A comparison of three doses of omeprazole in the treatment of equine gastric ulcer syndrome: A blinded, randomised, dose response clinical trial

Equine Veterinary Journal ISSN 425-1644 DOI: 1.1111/evj.12287 A comparison of three doses of omeprazole in the treatment of equine gastric ulcer syndrome: A blinded, randomised, dose response clinical trial B. W. SYKES*, K. M. SYKES and G. D. HALLOWELL BW Sykes Consultancy, Upper Orara, New South Wales, Australia School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington, UK. *Correspondence email: bensykes21@hotmail.com; Received: 21.1.13; Accepted: 6.4.14 Summary Reasons for performing the study: A previous study demonstrated that a dose effect between 1.6 and 4. mg/kg bwt of omeprazole per os s.i.d. is present in the treatment of equine gastric ulceration. In the same study, healing of glandular ulceration appeared inferior to healing of squamous ulceration. However, several limitations were recognised in that study and further investigation is warranted. Objectives: To further investigate the presence of a dose relationship in the treatment of gastric ulceration under conditions that may favour omeprazole efficacy such as administration prior to exercise and after a brief fast, and potential differences between the response of squamous and glandular ulceration to omeprazole therapy. Study design: A blinded, randomised, dose response clinical trial. Methods: Sixty Thoroughbred racehorses with grade 2/4 squamous and/or glandular ulceration were identified by gastroscopy. Horses were randomly assigned to receive either 1., 2. or 4. mg/kg bwt of enteric coated omeprazole per os s.i.d. 1 4 h prior to exercise. Gastroscopy was repeated at approximately 28 days. Results: The lower doses studied (1. and 2. mg/kg bwt) were noninferior to the reference dose (4. mg/kg bwt) in the treatment of squamous ulceration. Healing was greater in squamous ulceration than glandular ulceration (86% vs. 14%; P<.1). Improvement in ulcer grade was more likely in squamous lesions than glandular lesions (96% vs. 34%; P<.1). Worsening of the glandular ulcer grade was observed in 36% of horses. Conclusions: The results of this study suggest that, under the conditions studied, where omeprazole is administered before exercise and following a brief fast, doses of omeprazole as low as 1 mg/kg bwt per os s.i.d. may be as effective as higher doses. The proportion of glandular ulceration that heals with 28 days of omeprazole therapy is less than that of squamous ulceration. Keywords: horse; proton-pump inhibitor; squamous; glandular; gastric; ulcer Introduction Omeprazole is widely used in the treatment of equine gastric ulcer syndrome (EGUS) [1 5]. A previous study found a significant dose response in clinical patients with a dose of bwt per os s.i.d. resulting in a higher proportion of healed ulcers than a 1.6 mg/kg bwt per os s.i.d. dose [6]. However, 2 key factors, namely timing and the potential effect of fasting, may have resulted in suboptimal efficacy of omeprazole in that study. While it is widely believed that once daily dosing of omeprazole results in 24 h of acid suppression [7], and this is supported by experimental studies [8 14], a study by Merritt et al. [15] suggested that the duration of acid suppression may be <12 h. In the previous dose response study, omeprazole was administered approximately 2 h before exercise [6]. Given that exercise is thought to be the period during which damage to the squamous mucosa is most likely to occur [16], the long duration between administration and exercise may have reduced the efficacy of omeprazole at the doses studied. Furthermore, fasting may improve the bioavailability of omeprazole by as much as 3% [11]. The current authors have observed in the local racehorse population that most of the horses consume their evening meal within 4 h of feeding and propose that gastric fill is likely to be minimal if omeprazole is administered prior to exercise, which may favour absorption of omeprazole. Taking into account these 2 factors, administration of omeprazole closer to the high-risk period associated with exercise may result in enhanced efficacy across a range of doses. While the term EGUS is widely used to describe equine gastric ulceration, an attempt has recently been made to clarify the use of the terminology [17]. The term EGUS is all encompassing for ulcerative diseases of the stomach and within that broad definition more specific disease conditions or syndromes exist. The best understood of these is equine squamous gastric ulcer syndrome (ESGUS) for which the prevalence, risk factors and treatment are well described [1 5,18 23]. Equine glandular gastric ulcer syndrome (EGGUS) is less well recognised but recent studies have demonstrated a high prevalence in certain populations [18,2,24,25]. While glandular ulceration may result in delayed gastric emptying and secondary squamous ulceration in individual animals, it has been reported that the prevalence of squamous