Introduction

Cyclophosphamide is an alkylating cytotoxic drug commonly used in veterinary and human oncology and is available in an oral formulation. Cyclophosphamide requires oxidation in the liver, followed by further metabolism within the cells, to ultimately form the active metabolite phosphoramide mustard and acrolein.¹ At standard veterinary doses (200–250 mg/m²) as well as at higher doses (500–650 mg/m²) used prior to autologous bone marrow transplant (BMT) for dogs with lymphoma, cyclophosphamide is generally well tolerated, with myelosuppression, gastro-intestinal signs, and sterile hemorrhagic cystitis being the most commonly encountered adverse effects.^2,3 Accidental overdose of cytotoxic drugs is uncommonly reported; inadvertent cyclophosphamide administrations with a fatal outcome in one dog and survival in two dogs have been reported.^4,5

Case Report

A 6 yr old female spayed, mixed-breed canine weighing 20.7 kg (45.6 lb, 0.76 m²) was referred to Perth Veterinary Specialists with a 9-day history of severe, nonregenerative, cyclophosphamide-induced myelosuppression and sterile hemorrhagic cystitis.

Four weeks prior to referral, the dog's owner inadvertently obtained and administered cyclophosphamide, following a recommendation of cyclosporine for ventral abdominal pruritus and pododermatitis. Hemorrhagic cystitis was noted by the owner, and at the end of the third wk of treatment, the dog was presented to the referring veterinarian. Cyclophosphamide was administered at 2.5 mg/kg q 24 hr for the first wk, and 5 mg/kg q 24 hr for the second and third wk, equalling a total dose of 1,750 mg (equivalent to 2,303 mg/m² and dose intensity of 767 mg/m² per wk). Upon discovery of the mistake, a complete blood count (CBC) and an urinalysis were performed (day 0). Although not anemic, the dog had developed a grade 5 neutropenia and thrombocytopenia (Table 1). Hematuria (>200 cells/μL) and pyuria (<200 cells/μL) were detected on urinalysis, but there was no bacterial growth on culture. Cyclophosphamide was stopped, and the administration of broad spectrum antimicrobials (18.75 mg/kg amoxicillin/clavulanic^a acid orally q 12 hr and 7.5 mg/kg enrofloxacin^b orally q 24 hr) commenced. Seven days later (day +7), a repeat CBC showed worsening pancytopenia with no evidence of regeneration of any cell line (Table 1). The dog was then referred to Perth Veterinary Specialists for further evaluation.

TABLE 1  Serial Complete Blood Count Results During Episode of Cyclophosphamide Toxicity

View Fullscreen

On presentation (day +9), the dog was alert and responsive but quiet. Physical examination abnormalities included white mucous membranes, tachycardia, and a grade 2/6 left-sided systolic murmur. There was no evidence of mucosal or cutaneous hemorrhage. Pulmonary sounds and abdominal palpation were unremarkable, and the dog was afebrile (38.4°C). A repeat packed cell volume (PCV) showed a progressive anemia. Bone marrow biopsy and aspiration cytology were offered to determine if progenitor cells were present, indicating a potential for bone marrow and peripheral blood count recovery, but were declined by the owner in favor of commencing supportive care.

Blood typing was performed, and the dog received 350 mL dog erythrocyte antigen (DEA) 1.1 negative fresh whole blood via a temporary cephalic catheter. After the transfusion, the PCV had increased to 16%. Filgrastim^c (5 ug/kg q 24 hr) was administered subcutaneously, and tranexamic acid (TXA)^d (12.5 mg/kg q 8 hr) commenced; administration of broad spectrum antimicrobials continued. Upon the owner's request, the dog was discharged, and ongoing monitoring was continued daily as an outpatient.

On day +10, the dog remained clinically stable and afebrile; however, large amounts of hemorrhagic urine was still being produced. On day +11, +12, and +13, the manual peripheral PCV remained stable (Table 1), and the patient remained afebrile. The cardiac murmur was no longer audible on day +11. Although improved, macroscopic hematuria had persisted and, without veterinary opinion, the owner increased the dose of tranexamic acid to 25 mg/kg q 8 hr from day +12.

