Adverse Interaction Between Colchicine and Ketoconazole in a Chinese Shar Pei

Introduction

Multidrug treatment regimens increase the risk for adverse drug interactions. Absorption of many oral drugs involves efflux and metabolism in enterocytes and hepatocytes modulated by drug efflux transporters (e.g., multiple drug resistance [MDR]-associated proteins) and cytochrome P450 metabolism. Cytochrome P450 is the most versatile enzyme system involved with xenobiotic metabolism being responsible for the oxidative metabolism of a wide variety of compounds.¹ Either inhibition or induction of that enzyme system can importantly influence drug-drug interactions. Adverse drug-drug interactions often reflect overlapping substrate specificities that can complexly influence drug pharmacokinetics leading to accumulation of toxic drug concentrations or biotransformed drug adducts. Unfortunately, drug-drug interactions associated with MDR transport and/or cytochrome dependent metabolism involves so many agents that a concise list of avoidable interactions does not exist.

Colchicine, used for centuries to treat acute gout, is currently FDA approved for gout control and as a treatment of familial Mediterranean fever (FMF) to reduce the risk of systemic amyloidosis.^2–5 A similar chronic treatment strategy has been advised for familial shar pei fever (FSPF), which shares clinical features with FMF and is a risk factor for systemic amyloidosis despite different causal gene mutations.^4,6–11 Irrespective of the common use of colchicine in humans, there is limited information regarding its pharmacokinetics and the mechanism of reported adverse drug interactions.¹² Colchicine is reportedly metabolized by cytochrome P450s (CYP3A4 in humans) and is a substrate of MDR1 (variably referred to as P-glycoprotein, ABCB1-transporter, or adenosine triphosphate-binding cassette subfamily-B member-1).^5,12

Ketoconazole, an imidazole antifungal, can instigate adverse drug interactions as a result of its inhibition of certain P450 cytochromes involved with drug biometabolism as well as its inhibition of MDR1 transport.⁵

Case Report

A 2.5 yr old castrated male Chinese shar pei weighing 21.5 kg with a 2 yr history of episodic fever, lethargy, and shifting lameness was presumptively diagnosed with FSPF but had never been treated for the syndrome. Baseline health assessments included a routine hemogram and serum biochemical profile (performed using an overnight fasted blood sample), complete urinalysis, baseline total thyroxin concentration, and microagglutination titers for six Leptospira serovars. The complete blood cell count disclosed a normal erythrocyte count with microcytosis and normal morphology. Remaining hematologic, biochemical, and urinalysis parameters were within normal limits and the Leptospira serovar titers were negative (Table 1). Baseline total thyroxin was 1.7 µg/dL (reference range, 1–4). Ten months later, the dog was presented for a superficial skin condition described as a papulopustular cutaneous rash with alopecia extending over the cranial thorax. The lesion was interpreted as a pyoderma, likely complicated by a dermatomycosis. The dog was neither lame nor febrile (rectal temperature was 37.9 C), and further clinicopathological assessments were declined. Cefpodoxime^a (9 mg/kg per os [PO] q 24 hr) and ketoconazole^b (4.7 m/kg PO q 12 hr) were prescribed for 30 days by the local veterinarian.

TABLE 1 Tests Completed or Analyte Measured and Reference Range Pretreatment on Various Posttreatment Days Relevant to Initial Colchicine Administration

ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CK, creatine kinase; GGT, γ-glutamyltransferase; MCV, mean corpuscular volume; NA, not applicable; PCV, packed cell volume.

