Introduction

Carboplatin (cis-diammine-1,1-cyclobutane decarboxylate platinum [II]) is a platinum analogue chemotherapeutic agent that is widely used in veterinary medicine for neoplasms such as osteosarcoma, genitourinary carcinomas, primary lung tumors, gastrointestinal carcinoma, melanoma, and mesothelioma.^1–6 Its primary mechanism of action is the formation of DNA adducts, leading to kinking of the DNA strand and cell death, through the formation of irreparable double strand breaks or by triggering apoptosis.⁶ In veterinary patients, carboplatin is given as an IV injection, either through a butterfly or indwelling catheter. Reported side effects in dogs primarily include myelosuppression (both leukopenia and thrombocytopenia) and gastrointestinal upset.^1,2,6

Extravasation, defined as the inadvertent leakage of a drug outside of the vein and into surrounding tissue, of certain chemotherapeutic agents is a serious and potentially urgent situation.⁷ A drug’s ability to cause local tissue injury is dependent on its potential to act as a vesicant or irritant and the mechanism of injury (DNA binding versus non-DNA binding), in addition to the concentration and volume extravasated.⁸ A vesicant is defined as a drug that is capable of causing blisters and tissue destruction at the site of injury. Extravasation of a vesicant is typically associated with evidence of thrombosis of the smaller peripheral vessels injected with the possible vesicant in the affected limb, observed as early as a few hours after the injury.⁹ Tissue damage may then ensue either from a direct impairment of blood flow or as a consequence of the same cytotoxic damage that is desired for its antineoplastic properties. The affected area typically exhibits swelling from endothelial damage and subsequent ischemic necrosis upward of ∼7 days after administration.¹⁰ An inflammatory reaction is not typically associated with extravasation reactions, except around the periphery of the necrosis.^10,11 In the late stages of the injury, dense scar tissue may form that may require surgical debridement, which may lead to a delay in additional chemotherapy treatments and possibly alteration of limb function.^9–11 The vinca alkaloids (e.g., vincristine) and the anthracyclines (e.g., doxorubicin) are universally accepted as vesicants in veterinary and human medicine, with well-described physicochemical properties and supportive literature detailing the consequences of extravasation.^1,7,12,13 Other drugs that have been implicated as vesicants, primarily in human literature, include mechlorethamine, dactinomycin, and the taxanes.^7,14 Additionally, other drugs are implicated as vesicants primarily based on the concentration and amount extravasated, including mitoxantrone, cisplatin, and oxaliplatin.⁷

Platinum (II) compounds are a group of drugs that have largely been classified as irritants in human medicine; however, some platinum analogues have been classified as vesicants in certain situations.⁷ Cisplatin is considered to be a vesicant in concentrated solutions, whereas in dilutions it is qualified as an irritant only. Carboplatin, however, is the only established platinum (II)-based drug that has not been described as toxic to tissues.⁹ When extravasated, carboplatin has been categorized as an irritant to local tissues in the human literature.^9,15 Irritants are drugs that may cause pain and swelling at the site of injection without evidence of necrosis.^8,16 To the authors’ knowledge, carboplatin has not been implicated in the veterinary literature to cause extravasation reactions, and very little substantial information is available regarding this complication in human medicine.^9,15 The locally inert nature of carboplatin is one of the reasons that carboplatin is widely used for intracavitary treatment of neoplastic pleural or peritoneal effusion in veterinary patients, and even subcutaneous (SQ) or intratumoral injections.^17,18 Although most administrations outside the vein are not accompanied with complications, there are a few examples suggesting potential adverse perivascular reactions. In Simcock et al., there was a higher incidence of local infections (41% [7 out of 17]) when a single dose of carboplatin was given as an infusion via a subcutaneously placed catheter, compared with infection rates following clean surgeries (2.5–6%).^17,19,20 Because of the presence of an indwelling SQ catheter during the administration of carboplatin, however, it is unknown how much the carboplatin itself contributed to the increased infection rate. Additionally, carboplatin has been shown to cause ulcerations when administered intradermally to mice in 100% of subjects at concentrations ≥10 mg/mL. All the ulcers that formed in these subjects had vesicles associated with them that did not rupture, and all the lesions healed within 21 days of administration.²¹ Despite these instances, there is still considered to be a lack of concrete data to support carboplatin’s vesicant potential; therefore, extravasation reactions are not a common complication discussed with owners prior to administering it in our veterinary patients.

