Corresponding author: A. S. Moore, Veterinary Oncology Consultants, 379 Lake Innes Drive, Wauchope 2446 NSW, Australia; e-mail: firstname.lastname@example.org.
Background: Lymphoma is the most common malignancy affecting cats. A protocol employing vincristine, l-asparaginase, cyclophosphamide, doxorubicin, and prednisone (VELCAP-S) is effective and well tolerated in dogs with lymphoma. A 24-week variation of this protocol (VELCAP-C) was developed for treatment of cats.
Hypothesis: That VELCAP-C will result in survival times for cats with lymphoma that are similar to those obtained when cats are treated with a protocol that includes fewer chemotherapy agents.
Animals: Sixty-one cats with lymphoma.
Methods: Retrospective study. Outcomes evaluated were response to VELCAP-C therapy, toxicosis, and survival time. The effect of signalment, staging, CBC, and serum chemistry profile and dosage on these outcomes was examined.
Results: Six cats (10%) completed the protocol with a median survival of 1189 days. Forty-three percent (23 of 61) of the cats achieved complete response (CR) with a median survival time of 62 days. Cats that required a dose reduction of any drug during induction were more likely to achieve CR. Weight loss and hepatomegaly at diagnosis were negatively associated with response to treatment. Increased lactate dehydrogenase (LDH) serum activity at the time of initial treatment correlated with decreased survival times.
Conclusions and Clinical Importance: This multi agent protocol did not provide improved survival over historical data using protocols with fewer agents. Serum LDH activity levels might provide useful prognostic information for cats with lymphoma.
Lymphoma is a malignant disorder arising from the lymphoid organs and as of 1974 accounted for one-third of all feline tumors.1 Few useful prognostic factors in cats with lymphoma have been identified in the veterinary literature. Cats that are clinically ill (substage b),2,3 have advanced disease, including multiple organ involvement,4 and fail to achieve complete remission2,3,5 and cats that are FeLV positive have a poorer prognosis.2,4 Staging, as done routinely for dogs and human patients, is not consistently prognostic in cats.6 There is no association between anatomic site and prognosis.7 Achieving complete remission (CR) is a positive prognostic indicator for both remission duration and survival.6
Combination chemotherapy protocols that include l-asparaginase, doxorubicin, cyclophosphamide, vincristine, and prednisone are effective in the treatment of lymphoma in dogs.8–12 A similar protocol (VELCAP-S) to the one to being evaluated in this study has been used in dogs to acheive complete remission rates of 68% and an overall remission duration of 10 months.13
Although there are multiple studies to indicate the effectiveness of a multidrug approach in dogs, there are very few studies to evaluate their efficacy in cats. The addition of l-asparaginase with or without doxorubicin to the traditional COP protocol provided longer survival times in cats with alimentary lymphoma when compared cats treated with COP alone or single agent therapy.3,14–17 In contrast to the recent reports from the United States, a Netherlands study reported a median survival time of 266 days and response rates of greater than 75% in cats treated with the COP protocol, figures that appear similar to earlier reports from the United States despite a relatively low occurrence of FeLV in the Netherlands.5 The Netherlands population had a low prevalence of gastrointestinal involvement compared with more recent studies from the United States.5 The intent of this study was to evaluate the VELCAP-C protocol for efficacy and toxicosis and to identify prognostic factors associated with outcomes.
Methods and Materials
Cats receiving the VELCAP-C protocol for lymphoma between June 1994 and August 2001 at the Harrington Oncology Program at Tufts Cummings School of Veterinary Medicine and an associated private practice, Ashland Animal Care Center, and that had available medical records qualified for this retrospective study. Any cats receiving chemotherapy for lymphoma before starting VELCAP-C were excluded from evaluation. Inclusion in the study was by intent to treat. In addition, a group of cats was evaluated post hoc based on the treatment received or the portion of the protocol completed.
The VELCAP-C protocol is based on a similar protocol used to treat dogs with lymphoma (VELCAP-S)13 with additional doxorubicin treatments after induction (Table 1). Cyclophosphamide and doxorubicin were administered if the neutrophil count was above 3,000/μL on the day of treatment. Doses of doxorubicin or cyclophosphamide were reduced by 25% of the original dose if the neutrophil count 7 days after administration was <1,000/μL. All veterinarians involved in treating the cats used the same criteria for dosage reductions with regard to neutrophil count. Other dose reductions were made at the discretion of the treating veterinarian if toxicoses such as vomiting or anorexia were clinically important.
Table 1. VELCAP-C protocol for cats with lymphoma.
Prednisone (Pred.)=40 mg/m2 PO q24h for 7 days then every other day. Doxorubicin (DOXO)=25 mg/m2 IV; vincristine (VCR)=0.75 mg/m2 IV except on week 2 given at 0.5 mg/m2 IV; cyclophosphamide (CTX)=300 mg/m2 PO; l-asparaginase (l-ASP)=10,000 u/m2 IM.