View Fullscreen

Owner concern regarding frequent historical events consistent with FSPF and its relationship with reactive secondary amyloidosis prompted a prescription for colchicine^c the following week.^6,8 Initially, colchicine (0.025 mg/kg PO) was administered q 48 hr for 1 wk then the dose frequency was increased to q 24 hr once drug tolerance was demonstrated (i.e., lack of gastrointestinal signs).⁸ Within a few days of dose frequency escalation, the dog became inappetent, began vomiting, and developed diarrhea. Musculoskeletal signs differing from prior episodes of lameness were characterized by generalized weakness, hyperpathia, inability to navigate stairs, and abnormal sitting postures seemingly to achieve comfort. All medications were discontinued on day 21 of colchicine therapy when the dog was reassessed. Physical examination disclosed vague abdominal discomfort and abnormal sitting postures. Laboratory assessments completed on an overnight fasted blood sample (Table 1) disclosed persistent erythrocyte microcytosis, absence of an expected leukogram stress response in consideration of physical pain, and biochemical evidence of myonecrosis and hepatopathy. Presumptive diagnoses included a necrotizing myopathy, acute hepatopathy, and gastrointestinal toxicity secondary to colchicine and transient cortisol suppression by ketoconazole.

The dog was discharged for in-home monitoring without additional medications and the inappetence, lethargy, vomiting, and diarrhea resolved. Physical examination 4 days later confirmed normal vital signs and resolution of muscle pain and abnormal posturing. Laboratory assessments disclosed marked reductions in aspartate aminotransferase (AST) and creatine kinase (CK) activities, progressive increases in alanine aminotransferase (ALT), alkaline phosphatase, and γ-glutamyltransferase activities along with development of mild hyperbilirubinemia, hypercholesterolemia, and persistent erythrocyte microcytosis, which led to referral to Sage Veterinary Specialty and Emergency Centers for definitive assessment of a cholestatic hepatopathy and inspection for an antecedent hepatic disease including FSPF-associated amyloid deposition that might have increased susceptibility to colchicine toxicity.

At the time of referral, the dog was quiet and alert with neither musculoskeletal hyperpathia nor abdominal pain and had soft brown stool on rectal examination. Routine bench coagulation tests were unremarkable with no bleeding tendencies noted (prothrombin time was 16 sec; reference range, 12–17 sec and activated partial thromboplastin time was 69 sec; reference range, 71–102 sec). After discussing a watch and wait treatment option in consideration of an adverse drug interaction, the owners elected to pursue laparoscopic liver biopsies because of the sequential sustained liver enzyme activity and evolving hyperbilirubinemia. Their goal was to rule out an emerging disease process as well as hepatic amyloid deposition. Liver biopsies were collected with the patient under general anesthesia. The dog was premedicated with hydromorphone^d (0.1 mg/kg IV), propofol^e (4 mg/kg IV), and diazepam^f (0.25 mg/kg IV) and general anesthesia was maintained with isoflurane^g and O₂. IV fluids^h was provided intraoperatively for fluid support (10.8 mL/hr).

On gross inspection, the liver had a normal size, color, and surface contour. Multiple biopsies (one from each lobe) were collected for histopathology and copper quantification. Liver and aspirated bile were submitted for aerobic and anaerobic bacterial culture and sensitivity. Recovery from the procedure was uneventful, with postoperative analgesia provided by a 25 μg/hr fentanyl patchⁱ and buprenorphine^j administered as needed overnight (0.01 mg/kg IV q 8 hr, total of two doses administered), and cephalexin^k given IV 22 mg/kg q 12 hr. The dog was discharged the following morning with a fentanyl patch, tramadol^l (1.8 mg/kg PO q 8–12 hr as needed), ursodiol^m (6.7 mg/kg PO q 12 hr), s′adenosylmethionineⁿ, and ciprofloxacin^o (10.1 mg/kg PO q 12 hr) pending culture results and histopathological findings.