The purpose of this study is to describe a previously unreported complication of suspected extravasation reactions associated with carboplatin. Seven cases collected from multiple institutions are described, including the disease treated, injury observed, time period of onset of complications, and outcome of each case.

Materials and Methods

A retrospective review of medical records from dogs experiencing a suspected carboplatin extravasation injury were acquired by survey of veterinary oncologists via the American College of Veterinary Internal Medicine oncology list serve and with fliers posted at a well-recognized conference (Veterinary Cancer Society; St. Louis, Missouri). Criteria for inclusion was documentation of recent carboplatin administration in a vessel associated with the site of extravasation injury for treatment of various underlying oncologic diagnoses. The following six institutions contributed cases: University of Florida Veterinary Teaching Hospital, Gainesville, Florida; Animal Specialty and Emergency Center, Los Angeles, California; University of Pennsylvania, Philadelphia, Pennsylvania; VCA Great Lakes Veterinary Specialists, Warrensville Heights, Ohio; Animerge, Raritan, New Jersey; Blue Pearl Veterinary Specialists, Atlanta, Georgia.

Information extracted from the medical record review included signalment (age, sex, neuter status, breed), disease information (histopathologic diagnosis, date of diagnosis, primary treatment performed, presence of metastatic disease, patient outcome), carboplatin information (manufacturer, concentration), and drug reaction description. Information regarding the suspected carboplatin extravasation included the date of administration, method of administration (type [IV versus butterfly catheter], location), last known catheterization of the affected leg, drug information (dose, volume), whether or not flush was administered before and/or after treatment, any dilutions or premedications that occurred prior to treatment, chemotherapy protocol, if an extravasation was noticed, number of carboplatin doses administered, and whether or not a vesicant was previously administered to the affected vein. Information regarding skin reaction included the size and description of the wound, treatment employed, duration of reaction, and any comorbidities. Statistical or prognostic information was not obtained in this study because of the small sample size.

Results

Seven dogs met inclusion criteria. The breeds represented included two greyhounds, one Leonberger, one Rhodesian ridgeback, one Australian cattle dog, and two mixed-breed dogs. Ages ranged from 4 to 13 yr of age (median 10 yr old). Four were female spayed dogs, and three were male neutered dogs. Four dogs were diagnosed with osteosarcoma (left caudal scapula, left forelimb, right femur, and right distal tibia), two were diagnosed with apocrine gland anal sac adenocarcinoma (both right anal sac), and one had a grade 2 bronchoalveolar carcinoma (left caudal lung lobe). At the time of diagnosis, no cases exhibited any evidence of gross metastatic disease. See Table 1 for case descriptions.

TABLE 1 Patient Information Including Signalment

View Fullscreen

Chemotherapy was initiated in six of the seven cases as adjuvant treatment to surgery. One patient with osteosarcoma (right distal tibia; patient 3) was initially treated palliatively with hypofractionated radiation, but the owners ultimately elected to pursue an amputation with adjuvant chemotherapy. Six cases received single-agent carboplatin. The patient with bronchoalveolar carcinoma received concurrent carboplatin and gemcitabine every 3 wk. Preparation of the skin prior to administration of carboplatin included clipping hair overlying the vein and alcohol application to the catheter site in all cases. Three cases also included chlorhexidine application in their preparation (patients 1, 6, and 7). Five cases used a butterfly catheter for administration (23 gauge for two cases, 19 gauge for one case, 21 gauge for one case, and unknown for one case), and in two cases, IV catheters were placed (20 gauge × 0.75” and 20 gauge × 1.25”) to inject the chemotherapy. The veins used included the right cephalic vein in four cases, left cephalic in one case, and left lateral saphenous in two cases. The last catheterization date of the affected leg was known in four cases, which was a median of 35 days prior to administration of carboplatin (range 28–42). No known vesicants were previously administered in the affected leg. The patient with apocrine gland anal sac adenocarcinoma (patient 5) previously received IV mitoxantrone and oral piroxicam, but the injectable chemotherapy was not administered in the affected leg. Prior to administration of carboplatin and the injection site reaction, any other chemotherapy was administered a median of 21 days prior (range 21–28 days). In five of the seven cases, a 0.9% NaCl flush (unknown volumes) was administered before and after carboplatin chemotherapy administration, and a lactated Ringer’s solution flush (unknown volumes; patients 4 and 6) was used in two of the seven cases. The patient diagnosed with bronchoalveolar carcinoma treated with a combination protocol received a 4 hr gemcitabine infusion following carboplatin administration using the same IV catheter for both drugs. The catheter in this patient was flushed with 0.9% NaCl (unknown volumes; patient 7) before and after carboplatin.

