Randomized Trial of Cisplatin versus Firocoxib versus Cisplatin/Firocoxib in Dogs with Transitional Cell Carcinoma of the Urinary Bladder

Abstract

Background

Cisplatin combined with a nonselective cyclooxygenase (cox) inhibitor has potent antitumor activity against transitional cell carcinoma (TCC) in dogs, but this treatment is limited by renal toxicosis. Cox-2 is expressed in TCC, but only in limited sites within the kidney. A cox-2 inhibitor could enhance the antitumor activity of cisplatin with potentially fewer adverse effects on the kidney.

Hypothesis

Cisplatin/cox-2 inhibitor treatment will have greater antitumor activity but no more renal toxicosis than cisplatin alone in dogs with TCC.

Animals

Forty-four dogs with naturally occurring urinary bladder TCC.

Methods

Dogs were randomized to receive cisplatin (60 mg/m² IV q21d), firocoxib (5 mg/kg PO q24h), or the combination. Tumor measurements were determined before and at 6-week intervals during treatment. Renal function was monitored by serum creatinine concentration, iohexol clearance, and urine specific gravity. Toxicoses were graded according to Veterinary Co-Operative Oncology Group (VCOG) criteria.

Results

The remission rate with cisplatin/firocoxib (57%) was significantly (P = .021) higher than that with cisplatin alone (13%). Renal and gastrointestinal toxicoses were common in dogs receiving cisplatin, but there were no significant differences between dogs receiving cisplatin or cisplatin/firocoxib. Firocoxib alone induced partial remission or stable disease in 20 and 33% of dogs, respectively.

Conclusions

Firocoxib significantly enhanced the antitumor activity of cisplatin resulting in partial remission in more than half of the cases. The toxicoses inherent to cisplatin, however, were noted in dogs receiving this combination. Firocoxib had antitumor effects as a single agent and can be considered a palliative treatment for dogs with TCC.

Abbreviations

CBC

complete blood count

Cox

cyclooxygenase

CR

complete remission

GI

gastrointestinal

PD

progressive disease

PFI

progression free interval

PR

partial remission

QOL

quality of life

SD

stable disease

TCC

transitional cell carcinoma

VCOG

Veterinary Co-Operative Oncology Group

WHO

World Health Organization

Invasive transitional cell carcinoma (TCC) is the most common form of urinary tract cancer in dogs.[1] Canine TCC is rarely amenable to surgery because of frequent trigonal location in the bladder and the presence of metastases in approximately 20% of cases at diagnosis and in 50% of cases at death. Most dogs with TCC are treated with chemotherapy, cyclooxygenase (cox) inhibitors, or combinations of these drugs.[1] Mitoxantrone combined with piroxicam has been the most widely used treatment for TCC, resulting in remission in 35% and stable disease in 46% of cases.[2]

Cisplatin combined with piroxicam has had greater antitumor effects in dogs with TCC (remission rate 71%), but is not recommended because of frequent renal toxicosis.[3] The renal toxicosis is thought to be because of direct toxic effects of cisplatin on renal tubules and decreased renal blood flow related to inhibition of cox, especially cox-1, in renal blood vessels.[4, 5]

The antitumor effects of cox inhibitors are due, at least in part, to inhibition of cox enzyme activity.[6] The isoform of cox expressed in the majority of TCC masses is cox-2.[7] Although cox-2 is present in renal tissue, such as in the macula densa, cox-1 expression is more prominent in the kidney.[5, 8] Cox-2 inhibitors are expected to be a potentially safer choice when combined with cisplatin. In rats, cox-2 inhibitors did not worsen the renal toxicosis of cisplatin.[9] Cox-2 inhibitors have even been reported to have a protective effect against cisplatin toxicosis in rodents, presumably through their anti-inflammatory effects.[10, 11] Cisplatin combined with the cox-2 inhibitor, celecoxib, also has been safely tolerated in humans.[12]

This study was performed to determine the extent to which a cox-2 inhibitor, firocoxib, could enhance the antitumor activity of cisplatin in dogs with TCC while not increasing renal toxicosis. Firocoxib was chosen for its selective cox-2 inhibitory effects and excellent safety profile in dogs.[13, 14]

Methods

A multi-institutional (Purdue University, University of Missouri) prospective, randomized clinical trial was performed in dogs with TCC of the urinary bladder. The study was performed with Institutional Animal Care and Use Committee approval and according to Good Clinical Practice standards.

