Edinburgh Research Explorer Chemotherapy-induced diarrhoea in dogs and its management with smectite

7 Chemotherapy-induced diarrhea (CID) is a frequent chemotherapy adverse event in dogs. 8 Yet, there is currently no consensus regarding its management. Smectite is a natural medical clay, widely used in the treatment of acute diarrhea in humans. The objectives of this study were to assess the efficacy of smectite in the management of CID 11 in dogs, and to collect epidemiological data on CID. For each episode of diarrhea, dogs were randomized into 2 management groups: Smectite 13 group, receiving smectite at 0.5 g/kg PO per day divided in 2-3 doses initiated at the start of 14 CID; Control group, without initial medication. In both groups, rescue metronidazole was 15 prescribed if CID progressed or was not improved within 48 hours. 16 Sixty dogs were recruited and received 426 chemotherapy administrations between June 2017 17 and March 2019. The incidence rate of CID was 110/426 (25.8%, 95% CI: 21.7-30.2%), and 18 significantly differed between the chemotherapeutic drugs administered ( P < 0.001). 19 Metronidazole was administered in 5/54 events (9.3%, 95% CI: 3.1-20.3%) in the Smectite 20 group and in 40/56 events (71.4%, 95% CI: 57.5-82.3%) in the Control group ( P < 0.001). The time to resolution of diarrhea was shorter ( P < 0.001) in the Smectite group (median: 22 19.5 hours, interquartile range: 13.5-32 hours) compared to the Control group (median: 53 23 hours, interquartile range: 31.5-113.5 hours). The results of this study support the administration of smectite in the first-line management 25 of CID in dogs.


INTRODUCTION
Chemotherapy-induced diarrhea (CID) is a frequent chemotherapy adverse event in 29 dogs, 1 of which clinical severity can vary as described by the Veterinary Cooperative 30 Oncology Group criteria for adverse events (VCOG-CTCAE) version 1.1. 2 Yet, 31 epidemiological data is scarce since it is often reported as part of the gastrointestinal adverse 32 events, 3,4 and the collection of more accurate information may help improve its management. 33 In humans and rodent models, it was historically thought that CID arose solely from damage 34 induced by chemotherapeutic agents to the rapidly proliferating cells of the basal epithelium. 5

35
However, alteration of the mucus layer, increased intestinal permeability, mucosal 36 inflammation and changes in the intestinal microbiota seem to play a central role. [6][7][8] infectious agents (viral, bacterial or parasitic), and in particular Clostridioides difficile, may 38 also occasionally be involved in the pathogenesis of CID in humans. 9,10 To the best of the 39 author's knowledge, only one recent veterinary study investigated probiotics as a treatment of 40 CID, but failed to demonstrate a potential benefit. 11 In the absence of consensus for the 41 management of CID in dogs, it has been suggested that antibiotics such as metronidazole 42 could be used. 12

43
Although the mechanism of action of smectite is not clearly understood, it is an 44 activated natural aluminosilicate clay that can adsorb water and presents multiple other 45 interesting properties: it prevents toxins, bacteria and viruses adhering to intestinal 46 membranes, 9,[13][14][15][16][17] it has been shown to strengthen the mucosal barrier in vitro and in vivo, 18-20 47 to have some anti-inflammatory properties, 17,18,21,22 to decrease intestinal bacterial 48 translocation, 23 and to stabilise the intestinal microbiome. 16,17 49 Its role in the treatment of acute diarrhea in children and adults is well established, 24,25 50 and a beneficial effect has been found in different types of diarrhea and different 51 species. 21,22,[26][27][28] In a case series of 17 dogs with intractable CID, 10 dogs (58.8%) had 52 resolution of their diarrhea after 48-72 hours of treatment, and improvement was noted in the 53 remainder. 29

54
Chemotherapy dosage is typically limited by the occurrence of adverse events, and 55 efficient management of CID could decrease the need of dose reductions, and optimize 56 chemotherapy dose-intensity, whilst improving the quality of life of our veterinary patients.