and glandular ulceration within populations are unrelated to each other [18,2,25] suggesting that within the general population that the majority of squamous ulceration is primary in origin. The proposed pathophysiology of EGGUS differs from that of ESGUS [7]. Consistent with this, a previous study [6] suggested that the response of EGGUS to omeprazole therapy is inferior compared with the response of ESGUS and the effect observed warrants further investigation. The null hypotheses for this study are that there will be an equal healing response between ESGUS and EGGUS to omeprazole and that there will be a >2% difference in response to omeprazole therapy between the reference dose ( bwt) and the 2 lower doses (1 and bwt). Materials and methods Recruitment and examination Thoroughbred horses from 5 stables were examined during June 212. Prior to examination all horses were fed their normal evening feed, but any remaining food was removed 6 8 h prior to endoscopy. Water was not withheld and horses exercised normally on the morning of the examination at the trainers discretion. Horses were sedated with detomidine (1 2 μg Dozadine/kg bwt i.v.) a and examined using a 3 m flexible videoendoscope (Portascope) b. The squamous and glandular mucosa were scored by a single investigator (B.W.S.) using a 4 point scale as described by the EGUS Council [26]. Based on the results of the gastroscopic examination, horses that met the inclusion criteria were enrolled into the study. Inclusion criteria for the study were horses with grade 2 or greater ulceration of the squamous and/or glandular mucosa, in race training and expected to remain in work Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd 285

Comparison of 3 doses of omeprazole in EGUS treatment B. W. Sykes et al. for the next 4 weeks, not receiving any medical treatment for EGUS and otherwise considered to be free of other significant disease. Group allocation and blinding Once enrolled into the study, horses were stratified by trainer and randomly allocated to a dose group by pulling their names out of a hat. Horses were stratified by trainer to reduce the variability in the diet and management of the horses as these have been shown to be risk factors for ulcer development [24]. One investigator (K.M.S.) was responsible for randomisation while the remaining investigators remained blinded to the group allocation until scoring was completed and recorded. The trainers were not blinded to the treatment group. The study protocol allowed for randomisation to be broken in the event of an adverse event. Treatment protocols Horses were fed and exercised at the trainers discretion determined by their normal routine, housed in individual stalls, bedded on wood shavings and fed a diet typical of Australian racehorses [27]. All horses were fed twice daily, with the morning feed within 2 h of completing exercise, and the afternoon feed approximately 12 14 h before exercising. A commercially available, enteric coated omeprazole paste formulation (Gastrozol) a was administered per os 1 4 h prior to exercise each morning at a dose of 5, 1 or 2 mg. This is equivalent to 1, 2 and bwt, respectively, for a 5 kg horse. To comply with the local regulations for racing, where, in addition to the requirement for the observation of withdrawal times, the administration of all pharmaceuticals is prohibited on the day of racing, omeprazole was not administered on the day of racing. Follow-up endoscopy Repeat gastroscopy, as described above, was scheduled between Days 25 and 3. Some variation was allowed in the timing of the repeat gastroscopic examination to accommodate for the individual horses racing schedules. The squamous and glandular mucosa were scored and assessed separately for each horse. Where the starting grade was 2 for the particular mucosa, ulcer healing was defined as change to grade 1 [6]. Where a starting ulcer grade of 2 was present in the particular mucosa, horses were considered to have improved if the ulcer score for the region decreased by at least one grade. Where the starting score, for either the squamous or glandular mucosa, was 2 and subsequently changed to 1 the horse was considered to have both healed and improved for that mucosa [1]. Horses with a submaximal starting ulcer grade (<4) for the squamous or glandular mucosa were considered to have worsened if their ulcer grade for that particular mucosa increased by at least one grade. Data analysis Data for time to follow-up, age and weight were assessed using a Shapiro Wilk test and were normally distributed with the groups compared using analysis of variance (ANOVA). The rest of the data were not normally distributed. Wilcoxon paired test was used to assess changes in ulcer scores within groups over time. A Chi-squared test was used to evaluate differences in sex between the groups. Either a Chi-squared (if 8% of the groups have a frequency 5) or a Fisher s exact test (when <8% of the groups have a frequency of 5 or greater) was used