Influence of penicillin treatment of horses with strangles on seropositivity to Streptococcus equi ssp. equi‐specific antibodies

Abstract Background Antibiotic treatment of horses with strangles is reported to impair the development of immunity to subsequent exposure to Streptococcus equi ssp equi (S. equi). However, apart from a single clinical report, evidence‐based studies for this hypothesis are lacking. Hypothesis/Objective To determine whether penicillin treatment during clinical strangles influences the development or persistence of seropositivity to S. equi‐specific antibodies. Animals A natural outbreak of strangles with 100% morbidity in 41 unvaccinated mature Icelandic horses. Methods A prospective longitudinal study of acute clinical strangles from onset through full recovery approximately 10 months after the index case. Horses were monitored clinically 6 times for S. equi, as well as serologically for antibodies to antigens A and C of S. equi using an enhanced indirect ELISA. Seven horses received penicillin within 11 days of onset of fever (Group 1), 5 between 16 and 22 days after onset of fever (Group 2), and the remainder (Group 3, n = 29) received no antibiotics during clinical disease. The proportions of seropositive horses in each group were compared using an extension of Fisher's exact test with P < .05 as the level of significance. Results Although all horses were seropositive to S. equi within 2 months of the index case, significantly fewer horses treated early (Group 1) remained seropositive by 4 to 6 months (P = .04 and .02, respectively). Conclusions and Clinical Importance Findings support earlier suggestions that penicillin administered during acute strangles can interfere with persistence of humoral immunity to S. equi.

associated with fever, lethargy, and swollen abscessed lymph nodes over a period of weeks to months, 1,2 after which most horses recover uneventfully.
Penicillin is the drug of choice for treatment of streptococcal infections in horses. 3 However, there is controversy about whether to treat acutely infected horses with antibiotics or not, 2 because most cases of strangles resolve without specific treatment. 4 Nonetheless, recent reports suggests that, in some practice settings, the majority (78%) of horses with strangles receive some form of antibiotic treatment. 5 However, administration of antibiotics to horses with clinical strangles may simply delay the maturation of abscesses, or may increase the risk of developing metastatic or "bastard" strangles. 2 Furthermore, despite limited studies, 6 administration of penicillin to horses in early stages of strangles or as prophylaxis has been suggested to interfere with the development of protective immunity and thereby render such animals susceptible to reinfection. 2 When penicillin first became available to the medical profession, similar concerns about its impairment of the immune response arose.
Specifically, as related to treatment of streptococcal infections in people, 7 increased recurrence of infection was observed in human patients with scarlet fever (group A Streptococcus) treated with penicillin particularly early after the onset of clinical disease. Later work substantiated these observations, 8 with a suggested mechanism being that very early administration of penicillin interfered with the humoral immune response against Streptococcus. 9 Although humoral responses against S. equi correlate with protection, 10 to the best of our knowledge, no evidence-based studies document the effect of early penicillin treatment of strangles on the subsequent serologic response to S. equi.
A strangles outbreak with 100% morbidity occurred in a closed group of 41 mature Icelandic riding horses. Although antibiotic use was restricted during the clinical phase of the outbreak, because of the severity clinical signs or pressure from the owners, or both, 12 of the horses were prescribed courses of penicillin. It was hypothesized that penicillin treatment in clinical strangles would impair the development or long-term persistence of antibodies specific to S. equi.

| Horses included in the study
A strangles outbreak with 100% morbidity occurred in late April 2015 in a closed group of 43 adult Icelandic horses. Two horses were euthanized at the onset of the outbreak because of the development of clinical signs of strangles in combination with advanced age and poor dentition. The remaining 41 horses (mean age, 16.0 ± SD 6.6 years), including 32 geldings and 9 mares, were followed prospectively clinically, serologically and by upper respiratory microbiology from acute illness through clinical recovery as part of a larger study on the development of silent carriers. 11 All horses were S. equi positive by culture, qPCR positive, or both during the acute phase (first 7 weeks after the index case) and all but 1 horse had an episode of fever during that time. Because all horses were clinically affected during this outbreak, no isolation of clinically diseased horses was carried out and the same stable personnel handled all horses in the yard.
When outdoors, horses were held in smaller groups based on sex, but sharing of water buckets was possible. In the stable, all horses had individual boxes with low walls that allowed nose-to-nose contact with horses from other paddock groups. Three horses were lost for final sampling 10 months after the index case (see below, clinical sampling), with 2 horses euthanized; 1 because of laminitis and the other because of peritonitis. The remaining horse was moved to another farm after being determined to be S. equi negative on nasopharyngeal lavage (NL).