Response to therapy was defined as 1 of 2 categories: complete response (CR) and no response (NR). CR was defined as a complete resolution of clinical disease based on a physical examination and other noninvasive diagnostic tools such as radiographs or ultrasonography. No response was defined as any response less than a CR. The cats in the study were evaluated for response each time they were examined by a veterinarian for either chemotherapy or laboratory tests. Survival time was defined as the time from 1st treatment until the animal died or was censored. Cats were censored from statistical evaluation of survival when they were lost to follow-up or the protocol they were receiving changed. Cats that were still alive at the time of analysis had their survival times censored at the last known date of follow-up.
Diagnosis and Pre-Treatment Evaluation
The diagnosis of lymphoma in all cats was made by either cytologic or histopathologic evaluation of tissue or malignant effusion. All cats were assigned categories according to the scheme shown in Table 2. Substage categories “a” or “b” were applied to each cat based on absence or presence of clinical signs of disease at presentation. Cats with clinical signs including dyspnea, vomiting, diarrhea, anorexia or weight loss were included in the substage b category.
Table 2. Anatomic distribution of lymphoma in 61 cats.7
Number of Cases
Any structures within the mediastinal space
Any structure within the abdominal cavity
Gastrointestinal tract and associated lymph nodes, excluding liver and pancreas
Either or both kidneys
Organs such as the liver or spleen in the absence of gastrointestinal or renal involvement
Combination(s) of the above 3 subcategories
One or more peripheral lymph nodes
One peripheral lymph node only
A chain of adjacent lymph nodes in a restricted anatomical region
Many or all peripheral lymph nodes
Non lymphoid tissues such as the central nervous system, skin, larynx, nasal cavity
Combination of two or more of the above
Toxicosis was defined as any adverse effect severe enough to result in a dose reduction of any chemotherapeutic agent. These toxicoses included neutropenia (<1,000/μL at nadir or <3,000/μL at the time of treatment), anorexia, vomiting, or diarrhea. Toxicoses were determined and monitored by owner reports, physical examination findings, results of CBC, and serum chemistry profiles. Because of the retrospective nature of this study, toxicoses that did not result in dosage reduction were inconsistently reported and therefore were not evaluated.
Outcomes evaluated were response to therapy, toxicosis, and survival time. Factors including signalment, clinical signs, substage, category, subcategory, FeLV/FIV status, CBC, and serum chemistry findings were evaluated for their effect on response and survival times. The specific CBC and chemistry profile variables examined were hematocrit, neutrophil count, lymphocyte count, monocyte count, platelet count, total serum protein, and serum levels or activities of albumin, globulin, calcium, blood urea nitrogen, creatinine, alkaline phosphatase, alanine transferase, aspartate transferase, lactate dehydrogenase (LDH), potassium, sodium, chloride, phosphorus, and cholesterol. The effect on outcome of dose reduction due to toxicosis was also examined as a dichotomous variable, yes/no.
To determine the statistical influence of the factors listed above on the probability for a cat to enter complete remission, χ2 statistics were used for categorical variables: breed, sex, category/subcategory (Table 2), clinical signs, CBC, and biochemical findings (high, low or normal), substage, FeLV/FIV status, and dosage reduction, most of which were dichotomous. Independent T-test was employed to test the influence on whether a cat entered complete remission of variables that were continuous: age, CBC and biochemical findings, and weeks of protocol completed. Variables found to significantly affect remission on univariate analysis (P < .1) were analyzed by binary logistical regression to determine multivariate significance.
Kaplan-Meier survival statistics were used to determine overall survival times. Cox regression was used to determine the influences of previously mention variables on the duration of survival time. Variables that were found to be significant at the level of P < .1 in univariate analysis were entered into a forward logistic regression model to determine whether they retain significance under multivariate analysis (P < .05). Cats were censored from survival analysis at the last date of follow-up as outlined above.
Thirty-three of the patients had LDH serum activity measured before entering the protocol. Serum LDH activity were not measured in 28 cats because of a discontinuation of the measurement at the laboratory used and not because of selected factors in individual cats. Data from 2 groups of cats, based on the presence or absence of the LDH activity, were statistically compared for differences in the factors listed above by χ2 statistics for categorical data and independent T-test analysis for continuous variables. Because anorexia is possibly related to increases in serum LDH activity, χ2 analysis was performed on these 33 cats to evaluate the significance of anorexia in cats with increased LDH activity.
Sixty-one eligible records were reviewed. The median age of all cats was 10 years (range of 1–17 years). Forty-three of the 61 cats (70%) were castrated males and 18 (30%) were spayed females. Forty-five of the cats (74%) were domestic short-haired cats, 13 (21%) were domestic long-haired, and 3 (5%) were pure bred.