to evaluate healing, improvement and worsening of ulcers between groups, the effect of sex on ulcer healing, improvement or worsening and differences between glandular and squamous ulcer responses for each dose and overall. A Kruskal Wallis test was used to evaluate any differences in the numbers of race starts between treatment groups and any differences in the change in ulcer score over time between groups. Noninferiority was assessed using the upper limit of the 9% confidence interval of the differences in failure rate of the lower doses (1 and bwt) vs. the reference dose ( bwt). The confidence intervals to test for noninferiority were calculated using Wilson s method. Treatment effects were considered significantly different (i.e. noninferiority could not be demonstrated) if the upper limit of the 9% confidence interval (CI) exceeded 2%. An a priori noninferiority margin of 2% is commonly used when studies of this nature have not previously been published in the literature [28] and a recent, similar study evaluating placebo vs. misoprostol for the prevention of nonsteroidal anti-inflammatory drug (NSAID)-associated gastrointestinal injury in healthy human volunteers similarly set the a priori noninferiority margin at 17% [29]. Data are presented as mean ± s.d. and 95% CI where appropriate for continuous data and as median and interquartile ranges (IQR) for non-normally distributed data. Odds ratios (OR) and 95% CIs are displayed for binomial data. Commercially available statistical software packages were used (GraphPad 4. c, SPSS for Windows 15. d ). The 95% CIs are displayed using Jeffrey s intervals and calculated using online statistical software e. Noninferiority statistics were performed using online statistical software (Winpepi Compare 2 3.8) f. Significance was determined at P<.5. Results Horses Seventy-four horses were examined with 6 horses aged 2 7 years meeting the inclusion criteria. Three horses were lost to follow-up from both the 2 and bwt groups after being retired by the trainer for unrelated reasons and were excluded from data analysis. No difference in age was noted between groups (P =.4). Mean (± s.d.) weight was not significantly different between the groups (1 mg/kg bwt 52 ± 27 kg; bwt 512 ± 49 kg; bwt 58 ± 41 kg; P =.7). The distribution of female:male horses was not different between groups (1 mg/kg bwt 11:9; bwt 4:13; bwt 7:1; P =.2). There was no effect of sex on: ulcer healing (squamous male 87% [95% CI = 72 96], female 84% [95% CI = 64 95, P>.9]; glandular male 19% [95% CI = 7 39], female 7% [95% CI = 1 29, P =.6]); improvement (squamous male 94% [95% CI = 81 99], female 1% [95% CI = 88 1, P =.5]; glandular male 38% [95% CI = 2 59], female 29% [95% CI = 11 55, P =.7]); or glandular ulcer worsening (male 41% [95% CI = 24 59], female 29% [95% CI = 12 53, P =.5]). The horses had been in work for between 2 weeks and 12 months, although data were not specifically collected for analysis between groups, and, with the exception of the 6 horses described above, all horses remained in work without significant alteration in their standard routine. No adverse events were noted and blinding was maintained throughout the study. Gastroscopy The entire squamous mucosa was adequately observed in all examinations. Due to incomplete emptying of the stomach, the glandular mucosa was inadequately observed at enrolment in 5 horses and at follow-up in 3, 3 and 2 horses in the 1, 2 and bwt groups, respectively. In the remaining horses, residual fluid in the stomach obscured observation of the most ventral portion of the glandular body; however, the majority of the glandular body and entire pyloric antrum was visible. Where only the squamous mucosa was observed at enrolment, the horse was still considered eligible for inclusion with only the squamous mucosal score considered for statistical analysis. For horses where the glandular mucosa was not adequately observed at follow-up, the pre- and post treatment glandular mucosal scores were excluded from data analysis while the squamous scores were retained. Allowing for horses lost to follow-up and failure to observe the glandular mucosa on all occasions, the squamous mucosa was observed on both occasions in 2, 17 and 17 horses in the 1, 2 and bwt groups, respectively. The glandular mucosa was adequately observed on both occasions in 17, 14 and 15 horses in the 1, 2 and bwt groups, respectively. Days to follow-up and number of race starts Time to follow-up examination, 27.5 ± 1.5, 27.8 ± 1.4 and 27.9 ± 1.6 days in the 1, 2 and bwt groups, respectively, was not different between the groups (P =.7). The number of race starts for the 1 mg/kg bwt (; IQR: 2), bwt (1; IQR: 3) and bwt (1; IQR: 2) groups were not different between groups (P =.5). 286 Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd

B. W. Sykes et al. Comparison of 3 doses of omeprazole in EGUS treatment Number of horses 1 9 8 7 6 5 4 3 2 1 2. mg/kg 4. mg/kg 2. mg/kg 4. mg/kg Grade 4 3 2 1 Ulcer scores Squamous mucosa Glandular mucosa Grade /1 Grade 2 Grade 3 Grade 4 Fig 1: Distribution of equine squamous gastric ulcer syndrome and equine glandular gastric ulcer syndrome scores between the groups at enrolment. The overall prevalence of grade 2 or greater ulceration was 8% for the squamous mucosa and, in the horses where the glandular mucosa was adequately observed at enrolment, 65% for the glandular mucosa. The distribution of pretreatment squamous and glandular ulcer scores of horses included in the final data analysis are displayed in Figure 1. Squamous ulcer scores pre- and post treatment are displayed in Figure 2. No difference between the doses was observed in the change in squamous ulcer score (P =.8). Glandular ulcer scores pre- and post treatment are displayed in Figure 3. No difference between the doses was observed in the change in glandular ulcer score (P>.9). The improvement and healing response rates of the squamous and glandular ulceration are shown in Table 1. In the squamous mucosa no differences between doses were observed for either improvement (4 vs. bwt; OR =.18 [95% CI =.7 3.97] and 4 vs. 1 mg/kg bwt; OR =.16 [95% CI =.6 3.53]; P =.3) or healing (4 vs. bwt; OR =.38 [95% CI =.6 2.4] and 4 vs. 1 mg/kg bwt; OR =.2 [95% CI =.2 2.3]; P =.4). Similarly, in the glandular mucosa no differences between doses were observed for improvement (4 vs. bwt; OR =.45 [95% CI =.8 2.67] and 4 vs. 1 mg/kg bwt; OR =.84 [95% CI =.14 4.98]; P =.9), healing (4 vs. bwt; OR = 1. [95% CI =.5 18.32) and 4 vs. 1 mg/kg bwt; OR =.33 [95% CI =.3 3.78]; P =.5) or worsening (4 vs. bwt; OR =.53 [95% CI =.11 2.57] and 4 vs. 1 mg/kg bwt; OR =.67 [95% Grade 4 3 2 Fig 3: Box and whisker plot demonstrating median (solid line), interquartile ranges (represented by outer edges of box) and ranges (error bars) before and after treatment for equine glandular gastric ulcer syndrome grades when omeprazole was administered at 1, 2 and bwt per os s.i.d. No difference between the doses was observed (P>.9). CI =.14 3.11]; P =.4). The success and failure rates for healing and improvement are shown in Figures 4 and 5, respectively. Comparisons between the failure rates of the lower doses (1 and bwt) and the reference dose ( bwt), and the 9% CIs used for the determination of noninferiority are shown in Table 2. Against expectations, the lower doses were more effective than the higher dose; this was more consistent for squamous healing and improvement when compared with glandular healing and improvement (Figs 4, 5). Examination of the confidence interval around the point estimate of the difference in failure rate between 2 treatments provides an indication of whether the novel treatment (lower doses 1or2mg/kg bwt) are noninferior to the reference treatment (higher doses bwt). The a priori difference in failure rate that would be interpreted as clinically significant was set at 2%. As such, if the upper limit of the 9% CI for the difference in failure rate was less than 2% the novel treatment (lowerdose 1or2mg/kg bwt) would be considered noninferior to the reference treatment (high dose bwt). For example, in comparing the results achieved with the 1 and bwt doses with respect to squamous healing; there were fewer failures in the novel (1 mg/kg bwt) treatment when compared with the reference treatment ( bwt) and the point estimate of the difference in failure rate was -19%. The 9% CI of the difference in failure percentage was -4 to 3%. As the upper limit of the 9% CI was less than the a priori of 2% this suggests that the lower dose (1 mg/kg bwt) was noninferior to the reference dose ( bwt) for the healing of squamous ulceration. Noninferiority was similarly demonstrated for the bwt dose with regards to squamous healing and for both the 1 and bwt doses with regards to squamous improvement. Noninferiority was also demonstrated for the 1 and bwt doses with regards to glandular healing and improvement, respectively. Noninferiority could not be demonstrated for the 1 dose and bwt doses with regards to glandular improvement and healing respectively. 1 Fig 2: Box and whisker plot demonstrating median (solid line), interquartile ranges (represented by outer edges of box) and ranges (error bars) before and after treatment for equine squamous gastric ulcer syndrome grades when omeprazole was administered at 1, 2 and bwt per os s.i.d. No difference between the doses was observed (P =.8). Discussion When compared to the reference dose of bwt, noninferiority for the treatment of ESGUS was demonstrated for doses of 1 and bwt. The results for EGGUS were less consistent within noninferiority demonstrated in only 2 of the 4 parameters assessed. However, in the remaining parameters of EGGUS healing, where noninferiority could not be demonstrated, the lower doses had an equivalent or lower failure rate and the one-sided upper limit of the 9% CI approached the 2% a priori margin that was defined as significant. These findings suggest that, under specific conditions, doses as low as 1 mg/kg bwt may be as efficacious as a dose of bwt in the treatment of ESGUS. Further, they suggest that a similar Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd 287

