Fasting reduces the incidence of vincristine‐associated adverse events in dogs

Abstract Fasting has been shown to decrease chemotherapy‐associated adverse events (AEs), in part through insulin‐like growth factor (IGF‐1) reduction, and may induce a protective effect on normal cells during chemotherapy treatment in mice and people. The purpose of this study was to evaluate the effect of fasting on constitutional, bone marrow and gastrointestinal (GI) AEs, and serum glucose, IGF‐1 and insulin levels in dogs receiving vincristine. The study was a prospective, crossover clinical trial in tumour‐bearing dogs. Dogs were randomized to be fasted for 24 to 28 hours prior to and 6 hours following their first or second vincristine treatment, and fed normally for the alternate dose. A significant reduction in nausea, anorexia, lethargy and serum insulin was observed when dogs were fasted; however, no significant differences were found in other GI symptoms, neutrophil count, serum glucose or IGF‐1. Fasting prior to vincristine therapy is a safe and effective treatment modality that helped mitigate constitutional and GI AEs in tumour‐bearing dogs.


| INTRODUCTION
Cancer has emerged as the greatest health concern of dog owners, and maintaining quality of life (QOL) is paramount when pet owners are considering treatment options, especially chemotherapy. [1][2][3] The balance between toxicity, tumour response and patient QOL necessitates the continued investigation of methods to control chemotherapy-associated toxicity. In the absence of such options, owners may decline chemotherapy-based treatment for their pets or limit the extent and aggressiveness of these life-extending treatments.
Chemotherapy agents may cause mild to severe gastrointestinal (GI) side effects, and effective means to mitigate these effects remains elusive. 4 Vincristine is a commonly used chemotherapeutic drug with adverse events (AEs) such as lethargy, vomiting, diarrhoea or cytopenias can occurring in nearly 70% of treated dogs; however, the majority of these AEs are low grade. 5 In a retrospective analysis of owner-reported concerns during chemotherapy, concerns were identified at 48% of appointments, and were significantly related to diarrhoea and remission status. 2 Another study where owners assessed their dogs' QOL during chemotherapy treatment for lymphoma, the majority of owners felt their dogs QOL remained high and 80% would pursue chemotherapy treatment again. However, 32% felt QOL was acceptable but poorer than before the lymphoma diagnosis. 3 Due in part to these QOL concerns and potential adverse effects of chemotherapy, anti-emetic and antidiarrheal treatments are often prescribed. One of these treatments, maropitant citrate a selective neurokinin-1 receptor antagonist, has been shown in a randomized, double-blinded, placebo-controlled study to be an effective anti-emetic in canine cancer patients treated with doxorubicin. 6 However, a recent randomized, prospective clinical trial evaluating the use of maropitant for GI toxicity associated with vincristine and cyclophosphamide revealed no difference in vomiting, diarrhoea, anorexia or lethargy. 5 Further attempts to reduce the incidence of chemotherapy related toxicity have been made. In one study, canine patients fasted prior to doxorubicin administration experienced a reduced incidence of vomiting (10% vs 67% when fed normally). 7 Similarly, a case series of 10 people who voluntarily fasted prior to and after a chemotherapy cycle noted fewer chemotherapy-related toxicities compared with cycles in which they did not fast. 8 Fasting cycles have been shown in in vitro and in vivo models to provide both increased malignant cell apoptosis and protective benefits to normal cells when exposed to chemotherapy. 9-15 Furthermore, Lee et al showed 60% of tumour bearing mice treated with doxorubicin after fasting achieved longterm survival while the entire control group died of metastasis or chemotherapy related toxicities. 10 A recent study in breast cancer patients found a significant reduction in haematological AE when patients fasted prior to treatment. 16 It is proposed that the protective benefits of fasting on normal cells, including GI and haematopoietic cells, are in part because of reduced levels of insulin-like growth factor (IGF-1) and glucose. 10,17 The purpose of this study was to evaluate the effect of shortterm fasting on incidence and severity of GI, constitutional and haematological AEs in dogs treated with vincristine, as well as to identify any correlative effect of fasting on serum glucose, insulin and IGF-1 concentration. Thus, we hypothesized that short-term fasting prior to administration of vincristine would significantly reduce AE associated with chemotherapy, and would reduce serum glucose, insulin and IGF-1.