57
The primary objective of this study was to assess the efficacy of smectite in the management 58 in dogs with CID, taking the time to resolution of diarrhea as a primary outcome. A 59 secondary objective was to collect epidemiological data on CID in dogs. (2) to have pre-existing 68 diarrhea whether suspected to be related to the tumor or not. Owners were given an 69 information sheet and were required to sign a consent form confirming that they agreed to 70 enrol their dog in the study. Dogs were enrolled by the clinician in charge of the case, and 71 conditions required for enrolment were verified by the first author.

73
Dogs were randomized directly after enrolment, with a 1:1 allocation ratio into one of two

84
The smectite used in this study was in the form of hydrated calcium aluminosilicate clay 85 with a particle size < 80 µm (VBS Rx Clay ® , VBS Direct Ltd., Bulkeley, United-Kingdom).

86
Pots contained 100g of smectite and were provided with a 500 mg measuring scoop.

87
Dogs in the Smectite group were provided with a pot of smectite. If they developed 88 diarrhea, owners were instructed to give their dog smectite at 0.5 g/kg PO per day divided in Control group did not receive any first-line intervention.

95
In both management groups, a rescue protocol was implemented to comply with the local 96 ethical regulation. In case of progression of the diarrhea (increasing Waltham ™ faeces 97 consistency grade) or no improvement within 48 hours, metronidazole was prescribed at 10-98 15 mg/kg PO every 12 hours for 5 days.

99
Owners were asked to continue feeding their dogs with their normal routine diet during 100 the full chemotherapy protocol; including before, during and after the development of

105
Owners were given a diary to complete in case of diarrhea, from the last normal stool to at 106 least two non-diarrheic stools (see Supplementary form 1). They were asked to record in chemotherapy administration. In particular, rectal temperature was measured at each hospital 118 visit, whether it was scheduled or not. A QOL questionnaire which had been validated by a 119 previous study was completed by owners at every visit as part of the routine practice in our 120 institution, independently from this study (see supplementary form 2). 30 Predicted nadir 121 absolute neutrophil count (ANC) were obtained from complete blood counts (CBC) 122 performed 7 days after administration for all drugs except carboplatin, for which they were 123 performed after 10 days. Additional CBCs were also performed when clinically indicated.

124
Febrile neutropenia was defined as chemotherapy-induced neutropenia (ANC < 2.5 x 10 3 /µL) 125 in conjunction with fever (rectal temperature > 39.2 °C). 31 All chemotherapy-associated 126 toxicities were recorded and graded according to the VCOG-CTCAE version 1.1. 2 Lethargy, 127 anorexia and vomiting were used to assess constitutional/gastrointestinal adverse events. If 128 diarrhea was thought not to be related to chemotherapy based on the history, then the event 129 was excluded from the study.

139
Sample size determination was based on the hypothesis that the duration of diarrhea would be 140 significantly shorter within the Smectite group compared to the Control group. The minimum 141 detectable difference of the primary outcome between the two groups was set at 24 hours, with 142 an alpha risk of 5% and a beta risk of 20%, the number of diarrhea events to be included per 143 group was calculated to be 63. Based on our institution database, a period of about 1.5 years 144 was expected to collect the number of diarrhea events. An early stopping rule was established 145 in case of reaching the scheduled closure date, because the first author (Q.F.) was no longer 146 available to ensure the continuity of the study after this date.

147
Fisher's exact test was used to compare binary data, and Mann-Whitney test was used to 148 compare ordinal and continuous non-normally distributed data. Log-rank and Wilcoxon tests 149 were used to compare the time to resolution of the diarrhea between the two groups. The

160
Seventy-four dogs were initially assessed for eligibility for the study. Fourteen dogs were 161 excluded, 6 for having gastrointestinal tumors, 5 for having pre-existing diarrhea, and 3 162 because the owner declined to enroll their dog (Figure 1). Sixty dogs were prospectively 163 enrolled in this study (Table 1), and 426 chemotherapy administrations were recorded (Table   164 2). Chemotherapy protocols included 19-week CHOP-based protocol, 32 escalating-dose 165 vinblastine/prednisolone protocol, 33 and other single-agent lomustine, doxorubicin and 166 carboplatin protocols. One hundred and twelve diarrhea events were recorded, and only two 167 were excluded because they were suspected to be secondary to general anaesthetics.

