Actinobacillus equuli ssp. haemolyticus in a semi-occlusively treated horse bite wound in a 2-year-old girl.

We report on the isolation of Actinobacillus equuli ssp. haemolyticus from wound smears of a 2-year-old girl who was admitted to the hospital due to partial amputation of the distal phalanx of her right middle finger caused by a horse bite. A. equuli typically causes diseases in horses and only very few reports describing human infections (mostly associated with wounds) are available in the literature. Interestingly, although the bacteria could be found in consecutive samples taken at different points in time, there were no signs of advancing infection or inflammation. Moreover, the fingertip regenerated after 74 days under semi-occlusive dressings with very pleasant results. For strain identification two automated systems were employed producing discrepant results: VITEK 2 described the pathogens as Pasteurella pneumotropica while MALDI-TOF MS analysis revealed A. equuli. Sequence analysis of 16S rDNA gene finally confirmed A. equuli ssp. haemolyticus as the isolated strain. The antimicrobial susceptibility testing was performed according to the CLSI criteria for Pasteurella spp. Additionally we conducted a test according to the EUCAST criteria.

Introduction

Animal bites are an important issue in health care management. According to Thomas and Brook in the USA approximately 800,000 people seek medical aid due to animal bites each year [1]. Although there are no reliable data available, a rate of 30,000 to 50,000 cases of animal bite wounds per year are estimated in Germany [2]. Infections with both, aerobic and anaerobic organisms can lead to edema, erythema, lymphangitis, sepsis, necrosis but could also be the cause of osteomyelitis [1], [2], [3]. In recent years zoonotic infections such as rabies, tularemia, brucellosis or tetanus have become a rare incidence in Germany [2], [3]. The relevant organisms are mostly part of the oral cavity and differ depending on the animal species. While most bite injuries are caused by dogs and cats, horse bites are relatively uncommon. Bite wounds are normally characterized by a mix of animal oral flora, the victims skin flora and environmental organisms [1], [4], [5], [6]. The most common pathogens found in bite wounds belong to the family Pasteurellaecae, Staphylococcus spp. and Streptococcus spp. [1], [4], [5], [6].

Actinobacillus spp. are Gram-negative bacteria that are commensals of the oropharyngeal cavity of horses and sheep and are closely related to Pasteurellaceae [7], [8]. A. equulii leads to systemic infections in horses such as the sleepy foal disease, a mostly fatal septicemia of newborn horses. In addition to that, meningitis, mastitis, arthritis and endocarditis could be associated with A. equuli [9].

Recent epidemiological investigations suggested the species A. equuli to be divided into the two subspecies Actinobacillus equuli ssp. equuli ssp. nov. and Actinobacillus equuli ssp. hemolyticus ssp. nov. [10]. Moreover, Actinobacillus equuli ssp. equuli ssp. nov. is CAMP-Test negative whereas Actinobacillus equuli ssp. haemolyticus is positive [10]. In contrast to Actinobacillus equuli ssp. equuli ssp. nov. Actinobacillus equuli ssp. haemolyticus seems to be exclusively associated with horses [10]. The few reports affecting humans deal with wound infections and septicemia [7], [8].

Here we report the case of a 2-year-old girl who was bitten by a horse and had to undergo surgical treatment. The wound fluid was analysed at defined time points and bacterial growth could be detected revealing A. equuli ssp. haemolyticus, Bacillus cereus and mixed skin flora. In this case report we focus on the diagnostic procedures applied for identification of the causative pathogen and furthermore propose options for antimicrobial susceptibility testing of A. equuli.