Tissue copper concentration was within normal limits (238 ppm; reference range, 120–400 ppm)^p, and anaerobic and aerobic bacterial cultures were negative. Histologically, ring mitoses (Figure 1) characterized by condensed chromatin marginating the nuclear membrane were identified in scattered hepatocytes with approximately 9 ring mitoses observed in 10 high-power fields (HPFs) (×400). Mitotic figures (<1/HPF [i.e., ×400]) were abnormal, with haphazard distribution of chromosomes along the metaphase plate (Figure 1). Binucleate hepatocytes were frequently observed (approximately 12/HPF) along with rare apoptotic cells. Bile canaliculi were often distended with bile and most Kupffer cells contained cytoplasmic aggregates of yellow pigment consistent with either bile or bilirubin. Small numbers of lymphocytes and plasma cells occasionally infiltrated the interstitium of the portal tract without involvement of either the hepatic parenchyma or biliary structures. No hepatocellular copper or iron accumulation was detected with rhodanine red or Prussian blue stains, respectively. However, hepatocytes often contained light cytoplasmic aggregates of granular, brown material consistent with lipofuscin indicating oxidative membrane damage. A periodic acid-Schiff stain with and without amylase (diastase) digestion confirmed moderate cytoplasmic hepatocyte glycogen accumulation and diastase-resistant hepatocellular cytoplasmic granules consistent with lipofuscin. No amyloid was detected with Congo red, no evidence of parenchymal collapse was detected with a reticulin stain, and rare sinusoidal collagen fibrils were evident with Masson’s trichrome adjacent to the hepatic venule. Histological features were most consistent with colchicine hepatotoxicity as reported in humans, where metaphase arrest and ring mitoses reflect repression of mitotic spindle formation.⁴ Although there was no evidence of an unrelated liver disorder, contribution of idiosyncratic ketoconazole hepatotoxicity could not be discounted.

View Figure

• Download as Powerpoint
• Download Figure

The dog made a gradual recovery within 2 wk based on physical examination and follow-up laboratory assessments (Day 35, Table 1) that documented persistent erythrocytic microcytosis but a normal leukogram and marked reductions/normalization of liver and muscle enzyme activity and total bilirubin and cholesterol concentrations. A urine sample disclosed a specific gravity of 1.055, pH of 7, and benign sediment. Reassessment on Day 129 confirmed resolution of all clinicopathological abnormalities with the exception of the erythrocyte microcytosis. Fasting and 2 hr postprandial serum bile acids were normal (2 and 7 µmol/L, respectively [reference range, <25 µmol/L]) and MDR1 genotyping revealed no loss of function mutation that might explain predisposition to colchicine toxicity^q.

Discussion

Colchicine, a neutral lipophilic alkaloid extracted from two plants (Colchicum autumnale [autumn crocus, meadow saffron] and Gloriosa superba [glory lily]) has a narrow therapeutic/toxicity window with marked variability between individuals (humans) in drug disposition.⁵ However, there is little information regarding colchicine metabolism and disposition in dogs with only a single study describing 23% albumin binding.¹³ To the authors’ knowledge, there are no studies describing plasma colchicine pharmacokinetics in dogs. Considering the rapid gastrointestinal absorption of colchicine that follows oral dosing in humans and extensive first pass hepatic extraction and metabolism of colchicine that limit systemic drug bioavailability, approximately 25–50% of an oral dose is rapidly eliminated.^2,3 Wide extrahepatic distribution of unmetabolized drug occurs due to its low protein binding. The lipophilic nature of colchicine augments its transport across membranes, allowing access to its primary target of cytosolic tubulin, which also serves as a drug reservoir.⁵ In people and rodents, colchicine is predominantly eliminated through biliary excretion in feces, with enterocyte turnover contributing to its elimination. Excretion into bile is mediated in part by MDR1, achieving an approximate100:1 enrichment in bile relative to blood.¹⁴ Cellular extrusion of colchicine, including enterocyte drug release into intestinal lumen, is also mediated by MDR1.⁵ A lesser but significant role in human colchicine metabolism (approximating 5–20%) occurs by enterocyte and hepatic cytochrome P450-3E4 activity, which forms inactive metabolites.^3–5,15 Comparable information is not available for the dog.