177
No deviation was noted with the prescription of rescue metronidazole.

178
In the Smectite group, the median time to resolution of diarrhea was 19.5 hours (IQR: 13.5-179 32 hours; Figure 2). A median of 2 doses (IQR: 1-3) of smectite was administered before 180 resolution of diarrhea, which occurred at a median time of 18 hours (IQR: 13.5-26 hours) 181 from starting it. Smectite administration did not result in any constipation. Metronidazole was 182 administered in 5/54 events (9.3%, 95% CI: 3.1-20.3%). In 3 events metronidazole was 183 administered instead of smectite (protocol deviation), which was then administered as a 184 rescue in 1 case. In 2 events it was administered as a rescue protocol following administration 185 of smectite. The median QOL score was 9.0 (IQR:7.0-9.0), and the median 186 constitutional/gastrointestinal adverse event grade was 0 (IQR: 0-1). Continuous diarrhea was 187 followed by 1 hospitalization, 1 chemotherapy discontinuation and 1 chemotherapy dose 188 reduction.

189
In the Control group, the median time to resolution of diarrhea was 53 hours (IQR: 31.5-190 113.5 hours; Figure 2), which was significantly longer that in the Smectite group (P < 0.001).

196
Smectite was prescribed as a second rescue in 3 cases and was associated with resolution of 197 diarrhea within 24 hours in 2 cases. In the 3 rd case, the diarrhea continued and resolved   204 The overall incidence of diarrhea following chemotherapy administration was 110/426 205 (25.8%, 95% CI: 21.7-30.2%), and significantly differed between the chemotherapeutic drugs 206 administered (P < 0.001). Doxorubicin and vinca-alkaloids were associated with a higher 207 incidence, whereas cyclophosphamide was associated with a low incidence (Table 2). Only

230
The data prospectively collected in this study provides valuable information about CID in 231 dogs, which could be useful to further understand the development and management of this 232 common adverse event.

249
In the current study, CID was significantly associated with a lower QOL score, with a higher 250 incidence of hospitalization, and higher grade of lethargy, anorexia, vomiting and febrile 251 neutropenia toxicities. No significant difference between the Smectite and the Control group 252 was noted in the intention-to-treat analysis, but a causal relationship remains possible and 253 could be further explored with an appropriately powered multicentric study.

254
This study was designed to compare the resolution of CID following the administration of 255 smectite compared to no intervention since there is currently no consensus recommendation 256 regarding the management of CID in dogs. 12 However, for ethical concerns, an identical 257 rescue protocol with the administration of metronidazole was applied in both study groups.

258
Metronidazole was chosen because its use has already been suggested for the management of 259 CID in dogs, 12 and to comply with the local ethical regulation as it was already in our difficile being an increasing concern. 9,38,39 Antibiotics may also be associated with a poorer

309
This study had several limitations. It was non-blinded and the results are therefore subject to 310 bias. However, we believe the strong differences in several outcomes including the time to 311 resolution of diarrhea support the relevance of the findings. Also, no deviation in the 312 prescription of rescue metronidazole was noted, the deviations in completing the diaries were 313 similar in both groups, and owners completed the QOL questionnaires as part of the routine 314 practice in our institution. The second limitation was the early interruption of the trial, 315 increasing the risk of a type I error when analysing the primary outcome. This risk is however 316 minimal since the trial was stopped late in its course achieving nearly 90% of the 317 predetermined sample size after 21 months, and a type I error is considered very unlikely 318 with a P-value < 0.0005, 47 which was our case. Furthermore, the risk of introducing a bias is 319 considered low when a study is stopped independently from the result. 48 The third limitation 320 was the presence of protocol deviations, which may have ultimately affected the results of 321 this study by decreasing or masking the difference in the primary and secondary outcomes 322 between the two groups. A fourth limitation is that dogs were maintained on their routine diet

462
The difference in the incidence rate of chemotherapy-induced diarrhea among drugs was significant by The median faeces consistency grade at the start and at the worst of the diarrhea were 4 (IQR: 3.5-4.5) and 4.5