| Patient selection
Client-owned dogs presenting to Oregon State University Teaching Hospital between June 2016 and April 2018 with the intent of pursuing at least two doses of vincristine (Hospira, Lake Forest, Illinois) as part of their chemotherapy protocol were eligible for enrolment. For dogs receiving a multi-agent protocol, only AE associated with vincristine were evaluated.
Prior to enrolment, dogs diagnosed with lymphoma were permitted to begin treatment with prednisone (0.5-1 mg/kg PO q24, PAR Pharmaceutical, Chestnut Ridge, New York), cyclophosphamide (250 mg/m 2 PO rounded to nearest 25 mg capsule (Cipla, Sunrise, Florida ) and furosemide (1-2 mg/kg PO Vetone, Boise, Idaho). Dogs that experienced GI, constitutive or haematological AE following this treatment were not eligible for enrolment. Dogs on concurrent medications at the time of diagnosis with the potential to alter GI toxicities such as prednisone or a nonsteroidal anti-inflammatory were excluded unless they had been on that medication for greater than 1 week with no reported GI AEs and were anticipated to be on the medication throughout the study period. Dogs on probiotics were excluded.
All dogs had a physical examination, weight, complete blood cell count (CBC), and serum biochemistry performed prior to enrolment.
Dogs were excluded if they weighed <5 kg, had significant alterations on screening blood work suggestive renal or hepatic disease, had a concurrent metabolic disease, pre-existing chronic GI disease or were clinically suspected to have neoplastic GI involvement. An abdominal ultrasound was not a requirement to be included in the study.
Dogs were required to be fed twice daily or ad lib as part of their normal husbandry, and owners were required to feed a consistent diet during the study period. Dogs were excluded if they had previously been treated with vincristine, were positive for the multi-drug resistant gene (MDR-1, ABCB1) mutation, were typically fed once per day, had a diet change within 1 week of enrolment, or showed symptoms of nausea, vomiting or diarrhoea within 2 days prior to trial enrolment.
The study design, treatment protocol and informed client consent was approved by the Institutional Animal Care and Use Committee.
Once the study was complete, dogs were free to pursue further therapy as desired by their owner.

| Treatment protocol
A prospective, controlled, crossover clinical trial was implemented to evaluate the effect of short-term fasting on AE's from chemotherapy, and serum glucose, insulin and IGF-1. Each dog served as its own control with a crossover design. Randomization was performed via a coin flip. Dogs were fasted prior to their first chemotherapy treatment and then fed their regular feeding routine (twice daily, timing varied with normal husbandry habits) prior to their second chemotherapy treatment or were fed their regular feeding routine prior to their first chemotherapy treatment and fasted prior to their second treatment. Fasted dogs received no food after 12 PM the day prior to the chemotherapy treatment and had free access to water. All dogs were fed normally the evening of their chemotherapy appointment (approximately 6 PM), thus total fasting times ranged from approximately 30 hours for ad libitum fed dogs to 34 hours for dogs fed twice daily. All dogs were evaluated by physical exam, owner assessment of QOL survey and a CBC within 48 hours of each chemotherapy treatment. 18 Prophylactic anti-emetic (ie, maropitant, ondansetron and metoclopramide) or anti-diarrheal (ie, metronidazole and tylosin) medications were not permitted.
Dogs were treated within 1 hour of noon the day of their chemotherapy treatment. Dogs were treated with a standard chemotherapy dosage calculated by body surface area. Vincristine treatments were administered at 0.6 mg/m 2 intravenously (IV) and given at least 1 week apart. A CBC was scheduled 7 to 10 days after each dose of vincristine. A physical exam and owner assessment QOL survey were performed at each appointment.

| Response and toxicity assessment
Clients were sent home with AE journals after each chemotherapy appointment. They were instructed to note any changes in appetite or activity level in the journal as well as any episodes of vomiting, diarrhoea or nausea. GI and constitutional AE were graded based on owner journal documentation and patient history gathered at each appointment according to the Veterinary Cooperative Oncology Group-Common Terminology Criteria for Adverse Events (VCOG-CTAE) v1.1 scheme. 19 Similarly, haematologic AE were graded based on follow-up CBC data, 7 to 10 days post each vincristine dose. Grading was performed by the attending clinician. Dogs experiencing grade II or higher vomiting or nausea (according to the VCOG-CTCAE v1.1) could be treated as clinically indicated with ondansetron and/or maropitant, depending on clinician preference. The same anti-emetic drug(s) was required to be used as indicated for each dog for the duration of the study period. Dogs experiencing grade II or higher diarrhoea (VCOG-CTACE v1.1) were treated with metronidazole as indicated based on clinician preference.
Dogs were removed from the study if a grade III AE occurred that precluded continuation of chemotherapy at the same dose, progressive disease was noted, or an owner requested study withdrawal for any reason.

