2 test, which shows no significant difference (P>0.05) between the samples in terms of gender, hand injured or hand dominance. Table 5 demonstrates effective randomization between the two groups for individual finger injuries in terms of severity of fracture, presence or absence of the nail and presence or absence of injury to the nail bed. It is noted that when more than one finger was injured, double counting of the patient occurred in the test group. However, the table clearly shows the equality of the distribution between the two test groups, and as such, the results of this study are equally applicable across the range of injuries. Other factors examined included the age of the patient, the day of the week the injury was sustained and the digit injured. No other significant differences were identified except that patients were half as likely to injure their little finger as any other digit.
Follow-up rates
Follow-up for the 193 patients was 100% for the first review, reducing to 91% for the second and 67% for the third review attendances.
Compliance
The patients in the study were analysed on an intention-to-treat basis and as such compliance with study medication or placebo has to be assumed in the final analysis. Although it does not affect the results of the study it is interesting to note that compliance with flucloxacillin was only 84% and with placebo 91%. The commonest reason given for failure to comply in both groups was minor gastro-intestinal upset. The breakdown of patient outcome in terms of treatment group, compliance and infection rate is shown in Fig 2.
Infection rate
Seven infections were identified in 193 patients, giving an overall infection rate of 4%. All infections were easily treated with local wound toilet and oral antibiotics; no patient who developed an infection was withdrawn from the study. Six of the wound infections cultured Staphylococcus aureus and the seventh grew E. coli, coliforms and enterococcus. No patient developed osteitis or a deep wound infection and none were systemically unwell. Those that developed infections were no different from the remainder of the study group in terms of type of fracture or mechanism or severity of injury.
A difference of proportion test was utilized to assess the difference in infection rates between the two groups. There were three cases of infection in the 98 patients (3%) in the antibiotic group and four cases of infection in the 95 patients (4%) in the placebo group. This is a non-significant difference (P>0.05).
With such small numbers of patients developing an infection the normality assumption that the difference of proportion test is based on could be questioned. However, a review of the data does corroborate the statistical result, that there is no significant difference between the two test groups.
DISCUSSION
The cornerstone of optimal wound management in an Accident and Emergency Department includes an understanding of the mechanism of injury and amount of energy transferred to the tissues, and an appreciation of the degree of wound contamination. Thorough and meticulous wound toilet using careful surgical technique with débridement of devitalized tissue and, where appropriate, primary wound closure are always required.
There is no place for routine administration of prophylactic antibiotics as part of the management of the majority of wounds presenting to an Accident and Emergency Department (Day, 1975; Hutton et al., 1978; Illingworth, 1973; Thirlby et al., 1983). This principle has also been shown to apply to soft-tissue wounds of the hand (Grossman et al., 1981; Haughey et al., 1981; Peacock et al., 1988; Roberts and Teddy, 1977; Worlock et al., 1980). A number of authors have raised concerns that indiscriminate use of antibiotics for simple wounds may result in infection with resistant organisms, poor patient compliance, allergic reactions, other unwanted systemic side effects and unnecessary drug costs (Grossman et al., 1981; Haughey et al., 1981; Roberts and Teddy, 1977; Wavak, 1981; Worlock et al., 1980). However controversy exists regarding the need for prophylactic antibiotics in the treatment of open fractures of the distal phalanx of the digits. Most authors acknowledge this controversy although in the absence of good clinical trials many continue to administer prophylactic antibiotics. This caution may be an extrapolation from the standard orthopaedic teaching on treatment of open fractures of the long bones where use of prophylactic antibiotics has been shown to reduce the incidence of wound infection and osteitis (Gustilo and Anderson, 1976; Patzakis et al., 1974). Of greater concern is the perception that prophylactic antibiotics are inappropriately used as a substitute for adequate wound toilet and proper application of the principles of good surgical wound care by inexperienced medical staff (Davis and Stothard, 1990; Harrison and Hilliard, 1999).
A number of risk factors for the development of wound infection following injury are known, including delay to presentation for treatment, bite wounds, (especially human bites), diabetes and peripheral vascular disease. In keeping with other studies looking at wound management (Thirlby et al., 1983) patients considered for entry into our study were specifically questioned regarding these risk factors and excluded if present. Patients under the age of 16 years were also excluded to overcome the ethical issues surrounding informed consent from children taking part in clinical trials.
Our study confirms that the infection rate following open fracture of the distal phalanx managed acutely in an Accident and Emergency Department is low provided that proper attention is paid to the correct principles of wound care. Addition of prophylactic flucloxacillin in the management of this injury is unnecessary and confers no benefit in terms of reducing the already low rate of infection. None of our patients developed osteitis or serious deep infection. We realize that poor management of open fracture of the distal phalanx of the digit, often by inadequately trained junior staff, can lead to significant morbidity and we stress the importance of training, supervision and review by experienced interested clinicians to minimize morbidity and ensure favourable outcome for our patients.
Footnotes
Acknowledgements
We thank Lynn Carleton, Senior Pharmacist, Crosshouse Hospital for help with preparation and randomisation of the study medication. JS had the idea for the study and has overseen the entire planning and execution of the study and of the preparation of the manuscript. He is the guarantor of the study. JS, GMcN and JR participated in the planning of the study. JS and GMcN participated in the execution of the study. JR prepared the statistical analysis. JS wrote the first draft of the paper and all authors have contributed to the final version.
Funding
The study was supported by a grant from the Research Network, Ayrshire and Arran Health Board.
Competing interests
None.