Comparison of 3 doses of omeprazole in EGUS treatment B. W. Sykes et al. TABLE 1: The treatment response of squamous and glandular ulceration to 3 doses of omeprazole Squamous Glandular Response 95% Confidence interval Response 95% Confidence interval Squamous vs. glandular P value Healing (horses with an ulcer score grade 2 at enrolment) Improvement (horses with an ulcer score grade 2 at enrolment) Worsening (horses with an ulcer score < grade 4 at enrolment) Overall 86% (43/5) 74 94% 14% (5/35) 6 29% <.1 bwt 89% (16/18) 69 98% 23% (3/13) 7 5%.5 2. mg/kg bwt 94% (15/16) 74 99% 9% (1/11) 1 35% <.1 4. mg/kg bwt 75% (12/16) 51 91% 9% (1/11) 1 35%.1 Difference between doses Pvalue.28.52 Overall 96% (48/5) 88 99% 34% (12/35) 2 51% <.1 bwt 1% (18/18) 87 1% 31% (4/13) 11 58% <.1 2. mg/kg bwt 1% (16/16) 86 1% 45% (5/11) 2 73%.2 4. mg/kg bwt 88% (14/16) 66 97% 27% (3/11) 8 57%.3 Difference between doses Pvalue.11.63 Overall n.a. 36% (16/44) 23 51% n.a. bwt n.a. 38% (6/16) 17 62% n.a. 2. mg/kg bwt n.a. 43% (6/14) 2 68% n.a. 4. mg/kg bwt n.a. 29% (4/14) 11 55% n.a. Difference between doses Pvalue n.a..73 % 1 8 6 4 2 1 mg/kg Squamous 1 mg/kg Glandular Success Failure Fig 4: Success and failure rates for healing of equine squamous gastric ulcer syndrome and equine glandular gastric ulcer syndrome when omeprazole was administered at 1, 2 and bwt per os s.i.d. No difference between the doses was observed for healing of either equine squamous gastric ulcer syndrome (P =.28) or equine glandular gastric ulcer syndrome (P =.52). % 1 8 6 4 2 1 mg/kg Squamous 1 mg/kg Glandular Success Failure Fig 5: Success and failure rates for improvement of equine squamous gastric ulcer syndrome and equine glandular gastric ulcer syndrome grade when omeprazole was administered at 1, 2 and bwt per os s.i.d. No difference between the doses was observed for improvement of ulcer grade for either equine squamous gastric ulcer syndrome (P =.11) or equine glandular gastric ulcer syndrome (P =.63). effect may be observed in the treatment of EGGUS. Thus, the current findings suggest that the previously observed dose response [6] may not persist under conditions that favour omeprazole efficacy, such as administration prior to exercise and after a brief fast. Given the small number of animals studied and the wide 95% CIs for the data, indicating that the current results are imprecise, further work investigating the use of lower doses in a larger population of horses is warranted. The results clearly demonstrate significant differences between the healing response of ESGUS and EGGUS to omeprazole therapy. Following oral administration of omeprazole maximal serum concentrations occur between 45 and 9 min [11] with effective acid suppression achieved within 1 2 h [15]. The timing of administration in this study, 1 4 h before exercise was intended such that acid suppression would occur immediately prior to high-intensity exercise during which the majority of damage to the squamous mucosa is thought to occur [16]. In man, the inhibition of acid production for at least 2 h each day is recognised as a prerequisite factor for ulcer healing [3]. However, the duration of acid suppression required for ulcer healing in the horse has not been determined. High healing rates have been reported at a dose of bwt per os s.i.d. [1 6], yet in a recent study gastric ph remained >4. for <12 h at that dose [15], suggesting that the duration of acid suppression required for ESGUS healing in the horse differs from that required for healing of gastric ulceration in man. Another potential explanation for the efficacy of the low doses is the effect of fasting on bioavailability. While the horses in this study were not specifically fasted, as a population they consumed a high grain/low roughage diet typical of Australian racehorses [27]. The horses in this study were typically fed 12 14 h before morning exercise and the authors have observed that the majority consume their meals within 4 h resulting in an effective fasting period of around 8 1 h each day prior to administration of omeprazole the following morning. It is possible that this may have resulted in greater bioavailability of omeprazole. However, this was not specifically investigated in this study and warrants further investigation before any definitive conclusions can be drawn. The last potentially important factor for the differences observed between this and the previous dose response study [6] is the formulation of omeprazole used. Most commercially available omeprazole products, including Gastrogard TM and the preparation used in the previous dose response study, utilise a highly alkaline medium in which the omeprazole is suspended to buffer against gastric acidity [15]. However, the formulation used in the current study utilises enteric coated granules to protect the omeprazole. Direct comparison of the preparations is difficult because no published pharmacokinetic studies, or clinical trials comparing different products exist and caution should be exercised in 288 Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd

B. W. Sykes et al. Comparison of 3 doses of omeprazole in EGUS treatment TABLE 2: Noninferiority analysis of treatment failures when the lower doses (1 and bwt) were compared with the reference dose ( bwt). The lower doses treatment failure were noninferior to the reference dose treatment failure rate as the upper 9% CI (confidence interval) of the percentage difference was less than the 2% a priori margin Failure rates 1 mg/kg bwt bwt bwt Difference in failure (%) Novel treatment failure % minus control ( bwt) treatment failure % Upper and lower 9% confidence interval Squamous healing 6% (1/16) 25% (4/16) -19% -4 to 3% 11% (2/18) 25% (4/16) -14% -36 to 8% Squamous improvement % (/16) 12% (2/16) -12% -32 to 4% % (/18) 12% (2/16) -12% -32 to 3% Glandular healing 91% (1/11) 91% (1/11) % -24 to 24% 77% (1/13) 91% (1/11) -14% -38 to 13% Glandular improvement 55% (6/11) 73% (8/11) -18% -46 to 15% 69% (9/13) 73% (8/11) -4% -31 to 26% applying the results of this study to different formulations of omeprazole. Further work comparing the bioavailability and efficacy of different formulations is justified. The results of this study clearly demonstrate that the healing response of EGGUS to acid suppression therapy with omeprazole is less than that of ESGUS. In man, similar durations of acid suppression are required for healing of the reflux oesophagitis (analogous to squamous ulceration) and ulcerative gastritis (analogous to glandular ulceration) [3], but whether this is the case in the horse is not known. Accurate determination of the duration of acid suppression achieved following omeprazole administration under clinical conditions is required before specific conclusions relating to the duration of acid suppression required for healing of squamous and glandular ulceration in the horse can be made. An alternative explanation is that bacteria may play a role in the development or perpetuation of glandular ulceration and that acid suppression therapy alone may be inadequate. To date there remains significant conflict in the literature as to the role of bacteria in EGGUS. A recent study demonstrated that both gastric-adapted bacteria and opportunistic pathogens may play a role in squamous ulceration [31] but whether the situation is similar in the glandular mucosa is unknown. Helicobacter-like organisms have been identified in horses affected with EGGUS in some studies [32 34] whereas other studies have failed to identify such organisms [35,36] and to date no definitive role for bacteria in the development or perpetuation of glandular ulceration has been established. Antimicrobial therapy is commonly recommended as an adjunct to acid suppression therapy in refractory EGGUS [7,37] and, given the poor response of EGGUS to omeprazole therapy alone, studies evaluating whether antimicrobials improve the response rate are warranted. The potential for NSAIDs to cause EGGUS under clinical conditions is controversial and whether their use may have contributed to the worsening of EGGUS observed in the present study warrants consideration. An ulcerogenic capacity has been demonstrated for flunixin, phenylbutazone and ketoprofen at doses only 5% higher than typically recommended [38] while, at clinical recommended doses phenylbutazone and suxibuzone did not induce gastric ulceration when administered for 15 days [39]. Data regarding the use of NSAIDs was not recorded during the present study and this is a weakness of the study. However, the administration of NSAIDs was not identified as a risk factor for EGGUS in a recent study in Thoroughbred racehorses [24] and the authors have observed that few, if any horses, in the treatment population regularly receive such medications. Regardless, data specifically documenting the use of NSAIDs should be included in future study protocols investigating the therapeutic response of EGGUS to investigate any potential ulcer-generative effect. Data regarding time in work were also not recorded during the present study and is another potential weakness. While time in work has been identified as a risk factor for ESGUS in one study [24] it was, in contrast, not recognised as a risk factor for EGGUS in the same study. Further, time in work has not been demonstrated to be a factor in ulcer healing and as such, the omission of these specific data is unlikely to have influence the outcome of the present study. In conclusion, the results of this study suggest that, under the conditions studied, where omeprazole is administered before exercise and following a brief fast, doses as low as 1 mg/kg bwt per os once daily of the formulation given may be as efficacious as higher doses. However, due to the wide CIs, further studies are needed before definitive conclusions can be drawn. The study