Entry and Exclusion Requirements

Entry requirements included dogs with histopathologically confirmed TCC of the urinary bladder, measurable (by contrast cystography) lesions in the bladder, normal serum creatinine concentration, expected survival of ≥6 weeks, and written informed consent signed by pet owners. Dogs were excluded if they had received cisplatin at any time, if they had received a cox inhibitor within the previous 14 days, or if they had received a cox inhibitor for ≥1 month in the previous 3 months.

Case Evaluation and Tumor Staging

Pretreatment evaluation included physical examination with rectal examination (to detect urethral lesions and prostatomegaly), CBC, serum biochemical profile, urinalysis and culture, thoracic radiographs (3 views), abdominal ultrasound examination, and cystogram. This evaluation was repeated at 6-week intervals. For the first 10 dogs in each treatment group, iohexol clearance1 was determined before and after 6 weeks of treatment. A CBC, serum biochemical profile, and urinalysis were performed before each dose of cisplatin. A CBC was performed 7–10 days after each cisplatin treatment. Abdominal ultrasound examination was used to detect possible metastases. When feasible, fine needle aspirates or biopsies of lesions suggestive of metastases were obtained. The TNM stage was recorded according to World Health Organization criteria.[15]

The size of bladder lesions was assessed by cystography similar to what has been previously described.[16] Briefly, dogs were placed under general anesthesia, and a urinary catheter passed. The bladder was distended with air (4–9 mL/kg) and meglumine diatrizoate (0.22 mL/kg). Abdominal radiographs were made with the dog in left lateral, right lateral, ventrodorsal, and dorsoventral recumbency. Up to 5 prominent lesions that could be observed easily were selected for measurement, and cranial-caudal, dorsal-ventral, and medial-lateral measurements of the tumor lesions were recorded. These 3 values were multiplied to obtain the estimated volume of each lesion. The estimated tumor volume for the dog was determined from the sum of the volumes of the lesions. On follow-up examinations, the same degree of bladder distention with contrast was used, and the same lesions on the same views were measured. The presence of any new lesions was recorded.

Treatment

A computer-based randomization scheme, with cases stratified by enrollment site, was used. A unique allocation sequence using the PLAN procedure (SAS^® v 8.22) was generated and provided to each study site. Dogs were randomized to receive cisplatin alone, firocoxib alone, or cisplatin combined with firocoxib (cisplatin/firocoxib). Dogs in the single agent treatment arms that had tumor progression or unacceptable toxicoses were eligible then to receive the other single agent treatment. Dogs failing the study drugs were eligible to receive other treatment off study.

Cisplatin

Cisplatin3 was given at 60 mg/m² IV every 21 days. Saline diuresis (0.9% saline, 18 mL/kg/h IV) was given for 4 hours before and 2 hours after cisplatin administration. To reduce the risk of nausea, butorphanol4 (0.4 mg/kg IM) was given 30 minutes before and 90 minutes after cisplatin administration. If this treatment did not control the nausea, ondansetron5 (1.0 mg/kg IV 30 minutes before and 90 minutes after cisplatin administration) was given on subsequent cisplatin treatments. The study called for cisplatin treatment to be continued until 2 doses had been given after documentation of complete remission (CR) or until progressive disease (PD) was noted. In addition, cisplatin was discontinued if unacceptable toxicoses (deemed unacceptable by the attending clinician or pet owner) were noted. When cisplatin was discontinued, the dog was eligible to receive single agent firocoxib.