There were three different manufacturers of carboplatin in these seven cases, including Teva (three cases; patients 1, 4, and 6), Hospira (three cases; patients 2, 3, and 5), and Sagent (one case; patient 7). All patients were administered carboplatin at a 10 mg/mL concentration. The expiration date of the drug was only known for one patient (patient 1), and the drug was not expired. The median dosage of carboplatin administered was 300 mg/m² (range 250–350 mg/m²), and none of the doses were diluted prior to administration. The median total dose administered was 276 mg (range 221–450 mg). Storage conditions were known in four of seven cases, all of which were stored at room temperature. Premedication with SQ maropitant citrate was given in one patient (1 mg/kg SQ; patient 1). The median total number of carboplatin doses that had been administered at the time of the extravasation was 4 (range 3–5).

None of the patients were known to have experienced a local injection reaction immediately after the infusion, although one patient did react when the butterfly catheter was removed after administration of the flush (patient 3). No immediate treatment for extravasation was initiated in any case. Carboplatin was administered a median of 7 days prior to observation of clinical signs attributable to the suspected extravasation (range 4–15 days). Most of the patients (six out of seven) developed full-thickness necrosis around the vessel treated, although one patient only developed pitting edema (patient 3). Three of the cases that experienced full-thickness necrosis also exhibited swelling and edema in the same region, one of which was described to have waxing and waning signs for 3 days prior to the appearance of an ulceration (patient 7). See Figures 1–3 for visual representation of some of the wounds from patients described in this paper. Three dogs experienced lameness on the leg associated with the injection reaction, including the patient who only exhibited pitting edema (patients 1–3). The median longest diameter of the wound observed in those who experienced full-thickness necrosis was 3.25 cm (range 1.5–8 cm). Treatment differed based on the severity of reaction observed and was clinician dependent. The only dog who experienced pitting edema was treated with pain medications and warm packing of the affected region. Three of the six patients who experienced full-thickness necrosis received surgical debridement as part of their wound management. One of the three patients receiving surgery required three different debridements prior to wound healing (patient 1). Three patients received hyperbaric oxygen therapy (n = 1; patient 1) and/or cold laser therapy (n = 3; patients 1, 2, and 4). All patients with full-thickness necrosis and subsequent surgical debridement received wet-to-dry bandaging with various supplementary wound dressings (silver impregnated material, honey). Five of the six dogs with necrosis also received various antibiotics (amoxicillin/clavulanic acid, enrofloxacin, doxycycline) either for suspected wound infections or for prophylactic treatment, one of whom received concurrent topical silver sulfadiazine. It should be noted that wound cultures were not performed to confirm the presence or absence of infections, however. The sixth patient with necrosis received topical silver sulfadiazine 1% cream alone (patient 4). Five of the seven patients received a combination of anti-inflammatory drugs and pain medications (either carprofen or prednisolone, ± tramadol). Specific dosages were not supplied for antibiotics, anti-inflammatory drugs, or other analgesics. The wounds healed completely in six patients and partially in one patient (patient 1). The length of time required for complete wound healing was known in all six patients and took a median of 25.5 days after the local injection reaction was noticed (range 7–49 days). The patient who exhibited only partial wound healing was euthanized 74 days after the diagnosis of the injection reaction because of pulmonary metastatic disease. The wound, however, was mostly healed as it was contracted with epithelialization present and a central area of granulation tissue measuring only 5 mm at the time of death. None of the patients were known to have any comorbidities that may have affected their wound healing; however, the patient who only experienced partial wound healing was a greyhound.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Two patients received a subsequent dose of carboplatin, one through a surgically placed vascular access port placed in the jugular vein (patient 1). At the time of writing, five of the seven cases were alive. One was lost to follow-up 1 mo after the injection reaction, and one was euthanized for hemoptysis secondary to pulmonary metastatic disease 74 days after the suspected carboplatin extravasation.