clearly demonstrates that the response to omeprazole of EGGUS is inferior to ESGUS and a greater understanding of the efficacy of omeprazole under clinical conditions is needed. Lastly, given the poor response of EGGUS to monotherapy with omeprazole, the evaluation of adjunctive treatments such as antimicrobials is warranted. Authors declaration of interests No competing interests have been declared. Ethical animal research This study was performed in accordance with the New South Wales Department of Primary Industries guidelines for clinical studies and the New South Wales Research Act of 1985. Informed consent from the owner, or the trainer acting as an agent for the owner, was obtained at the time of enrolment to the study. Source of funding This study was funded by Axon Animal Health. Portascope.com provided a videoendoscope free of charge for the purposes of this study. Acknowledgements The authors would like to thank Axon Animal Health and Portascope.com for their generous support of the project, the trainers for their cooperation and Ashleigh Dand and Arja Pöntinen for their assistance in performing the examinations. Authorship B.W. Sykes contributed to the study design, data collection, study execution, data analysis and interpretation, and preparation and approval of the manuscript. K.M. Sykes contributed to data collection, study execution and preparation and approval of the manuscript. G.D. Hallowell contributed to study design, data analysis and interpretation and preparation and approval of the manuscript. Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd 289

Comparison of 3 doses of omeprazole in EGUS treatment B. W. Sykes et al. Manufacturers addresses a Axon Animal Health, Belrose, New South Wales, Australia. b Portascope.com, Bradenton, Florida, USA. c Graphpad Software, La Jolla, California, USA. d SPSS Inc, Chicago, Illinois, USA. e http://epitools.ausvet.com.au/content.php?page=ciproportion f http://www.brixtonhealth.com/pepi4windows.html References 1. Andrews, F.M., Sifferman, R.L., Bernard, W., Hughes, F.E., Holste, J.E., Daurio, C.P., Alva, R. and Cox, J.L. (1999) Efficacy of omeprazole paste in the treatment and prevention of gastric ulcers in horses. Equine Vet. J. 31, Suppl. 29, 81-86. 2. Doucet, M.Y., Vrins, A.A., Dionne, R., Alva, R. and Ericsson, G. (23) Efficacy of a paste formulation of omeprazole for the treatment of naturally occurring gastric ulcers in training Standardbred racehorses in Canada. Can. Vet. J. 44, 581-585. 3. Lester, G.D., Smith, R.L. and Robertson, I.D. (25) Effects of treatment with omeprazole or ranitidine on gastric squamous ulceration in racing Thoroughbreds. J. Am. Vet. Med. Ass. 227, 1636-1639. 4. MacAllister, C.G., Sifferman, R.L., McClure, S.R., White, G.W., Vatistas, N.J., Holste, J.E., Ericsson, G. and Cox, J.L. (1999) Effects of omeprazole paste on healing of spontaneous gastric ulcers in horses and foals: a field trial. Equine Vet. J. 31, Suppl. 29, 77-8. 5. Murray, M.J., Haven, M.L., Eichorn, E.S., Zhang, D., Eagleson, J. and Hickey, G.J. (1997) Effects of omeprazole on healing of naturally-occurring gastric ulcers in Thoroughbred racehorses. Equine Vet. J. 29, 425-429. 6. Sykes, B.W., Sykes, K.M. and Hallowell, G.D. (213) A comparison of two doses of omeprazole in the treatment of EGUS: a blinded, randomised, clinical trial. Equine Vet. J. 46, 416-421. 7. Hepburn, R. (211) Gastric ulceration in horses. In Pract. 33, 116-124. 8. Jenkins, C.C., Frazier, D.L., Blackford, J.T., Andrews, F.M., Mattsson, H., Olovsson, S.-G. and McCleod, M. (1992) Pharmacokinetics and antisecretory effects of intravenous omeprazole in horses. Equine Vet. J. 24, Suppl. 13, 84-88. 9. Jenkins, C.C., Blackford, J.T., Andrews, F., Frazier, D.L., Mattsson, H., Olovsson, S.-G. and Peterson, A. (1992) Duration of antisecretory effects of oral omeprazole in horses with chronic gastric cannulae. Equine Vet. J. 24, Suppl. 13, 89-92. 1. Andrews, F. and Jenkins, C.C. (1992) Effect of oral omeprazole on basal and pentagastrin-stimulated gastric secretion in young female horses. Equine Vet. J. 24, Suppl. 13, 8-83. 11. Daurio, C.P., Holste, J.E., Andrews, F.M., Merritt, A.M., Blackford, J.T., Dolz, F. and Thompson, D.R. (1999) Effect of omeprazole paste on gastric acid secretion in horses. Equine Vet. J. 31, 59-62. 12. Sandin, A., Andrews, F.M., Nadeau, J.A., Doherty, T.J. and Nilsson, G. (1999) Effects of intramuscular omeprazole on gastric acid secretion in horses over a twenty-four hour period. Equine Vet. J. 31, 5-53. 13. Andrews, F.M., Frank, N., Sommardahl, C.S., Buchanan, B.R., Elliott, S.B. and Allen, V.A. (26) Effects of intravenously administrated omeprazole on gastric juice ph and gastric ulcer scores in adult horses. J. Vet. Int. Med. 2, 122-126. 14. Haven, M.L., Dave, K., Burrow, J.A., Merritt, A.M., Harris, D., Zhang, D. and Hickey, G.J. (1999) Comparison of the antisecretory effects of omeprazole when administered intravenously, as acid-stable granules and as an oral paste in horses. Equine Vet. J. 31, 54-58. 