| Serum glucose, insulin, and IGF-1 measurement
Peripheral blood samples were obtained via jugular venipuncture immediately prior to each chemotherapy treatment to attain a fasted and fed sample (total of two time points) for each patient. Glucose was tested through Oregon State University Veterinary Diagnostic Lab. Glucose samples were processed (centrifuged and serum separated) within 1 hour of collection. Samples intended for insulin and IGF-1 measurement were allowed to clot and the serum was separated by centrifugation. The serum was promptly frozen and stored at −80 C until all samples were collected within the study period. Serum insulin and IGF-1 concentrations were batched and measured using ALPCO Diagnostics (Salem, New Hampshire) porcine/canine insulin enzyme-linked immunosorbent assay (ELISA) and IGF-1 ELISA, respectively.

| Statistical analysis
A power analysis was performed to estimate the number of dogs needed to detect a statistically significant difference in AE incidence between fasting and fed time points. Based on the literature, expectations were that dogs would have an AE incidence rate of 10% vs 40% to 60%. 5,7,20 Using an alpha level of 0.05 and a power of 0.80, a mini- thrombocytopenia. The Wilcox-signed rank tests were used to evaluate paired data on GI AE severity. Paired t tests were used to evaluate the effect of fasting and order of treatment on neutrophil count, and glucose concentration. Effects of fasting/feeding and order on glucose, insulin and IGF-1 levels were tested with a linear mixed model to account for within dog correlation of duplicates. The first arm (ie, order = 1) of the study was analysed separately for treatment effects with two parallel groups to evaluate treatment effects in the first treatment only, absent any potential order effects. A P-value of <.05 was considered significant.

| CELL LINE VALIDATION STATEMENT
This study did not involve cell lines; therefore, cell line validation work was not carried out.

| RESULTS
Eighteen client-owned dogs were enrolled. All dogs were diagnosed with multi-centric lymphoma via cytology and intended to initiate treatment with the CHOP protocol. Dogs were treated with cyclophosphamide (250 mg/m 2 PO rounded to the nearest 25 mg capsule) and furosemide (1-2 mg/kg PO) 7 to 10 days prior to enrolment with no reported GI, constitutive or haematological effects. Dogs were treated with prednisone pre-enrolment (0.5-1 mg/kg) and continued on the same dose throughout the study period. Seven patients were tested for MDR-1 gene mutation (ABCB1) through the Washington State University, and homozygous normal dogs were enrolled (n = 6).
Two dogs were removed from the study following enrollment and prior to vincristine administration. One because of an MDR-1 mutation and one because of elevated renal values at initial screening. One dog (dog 8) was removed from the study because of a grade IV neutropenia following the first vincristine dose. This dog's AE data were included within the first dose analysis. Fifteen dogs completed the study period. Two dogs (dog 9 and dog 10) were removed from the serum ELISA testing because of inadequate sample availability and one dog (dog 2) only had glucose concentration reported from the fasted dose. Thus, 15 dogs were included in the AE analysis and 13 were included in the serum analysis. Figure 1 and Table 1 outline dog specific details to summarize the study population, which con- Anti-nausea and anti-diarrhoea therapeutic intervention were not necessary, as no dogs experienced greater than grade two GI AEs.
However, one dog (dog 5) was progressively anaemic and was treated with gastroprotectants throughout the study protocol.