A repeat CBC on day +14 showed a severe but relatively stable anemia (hematocrit 15%); however, reticulocytes, nucleated red blood cells, and band neutrophils were detected (Table 1), indicating returning bone marrow function. Because of the normalization of the neutrophil count (4.46 × 10³ cells/μL), filgrastim and antibiotics were stopped (Table 2).

TABLE 2  Clinical Adverse Effects of Cyclophosphamide Overdose

View Fullscreen

A second fresh DEA 1.1 negative whole blood transfusion was required on day +15 due to clinically symptomatic anemia (PCV 12%), but, for the first time since presentation, macroscopic hematuria was no longer present (Table 2).

On day +21, the peripheral blood counts were near normal, indicating adequate bone marrow function (Table 1). The TXA dose was reduced (12.5 mg/kg q 8 hr); however, macroscopic hematuria recurred on day +24. The owner declined repeat urinalysis and sonographic evaluation of the bladder, and the dose of TXA was increased (25 mg/kg q 8 hr). Macroscopic hematuria resolved on day +29.

On day +51, the dog re-presented to the referring veterinarian for mucosal petechiation and recurrent macroscopic hematuria, despite continuing to receive TXA (25 mg/kg q 8 hr). Severe thrombocytopenia, confirmed by smear examination, was the sole abnormality on repeat CBC (Table 1), and the dog was immediately transferred back to Perth Veterinary Specialists for evaluation. Physical examination was unremarkable, apart from the previously detected mucosal petechiation. The owner declined further investigation, and the presumptive diagnosis of secondary immune-mediated thrombocytopenia (ITP) was made (Table 2). Vincristine^e (0.02 mg/kg IV), prednisolone^f (2 mg/kg orally q 24 hr), and melatonin^g (0.15 mg/kg orally q 12 hr) was administrated. Five days later (day +56), the manual platelet count had increased to 50–60 × 10³ cells/μL (Table 1). Prednisolone and melatonin were continued at the same dose. Macroscopic hematuria resolved again on day +72.

Thirty-two days after the onset of presumed ITP (day +83), the platelet count had returned to within the reference interval (Table 1). Melatonin and TXA were stopped, and over the following 6 wk, the platelet count remained within an acceptable interval (day +102), despite tapering and eventual withdrawal of prednisolone on day +125 (1.5 mg/kg q 24 hr, 1 mg/kg q 24 hr, 1 mg/kg q 48 hr). No hematological abnormalities or recurrence of macroscopic hematuria were observed in the months (day +137 and +150) following cessation of the prednisolone, melatonin, and TXA.

At the time of writing, the dog was clinically well and has not required any additional treatment for cyclophosphamide overdose. The last CBC performed on day +230 showed no abnormalities (Table 1).

Discussion

Cyclophosphamide is referred to as “stem cell sparing.”⁶ This is mainly because of high levels of aldehyde dehydrogenase expressed by hematopoietic stem cells (HSC), which appears to be the main mechanism of cyclophosphamide detoxification.⁷ It is for this reason that high-dose cyclophosphamide is myelosuppressive but not myeloablative.⁸ The ability to spare HSC, however, does not appear to be absolute, as multiple administrations of stem-cell-sparing drugs have been shown to reduce the repopulating ability of the marrow.⁶ Although the action of cyclophosphamide is not cell-cycle specific, there appears to be greater risk for cumulative damage to the HSC depending on the timing of the subsequent doses.⁶

After high-dose cyclophosphamide, the ensuing period of pancytopenia is relatively short, with hematopoietic recovery occurring in 10 to 17 days and, despite the risk, generally passes without significant morbidity or mortality.⁸ Assuming bone marrow function was normal prior to the administration of cyclophosphamide, supportive care alone should be sufficient for recovery, which is what has been reported in people and in dogs.^9,5 However, complications in bone marrow recovery have been reported to occur in people, in which case allogeneic bone marrow or HSC transplantation are pursued to restore bone marrow function.^1,10 In the case described here, finding a dog leukocyte antigen-identical donor (parent, sibling, or offspring) was considered impossible, given the dog had been rescued from a shelter with no known history.