Diagnosis and Pretreatment Evaluation
A diagnosis of lymphoma was made by biopsy of a lymph node or extranodal mass in 42 cats, by fine needle aspirates (lymph node, spleen, soft palate mass, abdominal mass, mediastinal mass, kidney) in 16, and by cytologic evaluation of related pleural effusion in 3 cats. No cat with a biopsy diagnosis had lymphocytic (low-grade) lymphoma. Anatomical categorization and number of study cats in each category are provided in Table 2.
Fifty-eight of the 61 cats (95%) were clinically ill (substage b) at initial treatment. The most common clinical signs upon presentation were weight loss (32 cats, 52%) and anorexia (27 cats, 44%). Less common clinical signs were vomiting, diarrhea, dyspnea, and neurologic signs.
Imaging studies were not performed in all cats; 30 cats had ultrasonographic studies performed and 30 had radiographic studies performed. Organomegaly was assessed by physical examination, imaging, or both. Hepatomegaly was identified in 9 (15%) and splenomegaly in 11 (18%). Fifty-four of the cats had serum chemistry profiles performed before treatment and 53 cats had a pretreatment CBC performed. Of these profiles, 33 included LDH concentrations and 13 cats (39%) had increased LDH activity. The population of 33 cats that had LDH activity measured was found statistically to be identical for all measured variables as those 28 cats whose LDH activity were not measured (P < .05). Feline leukemia virus serology was performed on 52 cats, with 14 (27%) testing positive for the virus. FIV serology was performed on 31 cats, with 5 cats, (16%) having positive results. Bone marrow aspirates were performed in 3 cats, with 1 having myelodysplasia and 2 having neoplastic infiltration.
Thirty cats (49%) required a dosage reduction at some point during the protocol. Two cats had a dosage reduction in more than 1 chemotherapy drug. Of the 61 cats entered into the protocol, 23 (37%) did not continue treatment beyond the 2nd week because of lack of response, loss to follow-up, or death. Of the 23 cats in this group, 6 (26%) had the dose of vincristine reduced or scheduled to be reduced because of toxicosis. The most common drug causing toxicoses that led to a dosage reduction was vincristine in 20 cats. All but one of these cats had gastrointestinal signs (vomiting or anorexia); the remaining cat had the dose of vincristine reduced because of elevated liver enzymes. Eleven cats had a dose reduction of doxorubicin, 2 for anorexia, 2 for neutropenia, 1 because of its small size, 1 because of vomiting after treatment, and 3 for reasons not recorded in the medical record. The dose of cyclophosphamide was reduced in 3 cats, 1 because of anorexia and 2 for reasons not recorded in the medical record. Overall, anorexia (12 cats, 20%) vomiting (10 cats, 16%), and neutropenia (6 cats, 10%) were the most common reasons for dosage reductions for all drugs. Only the 26 cats receiving a dose reduction for a known reason were included in statistical analysis of effects on remission and survival.
Remission and Survival
Twenty-six of the 61 cats (43%) treated with the VELCAP-C protocol achieved a CR. Nine factors were found to significantly (P < .1) affect remission on χ2 or T-test (univariate) analysis (age, breed, category, dosage reduction, packed cell volume, alanine transferase, aspartate transferase, hepatomegaly, and weight loss). These factors were analyzed with binary logistical regression to determine the factors significant in a multivariate setting (P < .05). Dose reduction, hepatomegaly, and weight loss were associated with response to treatment on multivariate analysis.
Cats that received a dosage reduction of any drug at some point in the protocol were more likely to achieve CR. Eighteen of the 30 cats (60%) receiving a dose reduction achieved CR, whereas 8 of the 31 cats (26%) that did not have a dosage reduction achieved CR (P=.001).
Those with hepatomegaly at the time of diagnosis were less likely to achieve a CR. Liver size was normal in 32 cats, of which 16 (50%) cats achieved a CR whereas only 1 of 9 cats (11%) with hepatomegaly achieved a CR (P= .003); 20 cats had undetermined liver size at the time of diagnosis, 9 (45%) of which achieved complete remission.
Cats with a history of weight loss before starting chemotherapy were less likely to achieve a CR. Weight loss was reported in 32 cats of which 10 (38%) achieved CR whereas 16 (62%) of 29 cats that had maintained their weight achieved a CR (P = .043).
The median survival time for all cats was 62 days with a range of 1–2,462+ days (Fig 1). Eighteen cats were censored during survival analysis. Twelve cats (20%) were lost to follow-up and 4 (6%) changed protocols between 7 and 2,426 days after starting therapy (median 208 days). Two cats were still alive 2,352 and 2,426 days after starting treatment. Six of the cats (10%) completed the 24-week protocol with a median survival time of 1,189 days and a range of 241–2,426 days. Factors that significantly (P < .1) affected survival times on univariate analysis were response, breed, hepatomegaly, packed cell volume, lymphopenia, serum activity of alkaline phosphatase, alanine transferase, aspartate transferase, blood urea nitrogen, chloride and LDH, and dosage reduction.