Cisplatin therapy was delayed if laboratory results on the day of planned treatment included a neutrophil count <2500/μL, a platelet count <100,000/μL, or a serum creatinine concentration >1.5 mg/dL. If treatment was delayed, laboratory tests were re-evaluated weekly for 4 weeks. Treatment was resumed when the neutrophil and platelet counts returned to acceptable numbers. If the serum creatinine concentration returned to normal, treatment was resumed at the original dose. If the serum creatinine concentration persisted at 0.1–0.5 mg/dL above the site's reference range, cisplatin was reinstituted at 50 mg/m². If marked myelosuppression (neutrophil count <1500/μL or platelet count <50,000/μL) or other major toxicosis had occurred, the subsequent dosage of cisplatin was given at 50 mg/m². Further reduction in dose (to 40 mg/m²) was made if unacceptable toxicosis was noted at the 50 mg/m² dosage. If return to acceptable laboratory findings did not occur within 4 weeks of the previous dose of cisplatin, then cisplatin was discontinued. In all 3 treatment arms, if a dog became ill, a medical history, physical examination, CBC, serum biochemical profile, urinalysis, and other tests indicated based on the clinical signs were performed. Any drug toxicoses, as well as other comorbid conditions, were treated as deemed appropriate by the attending veterinarian. This could include fluids administered IV, antibiotics, antiemetics, and other drugs specific for the condition that occurred.

Firocoxib

Firocoxib6 was given PO once a day in 57 or 227 mg tablets with half to whole tablets used to achieve a dosage of approximately 5 mg/kg. Pet owners recorded drug administration and observations of the dog daily. If any unacceptable toxicosis occurred with firocoxib, the drug was withdrawn, and dogs were eligible to receive cisplatin.

Combination Treatment

Dogs receiving cisplatin and firocoxib were treated with the same drug dosages as those used in the single agent arms. If toxicoses occurred, drug administration was adjusted as described for the dogs receiving single agent treatment.

Tumor Response and Toxicoses

Tumor response was defined as (1) complete remission (CR), complete resolution of any radiographic and clinical evidence of tumor; (2) partial remission (PR), ≥50% reduction in estimated tumor volume, no increase (of 50% or more) in the volume of any individual lesion, and no new lesions observed; (3) stable disease (SD), <50% change in estimated tumor volume, no increase (of 50% or more) in the volume of any individual lesion, and no new lesions; and (4) progressive disease (PD), ≥50% increase in tumor volume, or increase (of 50% of more) of the volume of any individual lesion, or the development of new lesions. Progression-free interval (PFI) was defined as the time from the first day of treatment until PD was noted. For dogs that were discontinued from treatment because of toxicosis before PD occurred, the PFI was censored at the date that subsequent treatment was initiated even if PD had not occurred at that time. The survival time was defined as the time from the first day of treatment until death. Permission to perform a necropsy at the time of death was requested. Toxicoses were graded according to Veterinary Co-operative Oncology Group (VCOG) criteria.[17]

Quality of Life (QOL)

Pet owners were asked to complete a QOL assessment for their dog before treatment and at 6-week intervals during treatment. Information included characterization of the urination (extent of straining, frequency of urination during the day and at night, blood in the urine), overall activity level, and appetite. In each category, a score of 1 to 5 was assigned with 5 being the best performance score.

Sample Size and Analyses

The primary endpoint was remission (partial or complete). The remission rate in the single agent arms was anticipated to be 15–20%, and the remission rate in the combination treatment arm was anticipated to be 50–75%. To detect a difference in remission rate of 35% between the single agent arm and combination treatment arm, it was expected that 34 dogs would be needed in each treatment group (power 0.90; 2-sided test). Interim analyses were scheduled to be conducted after 12 and 25 dogs had been entered in each treatment arm. If a significant difference in remission rate was noted at the time of the interim analysis, then no further enrollment would occur. It was considered inappropriate to continue enrolling more dogs in the trial if 1 treatment arm was clearly superior.

The 3 treatment groups were compared for differences in age, breed, weight, sex, TNM stage, urethral involvement, prostate involvement, and initial renal function (serum creatinine concentration, iohexol clearance, urine specific gravity). Analyses related to treatment included number of cisplatin doses; gastrointestinal (GI), hematologic, and renal toxicoses; tumor response; PFI; and duration of survival. Categorical variables were assessed by the Chi-square test of independence and the Fisher exact test. Continuous variables were analyzed by the Kruskal-Wallis nonparametric analysis of variance and the Wilcoxon rank sum test. The Kaplan-Meier product-limit method was used to estimate time until PD and overall survival time. Time-to-event comparisons between treatment groups were made by the log-rank test. Evaluation of hazard ratios for potential risk factors was done by forward stepwise multivariate Cox proportional hazard regression, selecting variables with values of P < .30 in univariate analysis for the multivariate model. Confidence intervals (95%) were calculated for hazard ratios. For all statistical analyses, a P value of <.05 was considered significant.