| DISCUSSION
This prospective crossover clinical trial evaluated the effect of fasting on vincristine-associated GI, constitutive and haematological AEs, as well as and serum glucose, IGF-1 and insulin levels. Our results demonstrate that fasting for 24 hours prior to treatment with vincristine significantly reduced the incidence of nausea, anorexia and lethargy, and significantly decreased serum insulin. In contrast, fasting prior to  Note: "X" represents no data collected, as this dog was removed from the study following the fed dose.
G1 phase may be less sensitive to doxorubicin, which preferentially targets cells in S phase. 7,25 Similarly, vincristine primarily targets cells in M phase and thus G1 cell cycle blockade may be protective. 23 in vitro studies in acute lymphoblastic leukaemia cells reported G1 cells were more susceptible to vincristine; however, a follow-up study found this effect limited to high doses and was not repeatable with another cell line. 26,27 Another proposed mechanism is that nutrient deprivation from fasting triggers various cellular pathways to invest energy in repair and maintenance rather than proliferation. This is supported by a study where mice fasted for 24 hours prior to a high dose of etoposide had less DNA damage noted in small intestinal stem cells 3 hours after treatment compared with the mice fed normally, despite similar levels of DNA damage between groups had been noted 1.5 hours after treatment. Thus, DNA repair is hypothesized to be more efficient during periods of nutrient deprivation. 15 The most investigated potential regulator of cellular stress resistance during fasting is IGF-1. Mice with low serum IGF-1 because of gene deletion are relatively resistant to high dose chemotherapy; this benefit was negated when IGF-1 is administered IV. 10 Serum IGF-1 has been shown to decrease in fasted states, however, robust reduction in humans requires 48 hours of fasting. [28][29][30] Fasting causes low serum insulin levels, which leads to growth hormone resistance in the liver and consequently inhibits hepatic IGF-1 production. Thus, one would expect glucose, insulin and IGF-1 to be positively correlated. In our study, we noted significantly reduced insulin when dogs were fasted; however, this was not associated with IGF-1 reduction. Dogs with lymphoma may have abnormal carbohydrate metabolism, as this group has been reported to have higher insulin levels compared with controls. 31,32 However, a recent study found no difference in serum insulin, IGF-1 or glucose in dogs with lymphoma compared with controls. 33 These parameters were not correlated in our study population, although based on previous reports, it is possible that longer fasting intervals would be necessary to observe such serum concentration changes. 21 Additionally, we report a significant difference in multiple GI AEs.
More robust improvement in the incidence of GI AEs may be because of a longer duration of fasting in our study (approximately 28 hours prior to treatment and 6 hours following treatment) and the relatively rapid ter-  this study excluded small dogs (<5 kg) because of concerns for potential hypoglycemia associated with fasting. As such, the patient population consisted of mostly large breed (>25 kg) neutered dogs, and conclusions from this study should only be made for medium to largesize dogs. Small breed dogs have increased mass-specific metabolic rates compared with larger dogs and fasting may have a more profound effect on serum glucose levels. 40 Further evaluation of the safety and efficacy of fasting in cancer-bearing small dogs is warranted. It is important to note that this study was not designed to assess the effects of fasting related to differences in patient cancer diagnosis, stage of disease or immunophenotype of lymphoma.
All dogs enrolled were diagnosed with lymphoma, and the majority were B-cell. It is possible the effect of fasting on AE associated with vincristine, and glucose, insulin and IGF-1 concentrations vary with patient stage and immunophenotype of lymphoma.
Finally, dogs were prescribed prednisone throughout the study period, which may have affected our results. However, dogs were required to have tolerated this medication for at least 1 week and were maintained at the same dose throughout the trial period. As a comparison, a study of dogs diagnosed with inflammatory bowel disease found no significant effect of prednisone on IGF-1. 41 Interestingly in some dogs, we noted a wide range of IGF-1 concentration between duplicate samples, and serum colour was grossly different between samples from fasted and fed time points. ALPCO ELISA serum insulin manufacturer claims to have no interference with triglyceride, bilirubin, and haemoglobin and IGF-1 serum ELISA is reported to have no interference of triglyceride, bilirubin and haemoglobin at concentrations below 100 mg/mL, 200 ug/mL, or 1 mg/mL, respectively. 42 Dogs with hyper-bilirubinemia were not included in our study; however, we were unable to measure lipaemia or haemolysis, thus we concede that it is possible that haemolysis and/or lipaemia may have affected our values. Similar variance has not been reported in other studies, but it is not always noted if assays were run in duplicate or if results were presented as the mean.

| CONCLUSION
Adverse effects related to chemotherapy remain a significant limitation to dose intensity, potential outcome as well as a cancer patient's QOL. Our findings suggest fasting is a reasonable treatment to consider for GI and constitutional AEs associated with vincristine in large breed dogs. Additional studies are warranted to validate the potential benefits of fasting prior to chemotherapy treatment in a larger cohort and explore optimal timing of re-feeding.