When all factors found to affect prognosis on univariate analysis were subjected to multivariate analysis (P <. 05), the only factor found to affect survival was increase in serum LDH concentration before starting chemotherapy. Thirteen cats (39%) that had high serum LDH levels had a median survival time of 33 days, whereas 20 cats (61%) with normal serum LDH had a median survival time of 91 days (Fig 2).
Cats receiving the VELCAP-C protocol did not appear to have longer survival times than those treated with previously reported protocols.16,18 Those cats that received dose reductions during the protocol were more likely to achieve CR. It is possible that those cats had a reduction in dosage needed it because they were being treated at a dose that was close to their maximally tolerated dose and therefore were more likely to show a tumor response. Alternatively, cats that were treated longer may have been more likely to receive a dose reduction. Fourteen cats (23%) completed only the 1st week of the protocol after receiving only vincristine and prednisone, whereas 25 cats (41%) completed 8 or more weeks of the protocol at which time they had received all the drugs. We would have expected to see a correlation between survival and dose reduction, but this was not borne out statistically, and neither was achieving CR associated with survival on multivariate analysis, which is in contrast to other studies.3 Although a portion of cats (10%) completed the protocol and had relatively long median survival times, this evaluation is probably less helpful in counseling owners than overall survival and remission times based on “intent to treat.” Vincristine most commonly required dose reduction, possibly because it was received first in the protocol; therefore all cats received the drug. The high incidence of toxicosis to vincristine was unexpected because this was not reported in protocols using similar doses.5,6,18,19 Toxicoses can be discouraging to an owner and may lead to premature euthanasia. It is difficult to know whether discontinuation in the 1st week of the protocol represents toxicosis or whether it is a reflection of advanced disease and possibly inadequate supportive care in the initial phase of induction. Most cats were sick (substage b) at the beginning of the protocol and therefore may have been more likely to show signs interpreted as drug toxicosis. More intensive supportive care at the beginning of the protocol may help support patients during toxicosis, perhaps limiting the number of cats euthanized prematurely. Based on our finding, it might be useful to continue treating cats with supportive care and a dosage reduction rather than stopping therapy. Alternatively, consideration could be given to increasing dosage levels of chemotherapy in cats that do not show toxicosis to treatment.
Twenty-five (41%) of the cats had alimentary lymphoma, and these cats are often more debilitated at the time of diagnosis with clinical signs such as weight loss and anorexia. Cats that had lost weight at the time of diagnosis were less likely to achieve CR. Although not significant in this study, loss of appetite is negatively prognostic in dogs12,13 and might be responsible for the weight loss seen before starting therapy in the cats in this study. This finding underscores the importance of providing supportive care, including nutritional support, for cats that are being treated for lymphoma. Hepatomegaly was also a negative prognostic indicator for response to therapy. Unfortunately, the detection of the hepatomegaly was subjective and varied, and the cause of the hepatomegaly was not documented, so the significance for this finding cannot be ascertained.
Increase in serum LDH activity on initial chemistry profile was a negative prognostic indicator for survival. Given the extensive distribution of LDH in tissues (skeletal and cardiac muscle, kidney, intestine, liver, lung pancreas, and bone), it is difficult to determine the source of an increase in serum LDH.20 High LDH activities have been correlated to high tumor burden and independently predict shorter survival times in humans with non-Hodgkin's lymphoma.21,22 Unfortunately, in this retrospective study it was not possible to assess if the serum activity of LDH correlated with tumor mass. LDH is nonspecific and can be increased for many reasons, including tissue hypoxia and anorexia.23 However, in this study, anorexic cats did not have a statistical difference in their LDH concentration when compared with cats without reported anorexia, nor did anorexia influence survival. Serum LDH levels may be an independent predictor of survival in cats as they are in humans.
There are inherent limitations given the retrospective design of the study. Not all the cats in the study underwent complete testing at the time of diagnosis, toxicoses were inconsistently recorded and so unable to be graded, and some cats were lost to follow-up. However, our identification of potential prognostic factors might assist in designing future studies.
This study was performed at Tufts University Cummings School of Veterinary Medicine, North Grafton, MA (Hadden, Cotter, Rand, Moore, Morrissey) and Ashland Animal Care Center, Ashland, MA (Davis).
This study was funded in part by NIH training grant T35 DK07635. This study was presented in part at the 23rd Annual Conference of the Veterinary Cancer Society, September 26, 2003, Madison, WI.