Results

Cases Enrolled and Treatment Given

The 1st interim analysis indicated a significant difference in the remission rate between dogs receiving cisplatin alone (13%) and cisplatin/firocoxib (57%) (P = .021), and enrollment was stopped at that time. Enrollment included 36 dogs at Purdue University and 8 dogs at the University of Missouri. Subject characteristics and TNM stage are presented in Table 1. Biopsy-confirmed nodal metastases were present in 6 dogs, and presumed nodal metastases were present in 5 dogs. Distant metastases included biopsy-confirmed lung metastases in 8 dogs, presumed lung metastasis (based on radiography) in 1 dog, and biopsy-confirmed metastases in the liver, kidney, and skin (1 dog each). There were no significant differences between the groups at enrollment with regard to age, sex, breed, weight, T stage, N stage, M stage, urethral or prostate involvement, serum creatinine concentration, iohexol clearance, or urine specific gravity. A summary of the cisplatin doses given to dogs in the trial is presented in Table 2.

Table 1. Subject characteristics and cancer stage for dogs participating in the clinical trial

Number of Dogs	Cisplatin	Firocoxib	Combination
	15	15	14
Age (years)
Median	10.0	10.8	11.5
Mean + SD	10.4 + 2.8	10.3 + 2.1	11.3 + 2.1
Range	6.0–15.0	6.0–13.8	8.0–14.0
Gender (number of dogs)
FS	10	9	9
FI	0	0	0
MC	5	4	5
MI	0	2	0
Breed (number of dogs)
Scottish Terrier	3	1	3
West Highland White Terrier	2	0	1
Shetland Sheepdog	1	2	0
Beagle	2	2	2
Other pure breed	3	6	7
Mixed breed	4	4	1
Weight (kg)
Median	14.9	16.7	14.3
Mean + SD	17.9 + 9.9	21.9 + 11.4	16.7 + 12.3
TNM stage (number of dogs)
T1N0M0	0	0	1
T2N0M0	8	5	8
T2N1M0	1	1	1
T2N0M1	2	2	1
T2N1M1	1	1	1
T3N0M0	2	2	1
T3N1M0	1	3	1
T3N0M1	0	1	0
T3N1M1	0	0	0
Any metastasis	5	8	4
Urethra involved	7	11	5
Prostate involved	2/5 male dogs	4/5 male dogs	2/5 male dogs

Table 2. Summary of cisplatin treatments given to dogs in the trial

Treatment Group	Cisplatin 15 Dogs	Cisplatin after Failing Firocoxib 5 Dogsa	Combination 14 Dogs
^a When dogs in the firocoxib alone arm had cancer progression, the dogs were eligible to receive cisplatin alone. Of 15 dogs in this treatment group, 5 dogs received cisplatin.
Number doses given to group
at 60 mg/m^2	33	7	27
at 50 mg/m^2	13	2	5
at doses <50 mg/m^2	0	0	3
Number of doses/dog
Median	2	2	2
Mean	3.1	1.8	2.6
Range	1–6	1–3	1–8
Dose reduction, number of dogs (% of treated dogs)	6 (40%)	2 (40%)	3 (21%)
Reason to discontinue cisplatin, number of dogs (% of treated dogs)
Cancer progression	7 (47%)	3 (60%)	2 (14%)
Toxicosis	8 (53%)	2 (40%)	10 (71%)
Other	0	0	2 (14%)

