The microbiology of crocodile attacks in Far North Queensland: implications for empirical antimicrobial therapy

Wound infections are common after crocodile attacks and, therefore, prophylactic antimicrobial therapy is advised. However, there are limited data to guide recommendations for the optimal empirical regimen.

In a study from 1992,1 six of 11 survivors of crocodile attacks from Australia’s Northern Territory developed wound infection. The organisms isolated included Aeromonas hydrophila and Enterococcus species, Clostridium species, Pseudomonas aeruginosa and Proteus species, Staphylococcus epidermidis and Burkholderia pseudomallei. As a result, the authors advocated an empirical antibiotic regimen of ceftazidime, penicillin and metronidazole, with the addition of flucloxacillin to treat the patient’s skin flora.1 The Australian Therapeutic Guidelines do not discuss crocodile attacks specifically, but suggest giving oral amoxycillin–clavulanate to people with animal bite wounds at high risk of developing infection, and to patients with established mild infection. For more severe infections, intravenous piperacillin–tazobactam is recommended.2

To validate these recommendations, we reviewed the medical records of 14 of the 15 patients attacked by crocodiles who presented to Cairns Hospital in Queensland, Australia, after 1990 (one chart had been destroyed). Patients were aged 8–70 years and 13 were males. Wild saltwater crocodiles were responsible for seven attacks, farmed saltwater crocodiles for five and wild freshwater crocodiles for two. At presentation, nine patients had wound swabs collected; skin and soft tissue infection was already clinically apparent in four people. Organisms were isolated in six patients where swabs were collected (Box). Eleven of the 14 patients had surgery, with three requiring repeat debridement. Three further patients underwent delayed primary closure and two others required joint washouts. All patients received empirical antibiotics, but the selected agents varied enormously: ceftriaxone was the most commonly prescribed, but metronidazole, gentamicin, doxycycline, cephazolin, flucloxacillin and penicillin were also administered. No patients developed metastatic infection and all survived, although two of them lost digits.

Our findings highlight the diversity of organisms isolated from wounds caused by crocodile attacks. These bacteria can originate from the crocodile’s oral flora, the patient’s skin or can be acquired from the water or soil during the attack. The oral and cloacal flora of Australian crocodiles contain a myriad of organisms, including Aeromonas hydrophila, Pseudomonas aeruginosa, and Proteus and Salmonella species.3 While Aeromonas hydrophila — an organism found in fresh and brackish water — was notably absent in our study, it was the most common isolate in the NT series.1

The excellent outcomes seen in this study are probably primarily explained by prompt, effective surgical care,4,5 but antibiotics may have prevented infective complications. Although the regimen recommended in the NT series would have covered most of the Queensland isolates, it is relatively complex. Based on the isolates in the two series, an empirical regimen of oral amoxycillin–clavulanate for high risk wounds and mild infections would appear appropriate, reserving intravenous piperacillin–tazobactam for more severe infections. These treatments accord with the recommendations of the Australian Therapeutic Guidelines. It is essential to collect tissue cultures to facilitate de-escalation or modification of therapy if rare or resistant organisms, such as Burkholderia pseudomallei or Vibrio species, are isolated. In addition, tetanus should also be considered, with immunisation where appropriate.