Discussion

This report describes the development of a full-thickness necrosis or severe local reaction in seven dogs following administration of carboplatin. Although leakage/extravasation at the time of the injection was not directly noticed, the severity and features of the local adverse reactions were not expected to be seen for this chemotherapy agent and are unlikely to be due to phlebitis only. There is a lack of substantial evidence connecting carboplatin to vesicant properties; therefore, extravasation reactions are not typically expected with this drug within the human and veterinary oncology communities.

Factors that contribute to a drug’s potential to cause tissue damage include the drug’s chemical properties, mechanism of action, concentration, and the amount of drug extravasated.^8,16 Molecular mechanisms of action associated with a known vesicant, mechlorethamine, have previously been proposed.^14,22,23 One suggested mechanism involves cleavage of fibrils connecting the basal epidermal cell layer to the basement membrane because of direct DNA damage and subsequent inhibition of glycolysis through nicotinamide adenine dinucleotide depletion.²² The second mechanism involves the loss of protection of free radicals because of local glutathione depletion, leading to direct tissue damage ²³ Neither mechanism has been specifically linked to carboplatin.

DNA-binding agents, such as doxorubicin, are more lethal to the tissues because they cause cell death of the local tissues directly encountered in addition to the neighboring cells via endocytosis of the drug–DNA complex released from the apoptosed cells.²⁴ Carboplatin is a platinum agent that exerts its primary mechanism of cytotoxicity by covalently binding to the purine bases of DNA, forming inter- and intrastrand crosslinks and kinking of the DNA molecule.⁶ Cisplatin and oxaliplatin, two other platinum agents, have been suggested to be vesicants, in a dose- and concentration-dependent manner.^7–9,24 Carboplatin has only been previously reported as an irritant in both human and canine literature to the authors’ knowledge.^8,15,24 It is not known why cisplatin and oxaliplatin have a higher propensity to cause extravasation reactions when compared with carboplatin. It may be related to the difference in their chemical structures. Carboplatin has a complex leaving group that differs from oxaliplatin, which is associated with carboplatin’s reduced activity in aqueous solutions and subsequently decreased toxicity to the kidney, ear, and nervous tissue. The different chemical properties of these drugs affect their movement across cell membranes. Cisplatin, with its simple structure, can readily cross cell membranes against a concentration gradient as it is neutralized at a physiologic pH. This difference in movement is related to each drug’s aquation chemistry. Oxaliplatin’s aquation chemistry is reportedly similar to that of cisplatin’s. Carboplatin, however, requires esterase activity for the carbosylato-leaving group to generate the reactive species of the compound. This requirement may contribute to its potential for toxicity in perivascular spaces and lack of vesicant potential.⁶

Marnocha et al. investigated the effects of intradermal administration of carboplatin in mice and showed that concentrations ≥10 mg/mL contributed to ulcer formation, which is the concentration that is most frequently used in practice.²¹ However, several reports suggest that mice may be more sensitive to extravasation of carboplatin compared with humans (and possibly dogs) based on species differentiation.^21,25 The reason for the lack of localized reactions in clinical practice, however, may be related to the location of injection. In Maroncha et al., one of the mice receiving 15 mg/mL of carboplatin was excluded from the study because of a suspected SQ injection as opposed to an intended intradermal injection. All of the mice receiving proper intradermal injections developed ulcers, whereas the mouse with the SQ injection developed only an initial bleb at the time of the injection without any observable tissue damage.²¹ Based on various reports in accordance with the findings in Maroncha et al., it appears that carboplatin is a safe drug to administer in sites outside of a vessel and perivascular tissue.^{8,15,17,18,24} Carboplatin has been frequently used in dogs for intracavitary, SQ, and intralesional injections without previously reported deleterious adverse events.^17,18 It should be noted, however, that in the study by Simcock et al., 41% of dogs receiving a slow SQ infusion of carboplatin delivered over 3–7 days following treatment of osteosarcoma via amputation or limb spare procedures developed an incisional infection at the site of injection.¹⁷ Most of these infections were uncomplicated, with the exception of those who received a limb spare, and resolved with various antibiotic therapies. Local irritation by the potential vesicant potential of carboplatin should be considered as a possible contributing factor to the higher rate of incisional infections observed in this study.