Antitumor Activity and Survival

The tumor responses at treatment day 42 are summarized in Table 3. There were no dogs for which the tumor response category improved after 42 days (ie, if cancer was stable at 42 days, it did not further regress to become classified as remission at a later date). Treatment response was assessable in 14 of 15 dogs receiving cisplatin, 12 of 15 dogs receiving firocoxib, and 11 of 14 dogs receiving cisplatin/firocoxib. In the cisplatin alone group, 1 dog developed severe GI toxicosis, and the owner elected to stop treatment after 21 days. In the firocoxib alone group, 2 dogs received firocoxib intermittently and then stopped receiving the drug altogether after 4 and 19 days. Both dogs had progressive GI toxicosis. At the time of death and necropsy, both dogs had widespread TCC metastases including intestinal metastases in the dog treated intermittently for 19 days. Another dog in this treatment group stopped firocoxib after 33 days because of increasing serum creatinine concentration associated with cancer progression and poor QOL. In the combination treatment group, 1 dog was stopped from treatment after 4 days because of severe GI toxicosis; a 2nd dog died unexpectedly after knee surgery; and a 3rd dog was stopped from treatment after 14 days because of pyelonephritis and increasing serum creatinine concentration.

Table 3. Tumor responses at 42 days of treatment in dogs in the study

Tumor Responsea	Cisplatin Number (%) of Dogs	Firocoxib Number (%) of Dogs	Cisplatin/Firocoxib Number (%) of Dogs
^a 16 Tumor response as defined in text. ^b 17 Tumor response at 6 weeks not available (see text).
CR	0 (0)	0 (0)	0 (0)
PR	2 (13)	3 (20)	8 (57)
SD	8 (53)	5 (33)	3 (21)
PD	4 (27)	4 (27)	0 (0)
NAb	1 (7)	3 (20)	3 (21)

In dogs with multiple lesions, not all of the lesions responded similarly to treatment, especially in dogs receiving firocoxib. Tumor response did not always reflect the clinical benefit to the dog. One dog receiving firocoxib alone, for example, had marked reduction in a trigonal mass and greatly improved urination, but was assigned a tumor response of PD because of a new lesion in the bladder apex. Another dog receiving firocoxib alone had complete radiographic resolution of 2 lung nodules, PR of a urethral mass, and SD of a bladder mass, and thus was assigned a tumor response of SD.

As noted in Table 2, treatment was discontinued in some dogs because of toxicity before development of PD. The PFI was calculated for dogs remaining on that treatment until PD was noted. The PFI was longer (P = .026) in dogs treated with cisplatin/firocoxib (median, 186 days) than in dogs treated with cisplatin alone (median, 87 days). The PFI for dogs receiving firocoxib (median, 105 days) was not significantly different (P ≥ .231) than PFI in dogs in the other treatment arms. Kaplan–Meier curves for PFI are presented in Figure 1. PFI was associated with M stage (hazard ratio, 5.78; 95% CI, 1.63–20.50; P = .007) and urethral involvement of the tumor (hazard ratio, 2.38; 95% CI, 1.02–5.55; P = .045).

Figure 1. Kaplan–Meier curves for progression free interval (PFI, days) in dogs treated with cisplatin alone (n = 7), firocoxib alone (n = 12), or the combination of the 2 drugs (n = 9). Note that dogs that discontinued treatment prior to the development of PD were not included in the Kaplan–Meier curves for PFI.

Of 15 dogs receiving cisplatin, 13 dogs went on to receive firocoxib after failing cisplatin, and tumor responses included 2 (15%) PR, 4 (31%) SD, and 7 (54%) PD. Of the 15 dogs initially receiving firocoxib, 5 dogs went on to receive cisplatin alone after failing firocoxib. Three of the 5 dogs had PD, and 2 dogs were not available for assessment of cisplatin response. After failing the study drugs, other therapies were given to 6 dogs in the cisplatin group (chemotherapy in 3 dogs, cox inhibitors in 3 dogs), 3 dogs in the firocoxib group (chemotherapy in 2 dogs, cox inhibitors in 1 dog), and 7 dogs in the combination treatment group (chemotherapy in 4 dogs, cox inhibitors in 3 dogs). Other therapies included piroxicam, deracoxib,7 mitoxantrone/piroxicam, carboplatin/piroxicam, doxorubicin, cyclophosphamide, gemcitabine, and an experimental apoptosis-inducing agent.

Survival did not differ significantly among dogs in the 3 treatment groups (P = .072) (Fig 2). The median survival was 338 days for dogs initially receiving cisplatin, 152 days for dogs initially receiving firocoxib, and 179 days for dogs receiving the combination.