| Clinical sampling
Data included here are from 5 herd visits, hereafter denoted serum sampling days post index case (T = 0) as T27, T46, T123, T193, and T313. Samplings T27 and T46 took place during acute clinical illness, approximately 4 and 7 weeks after the index case, and samplings T123-T313 were conducted after clinical signs had resolved (approximately 4, 6, and 10 months after the index case; Figure 1 Animal care approval for the study was obtained from the regional animal ethics committee.

| Statistics
Descriptive statistics were calculated and comparisons between groups conducted using the Freeman-Halton extension of the 2-tailed Fisher exact probability test or the Wilcoxon ranked sign test. The level of significance was set at P < .05.  Notes: Groups 1 versus 2 versus 3: a P B = .75, b P B = 1.0, c P B = .04, d P B = .02, e P B = .06. Freeman-Halton extension of the Fisher exact probability test where P B is the probability that the null hypothesis holds.

| RESULTS
proportion of animals that remained seropositive to either Ag A or Ag C (Tables 3-5). Of note, at T313, 14/38 were determined to be clinically silent carriers, but the proportion of S. equi positives between these groups did not differ (P = .99; Table 2).

| DISCUSSION
These results provide supportive evidence that antibiotic treatment during acute clinical strangles can impair persistence of humoral immunity against S. equi. Based on principles of immunology, these findings were not unexpected. However, even the delay of penicillin administration on average of >7 days after the onset of fever still influenced the duration of the serologic response. Nonetheless, the findings strengthen the evidence for strangles treatment impairing humoral immunity whereas previous suggestions relied solely on a single case report. 6 (Table 2 and Figure 1). These findings, in combination with seropositivity to either Ag A or Ag C by the first 7 weeks of the outbreak (Table 5), indicate an active infection by S. equi in all horses during this time.
Strikingly, by approximately 4 months after the index case (T123), fewer horses treated with penicillin earlier (Group 1) remained seropositive (Table 5), suggesting a more rapid waning of humoral immunity in these animals. Similarly, studies in people have shown that early penicillin treatment of streptococcal infections may interfere with the development of humoral immunity 8,9 or increase the risk of recurrence of clinical disease at a later date. 7 A key difference between these studies in humans and the strangles outbreak described here was administration of penicillin within a few days of infection in the human patients versus a mean of 7.7 days in the horses. Thus, our data suggest that immediate 2 or prophylactic 6 antibiotic treatment of strangles, although potentially curative, likely would markedly impair the development of humoral immunity to S. equi, leaving such treated animals susceptible to clinical disease from future challenges. 10 Without continued immune stimulation, antibody responses wane over time. Thus, it was unexpected to find that some animals, in particular those in Group 1, were once again seropositive by the final sampling (T 313; 10 months after the index case). At that time, more than 33% of the horses (14/38) were identified as long-term silent carriers. 16 However, silent carrier status was not associated with differences in seropositivity to S. equi among groups, because the proportion of longterm silent carriers was similar in all 3 groups (Table 2). Although the premises were closed to horse movement, for practical reasons resident horses were grouped together on the same paddock or summer pastures. Thus, nose-to-nose contact likely allowed interanimal transmission of S. equi, as suggested by sequencing studies. 16 Therefore, reinfection by silent carriers may have been key factor for those animals regaining seropositivity at the point of final sampling (Tables 3-5).

| CONCLUSION
Treatment of horses with penicillin during acute clinical strangles is associated with persistence of seropositivity to S. equi. Whether this limitation in the duration of antibodies against S. equi leads to increased susceptibility to recurrence of clinical strangles after re-exposure to an infective challenge of S. equi remains to be determined.

ACKNOWLEDGMENTS
All personnel and horse owners from the affected stable, Karin Nygren, Kamilla Ekström, and Cecilia Ekberg are thanked for their assistance during sample collection throughout the study.

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
This study was approved by the Swedish Ethical Committee on Animal Experiments (diary nr C 36/14). All horse owners provided informed consent for use of their animals in the study.