Case description

The 2-year-old, otherwise healthy, presumably immunocompetent and orderly vaccinated girl presented to our emergency department immediately (within around 35 minutes) after a horse had bitten off 30–40% of the distal phalanx of her right middle finger at the level of the lunula (Figure 1a (Fig. 1)), i.e. classified as Type III according to Allen [11]. No amputate was found at the site of accident. The amputation was treated with a semi-occlusive dressing [12] using a custom made silicone finger cap that forms a wet chamber around the wound while allowing for aseptic fine needle aspiration of wound fluid from a small reservoir that is connected to this chamber by a capillary (patent pending). A regular gauze dressing surrounded this finger cap for additional fixation. Since the actual horse bite was initially unknown and became clear only after questioning adult witnesses later on, the wound was cleaned and antiseptically irrigated with Octenisept® (Schülke & Mayr, Norderstedt, Germany) containing 0.1% octenidine dihydrochloride and 2% phenoxyethanol. An antibiotic prophylaxis was not administered. During the course of the treatment samples of the fluid were taken on a weekly basis while the fingercap was completely removed every two weeks to assess the regeneration progress. In total six samples were sent to the microbiology department for bacteriological analysis. Despite the infection with A. equuli ssp. haemolyticus, B. cereus and mixed skin flora which led to the formation of creamy yellow pus, we noted no advancing inflammation, no lymphangitis and no fever. Apart from an odor that increased with time and necessitated weekly changes of the outer gauze dressing, neither the girl nor her parents reported any unusual discomfort or pain and the hand was quickly used in all daily activities despite the injured finger and its dressing. After 74 days the fingertip had almost fully regenerated with complete scarless epithelialization, full function and very pleasant cosmetics with minimal shortening of the fingertip and with little clubbing of the nail (Figure 1a–e (Fig. 1)).

Figure 1

Photographs of the injured finger taken on days 2 (a), 28 (b), 56 (c), at the end of the treatment on day 74 (d) and finally at around 3 months after the injury (e)

After the smears were delivered to the laboratory a Gram staining was carried out showing both Gram-negative rods and cocci and Gram-positive rods. Bacterial colonies could be detected after 24 hours of incubation at 37°C and 5% CO2 on Columbia Blood Agar showing phenotypical similarities to Bacillus spp. In the subsequent analysis this strain could be identified as B. cereus (positive indole reaction and MALDI-TOF MS Score: 2.13). A second strain presented with small grey-white colonies showing a discreet β-hemolysis (Figure 2 (Fig. 2)). For identification two different automated systems were applied: first biochemical phenotyping by VITEK 2 (bioMérieux, Nürtingen, Germany) and second mass spectrometry MALDI-TOF MS (Bruker Daltonics, Bremen, Germany). First, VITEK 2 identified Pasteurella pneumotropica (identification score 96%) second MALDI-TOF MS led to A. equuli (MALDI-TOF MS Score: 2.149). To clarify the true nature of these bacteria a 16S rDNA sequencing analysis was conducted using a 3130 Genetic Analyzer (Life Technologies, Darmstadt, Germany). The subsequent comparison of the DNA sequence with the NCBI nucleotide collection (nr/nt) using BLAST algorithm (http://blast.ncbi.nlm.nih.gov) revealed 100% homology to the 16S rDNA sequence of A. equuli ssp. haemolyticus, thereby confirming the species proposed by mass spectrometry.

Figure 2

Growth of Actinobacillus equulii spp. haemolyticus on Columbia Blood Agar and Gram Staining. As demonstrated in a) the bacteria grow in small grey colonies with a sharp β-hemolysis while b) shows the typical Gram-staining showing the bacterial growth as both as gram-negative rods and cocci.

Since Actinobacillus spp. is a member of the family Pasteurella we decided to adapt the antimicrobial suspectibility testing according to the CLSI criteria for Pasteurella spp. The minimal inhibitory concentrations (MIC) were determined using E-Tests strips for the recommended antibiotics (bioMérieux, Nürtingen, Germany) [13], [14], [15], [16]. Since the European Committee for Antimicrobial Supectibility Testing (EUCAST) published guidelines for Pasteurella multocida (antibiotic panel tested by disk diffusion method on Mueller-Hinton agar with 5% sheep blood) in 2013 we decided to compare both, EUCAST and CLSI results. For evaluation, the EUCAST-breakpoints published for non-related species were applied [14]. However, since for non-species related breakpoints only MICs are available we additionally applied E-tests to detect the MICs. These results are demonstrated in Table 1 (Tab. 1) and Table 2 (Tab. 2).

Table 1

Antibiotics tested according to CLSI criteria. The antibiotics were tested according to the CLSI-guidelines published for Pasteurella spp. Ampicillin, Amoxicillin/Clavulanic Acid, Moxicillin, Trimethoprim/Sulfamethoxazole and Tetracycline were tested to be susceptible (S) and Erythromycin resistant (R).