Figure 2. Kaplan–Meier curves for survival (days) in dogs treated with cisplatin alone, firocoxib alone, or the combination of the 2 drugs.

Toxicoses and QOL

GI toxicoses observed in the trial are summarized in Table 4. Evidence of grade 3 or 4 toxicoses occurred in 6 of 15 dogs (40%) receiving cisplatin alone, 1 of 15 dogs (7%) receiving firocoxib alone, and 7 of 12 dogs (58%) receiving the combination. GI toxicoses were more common in dogs receiving cisplatin (P = .002) or cisplatin/firocoxib (P = .022) compared with dogs receiving firocoxib alone. There was no significant difference in GI toxicoses between dogs receiving cisplatin alone or cisplatin/firocoxib. Evidence of weight loss was noted in dogs receiving cisplatin (0.95 ± 1.1 kg loss over 42 days) and cisplatin/firocoxib (0.59 ± 0.65 kg loss over 42 days), but not in dogs receiving firocoxib alone (0.1 ± 1.0 kg increase over 42 days).

Table 4. Summary of gastrointestinal and hematologic toxicoses observed in dogs in the trial following VCOG criteria[17]

VCOG Score for Toxicoses	Cisplatin Alone Number of Dogs	Firocoxib Alone Number of Dogs	Combination Number of Dogs
Gastrointestinal toxicosis
0	1	10	4
1	6	2	1
2	2	2	0
3	4	1	6
4	2	0	1
Data not available	0	0	2
Hematologic toxicosis
0	6	14	7
1	5	1	3
2	1	0	1
3	2	0	0
4	1	0	0
Data not available	0	0	3

Hematologic toxicoses are summarized in Table 4. Bone marrow suppression occurred in 60% of dogs receiving cisplatin and 40% of dogs with follow-up receiving cisplatin/firocoxib. Hematologic toxicosis did not differ (P = .753) between dogs receiving cisplatin alone and dogs receiving cisplatin/firocoxib. One dog receiving firocoxib had mild thrombocytopenia (170,000 platelets/μL) of unknown origin on 1 visit.

Renal toxicoses are summarized in Figure 3. The amount of increase in serum creatinine concentration between day 0 and 42 was greater (P = .03) for dogs treated with cisplatin alone or cisplatin/firocoxib than with firocoxib alone. After 6 weeks of treatment, the serum creatinine concentration was above the reference range in 5 of 15 dogs receiving cisplatin, 2 of 13 dogs receiving firocoxib, and 5 of 11 dogs receiving the combination treatment. The increase in serum creatinine concentration did not differ (P = .979) between dogs treated with cisplatin alone or with cisplatin/firocoxib.

Figure 3. Change in serum creatinine concentration (mg/dL) and iohexol clearance (mL/min/kg) after 42 days of treatment.

Before treatment, iohexol clearance was below the reference range in 1 dog receiving cisplatin, 3 dogs receiving firocoxib, and 2 dogs receiving the combination. All of the dogs had normal serum creatinine concentration pretreatment. There were no associations between serum creatinine concentration or iohexol clearance at the start of treatment and development of bone marrow or GI toxicoses during treatment or tumor response to treatment. There were no significant differences among the 3 treatment groups with regard to the change in iohexol clearance with treatment. After treatment, iohexol clearance was below the reference range (2.9–8.1 mL/min/kg) in 5 of 9 dogs receiving cisplatin (3 with increased serum creatinine concentration), 3 of 7 dogs receiving firocoxib (all with normal creatinine concentration), and 9 of 10 dogs receiving cisplatin/firocoxib (4 with increased serum creatinine concentration). In all dogs in which the serum creatinine concentration was above normal, iohexol clearance also was below the reference range.

There was no significant change in urine specific gravity with treatment in any of the 3 groups. The mean ± SD of specific gravity before and after treatment was 1.024 ± 0.020 and 1.024 ± 0.022 for dogs receiving cisplatin, 1.025 ± 0.020 and 1.029 ± 0.011 for dogs receiving firocoxib, and 1.021 ± 0.009 and 1.033 ± 0.017 for dogs receiving the combination treatment.