None of the patients were known to have comorbidities that would have affected their skin and healing ability. It is interesting to note that two of the patients were greyhounds, a breed that may have impaired wound healing capabilities because of impaired clot strength and bleeding tendencies.^26,27 Five of the patients were administered carboplatin with a butterfly catheter instead of a flexible indwelling catheter. Butterfly catheters are rigid and not as forgiving of small movements, both of which increase the risk of puncture through the other side of the vessel. In at least one of the patients with complete documentation, consecutive doses of carboplatin were administered in the affected leg (left lateral saphenous vein, patient 1; Table 1). This patient may have been recovering from phlebitis from their first catheterization, thus increasing the risk of complications. This was also a location with minimal soft tissue coverage, both of which are risk factors for extravasation.^8,10 Additionally, patient 7 received a 4 hr infusion of gemcitabine following the administration of carboplatin using the same catheter. It should be noted that gemcitabine has not been previously described as a vesicant, so it is not suspected to have contributed to the development of the wound in this patient; however, gemcitabine should not be completely eliminated as a contributing factor for a suspected extravasation reaction in this patient.

This report raises the concern for carboplatin to be able to cause severe injection reactions, and further investigation into the identification of risk factors for our patients is warranted. Most extravasations can be prevented with the systematic implementation of careful, standardized, and evidence-based administration techniques (e.g., the use of healthy veins with adequate soft tissue coverage, avoidance of administration of a vesicant prior to chemotherapy, use of a flexible indwelling catheter, rotation of limbs in between chemotherapy treatments, use of experienced technicians, use of sedation when needed, and avoidance of fluid pumps to allow prompt detection of resistance).^7,24 Early recognition of extravasation is key in the management of its potential consequences. In this case series, extravasation was not detected/recognized at the time of injection. Clinical signs attributable to extravasation documented in human medicine often include burning during the infusion as an initial symptom, followed by pain, erythema, pruritus, and swelling. Eventually, these signs can progress to infection, ulceration, severe necrosis, and tissue destruction.^7,8,11,16 In human oncology, extravasation must be suspected if any of these specific signs or symptoms are present. Only one patient reacted with a painful response at the end of the postsaline infusion, but this was the only patient who did not develop necrosis as a result of suspected extravasation. The remaining six patients developed full-thickness necrosis at the site of injection ranging between 4 and 15 days after administration. Four of the seven patients exhibited soft tissue swelling and edema that caused varying degrees of limb dysfunction or pain.

No known antidotes to carboplatin extravasation exist, although 4 out of 7 of the patients’ wounds healed completely with supportive care alone. Treatments previously recommended for extravasation reactions are largely based upon anecdotal recommendations and case reports.^8,12,16 Some recommendations include warm or cold compresses based on the mechanism of tissue injury, dimethyl sulfoxide to promote vasodilation and free radical scavenging, hyaluronidase to facilitate detachment of the drug from the damaged tissue, sodium thiosulfate for neutralization of the extravasated drug (if applicable), surgical debridement, and possibly steroids for their anti-inflammatory properties. The utility of these treatments, however, is highly dependent upon the type of drug extravasated and the mechanism of extravasation injury. Patients in this study received a variety of supportive care treatments that resulted in complete healing in all cases, excluding the patient censored for death prior to complete wound healing, a median of 25.5 days after initial observation of the wound. Because of the small sample size and variability of cases included, it is not possible to determine the effectiveness of an individual treatment option in this case series.

Conclusion

There are several weaknesses and limitations to this case series. There are a small number of cases represented with a varying spectrum of diagnoses and treatments performed. Additionally, the retrospective nature of this study prohibits acquiring complete medical records, and the lack of histopathologic confirmation supporting the extravasation reaction weakens the suspicion of carboplatin’s vesicant potential as the underlying cause of the wounds observed.

This case series highlights the possibility that carboplatin may behave as a vesicant in some patients, although true risk factors or predispositions have not been identified. Based on this case series, the authors would recommend that any reaction following carboplatin injection should be documented closely with serial size measurement and photographs in addition to prompt and prudent management of the wound bed similar to those caused by known perivascular irritants or chemical burns. To fully characterize the vesicant potential for carboplatin, a larger case series should be reported that includes more detailed information about administration to identify risk factors, in addition to histopathological confirmation to support the suspicions of a true extravasation reaction.