Table 2

Antibiotics tested according to the EUCAST criteria. The proposed test panel for Pasteurella multocida was chosen. The gained MICs were interpretated according to the EUCAST non-species related breakpoints. Ampicillin, ampicillin/clavulanic acid, cefotaxim and ciprofloxacin were tested as susceptible (S). For Trimethoprim/Sulfamethoxazole and Tetracycline there is an insufficient evidence (IE) for being a good therapeutic target.

Discussion

In the reported case A. equuli ssp. haemolyticus and B. cereus were repeatedly detected in the wound fluid of a 2-year-old girl who underwent treatment after being bitten by a horse. Despite the positive bacterial analyses no signs of advancing infections or inflammation were detected. The overall outcome was beneficial and it was decided not to apply any antibiotics. Our observations might be seen in the context of a positive role (and the novel understanding) of the microbiome in wound healing [17]. In this respect processes such as bacteria-bacteria antagonism, a microbial influence on the initiation of re-epithelialisation, epidermal-cell proliferation and neo-vascularisation might play a significant role [17], [18]. To date there are no reports demonstrating a role of A. equuli ssp. haemolyticus in wound healing but nevertheless a potential role under certain conditions (e.g., occlusion) may not be denied. On the other hand, different (possibly undetected) microorganisms might have prevented the initiation of wound infection. In contrast, when detected in wounds, B. cereus is seen as an environmental contaminant [19]. For that reason, further investigation and research is needed to elucidcate the conditions and the extent, to which bacteria might have contributed to such a favourable outcome and inhibition of advancing infection in the case described here.

Regarding strain identification both mass spectrometry and 16S rDNA sequencing correctly identified the species Actinobacillus whereas the biochemical phenotyping by VITEK 2 resulted in Pasteurella pneumotropica. This discrepancy might be explained by the fact that only four biochemical tests discriminate between Actinobacillus spp. and Pasteurella spp.: β-galactosidase, urease activity, growth on McConkey agar and indole production [7]. Therefore, it may be assumed that the strain tested here might not have possessed all properties which would have been necessary for the correct biochemical identification. Moreover, MALDI-TOF MS has been shown to reliably identify A. equuli ssp. haemolyticus in a previous study [20], [21].

Due to the close relationship of Actinobacillus spp. and Pasteurella spp. we decided to adapt the antimicrobial testing according to the CLSI-criteria for Pasteurella spp. Our strain was susceptible to ampicillin, ampicillin/clavulanic acid, moxifloxacin, trimethoprim/sulfamethoxazole and tetracycline, but showed resistance towards erythromycin (Table 1 (Tab. 1)). This observation is in accordance with reports on the resistance profile of Pasteurella multocida which demonstrate ineffectiveness for erythromycin, while the other antibiotics mentioned here are effective therapeutical agents [22]. Since the European Committee for Antimicrobial Standardisation Testing (EUCAST) proposed an antibiotic panel for Pasteurella multocida in 2013 we additionally conducted an antimicrobial testing according to this panel. For evaluation we used the proposed MICs for non related species. However, since zone diameter breakpoints are only defined for Pasteurella multocida we performed an E-test for each antibiotic. The strain tested here was found to be susceptible to ampicillin, ampicillin/clavulanic acid, cefotaxime and ciprofloxacin. Insufficient evidence as an appropriate therapeutic target needs to be proposed for trimethoprim/sulfamethoxazole and tetracycline (Table 2 (Tab. 2)).

Conclusion

In conclusion we could demonstrate the isolation of A. equuli ssp. haemolyticus in wound fluids of a 2-year-old girl. Interestingly, despite of repeated isolation of bacteria that might cause wound infections no therapeutically significant infection was seen (e.g., advancing erythema, swelling, lymphangitis, fever, lymphadenopathy, etc.). In this respect further research is needed to understand potential positive properties of the wound associated microbiome and specific conditions of Actinobacillus equuli ssp. haemolyticus for wound healing. Furthermore, it could be shown that MADLI-TOF MS is a reliable tool for correct identification of this pathogen. Further studies with larger numbers of bacteria will have to clarify the best diagnostic approach. Finally, we propose that an appropriate antimicrobial susceptibility testing can be performed according to both, CLSI-guidelines for Pasteurella spp. or EUCAST-test criteria.

Notes

Competing interests

The authors declare that they have no competing interests.