QOL data are summarized in Table 5. The QOL was maintained or improved in the majority of dogs in all 3 treatment groups.

Table 5. Change in QOL with treatment. A score from 1 to 5, with 5 being the best performance, was assigned to each performance category

	Cisplatin	Firocoxib	Combination
Quality of urination
Decrease with treatment, number of dogs	1	1	1
Increase with treatment, number of dogs	7	8	9
No change with treatment, number of dogs	5	4	1
Data not available, number of dogs	2	2	3
Pre treatment score, mean ± SD	2.6 ± 0.5	2.5 ± 0.6	2.6 ± 0.3
Post treatment score, mean ± SD	2.9 ± 0.5	2.9 ± 0.8	3.2 ± 0.6
Activity level
Decrease with treatment, number of dogs	3	1	0
Increase with treatment, number of dogs	1	2	4
No change with treatment, number of dogs	9	9	7
Data not available, number of dogs	2	3	3
Pre treatment score, mean ± SD	3.5 ± 0.9	3.4 ± 0.9	3.1 ± 1.2
Post treatment score, mean ± SD	3.5 ± 0.7	3.7 ± 0.9	3.6 ± 1.0
Appetite
Decrease with treatment, number of dogs	5	0	1
Increase with treatment, number of dogs	2	3	2
No change with treatment, number of dogs	6	9	8
Data not available, number of dogs	2	3	3
Pre treatment score, mean ± SD	2.9 ± 1.0	3.5 ± 0.7	3.1 ± 1.3
Post treatment score, mean ± SD	3.4 ± 0.7	3.9 ± 0.3	3.3 ± 1.2
All categories combined
Decrease with treatment, number of dogs	4	1	1
Increase with treatment, number of dogs	6	8	10
No change with treatment, number of dogs	3	3	0
Data not available, number of dogs	1	3	3
Pre treatment score, mean ± SD	2.9 ± 0.4	2.8 ± 0.6	2.8 ± 0.4
Post treatment score, mean ± SD	3.1 ± 0.4	3.1 ± 0.6	3.2 ± 0.5

Discussion

The study confirmed that a cox-2 inhibitor can significantly enhance the activity of cisplatin in dogs with TCC, similar to that reported with nonselective cox inhibitors.[3] Cisplatin/firocoxib resulted in remission in 57% of dogs and SD in 20% of dogs for a cancer control rate of 77%. The PFI was significantly longer in dogs receiving cisplatin/firocoxib (median, 186 days) than in those receiving cisplatin alone (median, 87 days). The mechanisms by which cox inhibitors enhance the activity of cisplatin remain to be completely defined. Induction of apoptosis has been a consistent finding in previous studies of cox inhibitors in cancer, including studies of dogs and humans with invasive TCC.[18-20] Induction of apoptosis is crucial for cisplatin-induced DNA damage to cause cell death.[21] Cox inhibitors may enhance the ability of cancer cells to undergo apoptosis, thus enabling cisplatin to kill the cancer cells. Other mechanisms are also possible.[18, 20, 21]

Although survival did not differ significantly among treatment groups, the survival in dogs receiving cisplatin alone followed by firocoxib alone (median, 338 days) compared with that of dogs receiving the 2 drugs together (median, 179 days) was intriguing. In an earlier study, dogs receiving cisplatin alone followed by piroxicam alone had a median survival of 309 days, compared with 246 days in dogs receiving cisplatin/piroxicam.[3] The favorable survival with the sequential use of cisplatin and cox inhibitors could be, at least in part, because of decreased risk for the development of resistance to both drugs simultaneously, as well as avoiding overlapping toxicoses.

As expected, cisplatin and cisplatin/firocoxib were associated with GI, bone marrow, and renal toxicosis. In most instances, the toxicoses with cisplatin/firocoxib were no worse than those with single agent cisplatin. There were no significant differences among the groups in GI toxicosis, weight loss, bone marrow suppression, or increase in serum creatinine concentration. Although not significantly different, the decrease in iohexol clearance appeared slightly greater in dogs receiving the combination. It is likely that in dogs receiving the combination treatment, the antitumor effects would be even greater if cisplatin could be continued longer. With the combination treatment, 71% of dogs stopped receiving cisplatin because of toxicoses before cancer progression. Continued efforts to identify ways to decrease toxicoses of cisplatin/cox inhibitor therapy are indicated. Decreasing cisplatin dosage or substituting carboplatin for cisplatin could be considered, but an approach that could have more impact would be to employ emerging agents that could provide some protection from cisplatin toxicoses to the kidneys and bone marrow.