15. Merritt, A.M., Sanchez, L.C., Burrow, J.A., Church, M. and Ludzia, S. (23) Effect of GastroGard and three compounded oral omeprazole preparations on 24 h intragastric ph in gastrically cannulated mature horses. Equine Vet. J. 35, 691-695. 16. Lorenzo-Figueras, M. and Merritt, A. (22) Effects of exercise on gastric volume and ph in the proximal portion of the stomach of horses. Am. J. Vet. Res. 63, 1481-1487. 17. Merritt, A. (29) Appeal for proper usage of the term EGUS : equine gastric ulcer syndrome. Equine Vet. J. 41, 616. 18. Begg, L.M. and O Sullivan, C.B. (23) The prevalence and distribution of gastric ulceration in 345 racehorses. Aust. Vet. J. 81, 199-21. 19. Dionne, R., Vrins, A. and Doucet, M. (23) Gastric ulcers in Standardbred racehorses: prevalence, lesion description, and risk factors. J. Vet. Int. Med. 17, 218-222. 2. Luthersson, N., Nielsen, K.H., Harris, P. and Parkin, T.D.H. (29) The prevalence and anatomical distribution of equine gastric ulceration syndrome (EGUS) in 21 horses in Denmark. Equine Vet. J. 41, 619-624. 21. Luthersson, N., Nielsen, K.H., Harris, P. and Parkin, T.D.H. (29) Risk factors associated with equine gastric ulceration syndrome (EGUS) in 21 horses in Denmark. Equine Vet. J. 41, 625-63. 22. McClure, S.R., White, G.W., Sifferman, R.L., Bernard, W., Doucet, M.Y., Vrins, A., Holste, J.E., Fleishman, C., Alva, R. and Cramer, L.G. (25) Efficacy of omeprazole paste for prevention of gastric ulcers in horses in race training. J. Vet. Int. Med. 226, 1681-1684. 23. White, G., McClure, S. and Sifferman, R. (27) Effects of short-term light to heavy exercise on gastric ulcer development in horses and efficacy of omeprazole paste in preventing gastric ulceration. J. Am. Vet. Med. Ass. 23, 168-1682. 24. Habershon-Butcher, J.L., Hallowell, G.D., Bowen, I.M. and Sykes, B.W. (212) Prevalence and risk factors for ulceration of the gastric glandular mucosa in Thoroughbred racehorses in training in the UK and Australia [Abstract]. J. Vet. Int. Med. 26, 731. 25. Murray, M.J., Nout, Y.S. and Ward, D.L. (21) Endoscopic findings of the gastric antrum and pylorus in horses: 162 cases (1996-2). J. Vet. Int. Med. 15, 41-46. 26. Andrews, F., Bernard, W., Byars, D., Cohen, N., Divers, T., MacAllister, C., McGladdery, A., Merritt, A., Murray, M., Orsini, J., Snyder, J. and Vatistas, N. (1999) Recommendations for the diagnosis and treatment of equine gastric ulcer syndrome (EGUS). Equine Vet. Educ. 11, 262-272. 27. Richards, N., Hinch, G.N. and Rowe, J.B. (26) The effect of current grain feeding practices on hindgut starch fermentation and acidosis in the Australian racing Thoroughbred. Aust. Vet. J. 84, 42-47. 28. Bermingham, E., del Castillo, J.R.E., Lainesse, C., Paloske, K. and Radecki, S. (212) Demonstrating bioequivalence using clinical endpoint studies. J. Vet. Pharmacol. Ther. 35, 31-37. 29. Lee, K.N., Lee, O.Y., Choi, M.G., Choi, S.R., Lee, D.H., Lee, Y.C., Kim, T.N., Choi, S.C., Rew, J.S. and Seol, S.Y. (211) Prevention of NSAID-associated gastroduodenal injury in healthy volunteers a randomized, double-blind, multicenter study comparing DA-961 with misoprostol. J. Korean Med. Sci. 26, 174-18. 3. Shin, J.M. and Sachs, G. (28) Pharmacology of proton pump inhibitors. Curr. Gastroenterol. Rep. 1, 528-534. 31. Al Jassim, R., McGowan, T., Andrews, F. and McGowan, C. (28) Gastric ulceration in horses: the role of bacteria and lactic acid; Australian Government: Rural Industries Research and Development Corporation, Kingston. pp 1-26. 32. Contreras, M., Morales, A., Garcia-Amado, M.A., De Vera, M., Bérmudez, V. and Gueneau, P. (27) Detection of Helicobacter-like DNA in the gastric mucosa of Thoroughbred horses. Lett. Appl. Microbiol. 45, 553-557. 33. Morales, A., Garcia, F. and Bermudez, V. (21) Detection of Helicobacter-like organisms in Thoroughbred horses from Venezuela. Braz. J. Vet. Path. 3, 52-55. 34. Fox, J. (22) The non-h. pylori helicobacters: their expanding role in gastrointestinal and systemic diseases. Gut 5, 273-283. 35. Husted, L., Jensen, T.K., Olsen, S.N. and Mølbak, L. (21) Examination of equine glandular stomach lesions for bacteria, including Helicobacter spp by fluorescence in situ hybridisation. BMC Microbiol. 1, 84. 36. Martineau, H., Thomson, H. and Taylor, D. (29) Pathology of gastritis and gastric ulceration in the horse. Part 1: range of lesions present in 21 mature individuals. Equine Vet. J. 41, 638-644. 37. Nadeau, J.A. and Andrews, F.M. (29) Equine gastric ulcer syndrome: the continuing conundrum. Equine Vet. J. 41, 611-615. 38. MacAllister, C.G., Morgan, S.J., Borne, A.T. and Pollet, R.A. (1993) Comparison of adverse effects of phenylbutazone, flunixin meglumine, and ketoprofen in horses. J. Am. Vet. Med. Ass. 22, 71-77. 39. Andrews, F., Reinemeyer, C.R. and Longhofer, S.L. (29) Effects of top-dress formulations of suxibuzone and phenylbutazone on development of gastric ulcers in horses. Vet. Ther. 1, 113-12. 29 Equine Veterinary Journal 47 (215) 285 29 214 EVJ Ltd