The administration of recombinant human granulocyte colony-stimulating factor (rhG-CSF) has been used in dogs to stimulate recovery from neutropenia prior to autologous BMT, but prolonged administration of 3 wk or more in dogs has been reported to result in the development of neutralizing antibodies to rhG-CSF and endogenous canine granulocyte colony-stimulating factor, leading to severe, persistent neutropenia.^2,3,11 In one study, no dog developed persistent neutropenia after administration of rhG-CSF at 5 ug/kg q 12 hr for 7 days, suggesting that the induction of neutralizing antibodies is unlikely to occur in this period.² As it was considered possible that this dog's bone marrow function could recover without stimulation, to minimize the risk of developing neutralizing antibodies, 5 ug/kg rhG-CSF was administered q 24 hr and was stopped once peripheral blood count recovery was evident.

Two dogs were recently reported to have survived after inadvertently receiving cumulative cyclophosphamide doses of 1,078 mg/m² and 1,080 mg/m² over 9 and 11 days, respectively.⁵ One dog was managed as an outpatient with broad spectrum prophylactic antibiotics and granulocyte colony-stimulating factor (G-CSF) (the source was not specified). The other dog was hospitalized due to suspected neutropenic sepsis and received intravenous fluid and electrolyte support, broad spectrum antibiotics, G-CSF, as well as platelet and packed red blood cell transfusions. One other case report documented a fatal outcome in a dog after inadvertent cyclophosphamide administration.⁴ In that case, the dog succumbed to neutropenic sepsis and hemorrhagic complications after receiving a cumulative cyclophosphamide dose of 1,521 mg/m² over 21 days. The approach to treatment was similar in all of these reported cases and in the case presented here; therefore, factors such as total cumulative dose and intrinsic differences in cyclophosphamide metabolism and bone marrow recovery most likely contributed to the differing outcomes. The lack of neutropenic sepsis in this case was also most likely a principal factor associated with survival. Eighty percent of identified infections are caused by opportunistic endogenous flora; however, in this case, avoiding prolonged hospitalization reduced the risk of acquiring a nosocomial infection.¹

Sterile hemorrhagic cystitis (SHC) is a well-recognized adverse effect of cyclophosphamide.¹ The inflammatory effect of prolonged contact of acrolein on the bladder mucosa leads to submucosal edema, hemorrhage, necrosis, and fibrosis.¹ SHC is characterized by lower urinary tract signs of hematuria, stranguria, and pollakiuria and is diagnosed by exclusion of urinary tract infection, neoplasia, and urolithiasis, with supportive cytological, histopathological, and/or sonographic or cystoscopic appearance.¹ Bacterial infection was excluded by negative culture; however, no further diagnostic tests were performed to support the diagnosis of SHC. It is possible that the urinary tract hemorrhage seen at initial presentation was solely due to severe thrombocytopenia; however, the persistent macroscopic hematuria in the presence of adequate numbers of platelets for primary hemostasis seen on day +21 was supportive of SHC being present in this individual. One study reported that previous or preexisting immune-mediated disease was a predictive factor for the development of SHC, possibly due to exacerbation of the toxic effects of acrolein on the bladder mucosa.¹² This dog's suspected underlying allergic dermatitis may have acted as a predisposing and/or exacerbating factor to the development and severity of SHC.

In this case, TXA, an antifibrinolytic agent widely used to limit bleeding in human victims of trauma and surgical patients, was administered to reduce the hemorrhage from the urinary tract. TXA binds with plasminogen and prevents its conversion to plasmin, thus limiting the fibrinolytic pathway. In humans, at recommended doses, TXA is effective with negligible side effects, however, is contra-indicated in patients with thromboembolic disease, with inadvertent high doses being associated with acute arterial or deep vein thrombosis.¹³ Within the veterinary literature, there is limited information regarding the use of TXA. One study reported that apart from inhibition of emboli resolution, no other adverse effects were observed in 37 dogs administered TXA at doses up to 110 mg/kg q 12 hr for 30 to 40 days.¹⁴ Recently, another study found the recommended doses for TXA in horses was up to 20 times higher than necessary to inhibit fibrinolysis.¹⁵ The dosing scheme used for horses and this case was based on in vitro data for healthy human donors and may, therefore, represent a gross overdose in nonhuman species or in patients with procoagulation disorders. There were no adverse effects associated with the use of TXA in this case; however, in the absence of phase I studies or regular monitoring with thromboelastography, the potential for serious or even fatal complications should be considered prior to use in dogs with uncontrollable hemorrhage, including severe SHC.