This study also was important in demonstrating antitumor activity of firocoxib as a single agent, indicating that firocoxib can be used palliatively in dogs with TCC. Partial remission occurred in 20% and SD in 33% of dogs receiving single agent firocoxib. This remission rate compares favorably to that observed with piroxicam. Although CRs have been observed with piroxicam,[1] CR did not occur with firocoxib treatment in this study, nor have CRs been reported with other cox-2 inhibitors in dogs with TCC.[22] It is not known if the finding of CRs in dogs treated with piroxicam is because of (1) having larger numbers of dogs treated with the drug and therefore a better chance to observe CR, (2) the longer half-life of piroxicam (40 hours) providing constant cox inhibition, (3) the inhibition of both cox-1 and cox-2 by piroxicam, or (4) other mechanisms. Although survival for dogs receiving firocoxib alone (median, 105 days) appeared shorter than survival reported with the nonselective cox inhibitor, piroxicam (195 days),[1] this may be because of the more advanced stage cancer in dogs receiving firocoxib. Dogs receiving firocoxib had metastases in 53% of cases, urethral involvement of the TCC in 73% of cases, and prostate involvement in 80% of male dogs. The presence of TCC in these 3 areas appears higher than that reported for dogs in other studies,[1] and TCC in any of these 3 areas has been associated with a worse prognosis.[23] Firocoxib generally was well-tolerated. Gastrointestinal signs were noted in 5 dogs receiving firocoxib, but the likely cause of the GI upset was extensive metastases in 2 dogs, and renal failure because of cancer progression in another dog.

Another positive finding of the study was that QOL was maintained or improved in most dogs in the study. Although there were no significant differences among the treatment groups, the results related to urination and activity level in dogs receiving firocoxib alone or cisplatin/firocoxib were notable. When combining the scores across categories in dogs with follow-up available, 67% of dogs receiving firocoxib alone and 91% of dogs receiving the combination had a higher QOL score after 6 weeks of treatment than they had before treatment.

In conclusion, firocoxib had antitumor effects as a single agent and enhanced the activity of cisplatin resulting in remission in 57% of dogs with TCC receiving the combination treatment. Toxicoses inherent to cisplatin will influence the decision of whether to pursue this treatment in a given dog. Careful monitoring of renal function and other evidence of toxicoses in dogs receiving cisplatin/firocoxib is crucial. When using cisplatin/cox inhibitor treatment, decrease in cisplatin dose or withdrawal of the drug likely will become necessary in most dogs over the course of treatment.

Acknowledgments

The authors thank Drs Kim Selting, Nellie Owen, Dudley McCaw, and Jim Lattimer at the University of Missouri, and clinicians and staff in the Purdue Comparative Oncology Program at Purdue University for their case management and data entry related to this clinical trial. We also thank Dr Wilson Rumbeiha at the Michigan State University Diagnostic Laboratory for assisting with iohexol clearance assays, and Dr Diane Larsen at Merial Limited for manuscript and data review.

Conflict of Interest Declaration: Drs Hanson and Alva are employees of Merial Limited, Duluth, GA. Merial Limited provided funding for the study. Merial Limited manufactures and sells firocoxib.

Footnotes

1. 1

   Iohexol measurement performed at Michigan State University Diagnostic Laboratory, East Lansing, MI

2. 2

   PLAN procedure, SAS^® v 8.2a, Sas Institute Inc, Cary, NC

3. 3

   Platinol^®, Bristol-Myers Squibb, Princeton, NJ

4. 4

   Torbugesic^®, Fort Dodge Laboratories, Fort Dodge, IA

5. 5

   Zofran, Glaxo Smith Kline^®, Research Triangle Park, NC

6. 6

   Previcox^®, Merial Limited, Duluth, GA

7. 7

   Deramaxx^®, Novartis Animal Health, Greensboro, NC