Patients with SHC suffer significant discomfort, but, in the absence of a concurrent coagulopathy, the condition is not considered to be life threatening. Urinary tract hemorrhage is a potentially fatal complication in people with severe thrombocytopenia.¹⁶ In this case, the SHC at initial presentation was most likely exacerbated by concurrent severe thrombocytopenia, leading to significant external hemorrhage via the urinary tract. Fresh whole blood was an appropriate choice because of the clinical and clinicopathological evidence of external hemorrhage. The requirement to raise platelet levels was not considered to be as critical as correcting the anemia, as the risk of bleeding with thrombocytopenia is somewhat affected by the degree of anemia.¹⁷ Blood typing, but not cross-matching, was performed prior to both transfusions, which is not ideal when the requirement for multiple transfusions is possible. However, as high-dose cyclophosphamide is capable of inhibiting humoral and cell-mediated immune responses, and has been shown to reduce the risk of graft-versus-host disease, the risk of a transfusion reaction may have been reduced in this dog.¹⁸

The subsequent development of a second episode of clinical grade V thrombocytopenia on day +51 was not expected and has not been reported as a sequela of chemotherapeutic overdose in humans. In one study, two dogs developed mild (grade 1 and 2), long-term thrombocytopenia after receiving 500 mg/m² cyclophosphamide prior to autologous bone marrow transplant.³ One of these dogs had a positive anti-platelet antibody titer, suggesting an immune-mediated etiology. The development of presumed ITP in these two cases, and in the case presented here, may have been associated with the administration of blood products or antibiotics or due to severe immunosuppression of T regulatory lymphocytes (Treg). Tregs are important for suppressing the immune responses of other cells and when depleted may allow for the development of autoimmunity. Continuous low-dose administration of cyclophosphamide over 28 days has been shown to significantly reduce pathologically elevated numbers of Tregs in dogs with soft tissue sarcomas; however, whether the same effect occurs after a single high-dose bolus, or high-dose accumulation over less than 28 days, or in noncancer-bearing dogs has not been established.¹⁹

Based on the severity of thrombocytopenia (<10 × 10³ cells/μL) in the absence of other cytopenias or clinical evidence of consumption, the second episode of thrombocytopenia in this case was presumed to be immune mediated. Unfortunately, confirmation by antiplatelet antibody titers and/or assessment of megakaryopoiesis by bone marrow biopsy were not performed; however, the relatively rapid improvement of platelet numbers with immunosuppressive doses of prednisolone was supportive of an immune-mediated etiology. Of the two previously reported dogs that developed mild thrombocytopenia, one recovered spontaneously, and the other recovered with intermittent prednisolone treatment.³ Of the other two dogs that survived cyclophosphamide intoxication, neither were reported to develop a second episode of thrombocytopenia after initial recovery; however, the follow-up time of these dogs (+25 days and +34 days) may have been inadequate to detect this phenomenon.⁶

Melatonin stimulates thrombopoiesis and has been successful in the treatment of people with refractory idiopathic thrombocytopenic purpura.²⁰ To the best of the author's knowledge, there is no evidence to support the use of melatonin in canine ITP; however, due to the severity of the thrombocytopenia and safety at similar doses used for the treatment of dogs with alopecia, treatment was commenced. In this case, it is not clear what impact, if any, melatonin had on resolution of ITP.

One pertinent aspect to the management of this case was the exhaustion of available resources. Fresh whole blood, or other blood products, were not freely available, and a new donor had to be found for each blood transfusion. It is difficult to be certain when a patient's peripheral blood count will recover; therefore, at initial presentation, there is no indication of how long supportive care will be required. Recovery may take months, and pet owners need to be aware that the limited availability of blood products and resources may preclude long-term supportive care.

Conclusion

Even at high cumulative doses of cyclophosphamide, individual dogs may be successfully managed with supportive care. In addition to prophylactic broad-spectrum antibiotics and blood products, rhG-CSF and TXA may be effective and safe for use in dogs with drug-induced myelosuppression and hemorrhage.