In humans, multiple MDR1 expression modulators and competitive substrates as well as inhibitors of P450 cytochromes involved with colchicine metabolism impart clinically significant colchicine drug-drug interactions.⁵ Best characterized are adverse interactions with macrolide antibiotics (erythromycin, clarithromycin), cyclosporine, ketoconazole, itraconazole, and hypolipidemic statins, as well as natural grapefruit juice.³ In vitro cell line studies also have identified additional drugs that may modify colchicine clearance including but not limited to prazosin, vinblastine, quinidine, H₂ blockers, glucocorticoids, and verapamil.^5,16

Cells with high proliferative rates are disproportionately affected by colchicine owing to formation of a tubulin-colchicine complex that physically interferes with elongation of the microtubule polymer necessary for cell division. Colchicine arrests microtubule elongation at low concentrations and causes microtubule depolymerization at higher concentrations.⁵ Through that mechanism, colchicine can influence a multitude of cellular processes in addition to cell division (e.g., signal transduction, gene expression, cell shape, cell migration, and cytosolic protein assembly within the Golgi apparatus, which is essential for endocytosis and exocytosis).⁵ The classic histological lesion associated with colchicine hepatotoxicity is inhibition of cell spindle formation essential for chromosome movement, leading to metaphase arrest and ring mitoses, which was observed in hepatocytes from the dog described in this report.^5,17

Therapeutic effects of colchicine in treatment of FMF and FSPF are thought to involve reduced expression of adhesion molecules on neutrophil membranes limiting their participation in local inflammation.^4,5 A paucity of MDR1 expression in neutrophils aligns with their three-fold greater intracellular colchicine concentration compared with lymphocytes and their predisposition to colchicine effects. In FMF, colchicine also modulates chronic inflammation, driven by pyrin, through formation of a cytosolic colchicine-pyrin interaction.⁵ Because colchicine modulation of gene expression (at high dosages) has a delayed onset in dampening inflammation during cyclic FMF, chronic drug administration, rather than either targeted treatment of disease flare or recurrence, is recommended.⁵ A similar chronic treatment strategy has been advised for FSPF.⁸

It is possible to measure colchicine plasma concentrations by liquid chromatography-mass spectrometry, high-performance liquid chromatography, or radioimmunoassay methods. However, plasma colchicine measurements are neither readily available in the clinical setting for humans nor useful in assessment and management of suspected toxicity because there is no established correlation of illness severity with plasma colchicine concentrations. That finding likely reflects the fact that intracellular, rather than plasma colchicine concentrations, are most relevant.³ In the case described herein, a laboratory providing a validated canine assay could not be identified, and drug administration had been discontinued for nearly 1 wk before referral where consideration of a potential adverse drug interaction was suspected.

Colchicine influences function of multiple cell types and thus can impart multiorgan toxicity; organs having high cell turnover rates are most vulnerable (gastrointestinal tract, bone marrow).^7,8 Three phases of colchicine toxicity have been characterized in acutely intoxicated humans. The initial phase involves gastrointestinal signs manifesting within 24 hr of exposure, characterized by nausea, vomiting, diarrhea, abdominal pain, hypovolemia and leukocytosis. Some of these features were recognized in the dog of this report. The second phase develops within 7 days of toxic drug exposure and involves multiorgan injury (e.g., respiratory, cardiac, neurologic, renal, hepatic, skeletal muscle, and bone marrow). This phase may initiate disseminated intravascular coagulation, pancytopenia, hemolysis, and metabolic derangements (electrolytes, hypoglycemia, metabolic acidosis). A recovery phase follows within 7–21 days of toxic drug exposure characterized by resolution of organ system derangements and a rebound leukocytosis.³ In the dog described herein, gastrointestinal toxicity was identified within days of colchicine dose escalation with rhabdomyolysis and hepatotoxicity suspected based on the profound muscle pain and markedly increased CK and escalating AST and ALT activities.

Although uncommon, colchicine-induced myotoxicity is well characterized in humans but has neither been recognized nor described in dogs.^3,18 In people, it associates with an abrupt onset proximal limb muscle weakness and hyperpathia and marked increases in CK, ALT and AST activities, as observed in the case described herein.^6,18 Myopathology has been confirmed in a small number of humans using electrophysiological assessments (e.g., positive sharp waves and fibrillation potentials) and muscle biopsy. A myocyte vacuolar change in the absence of either necrosis or inflammation is thought to reflect accumulation of lysosomes and autophagic vacuoles from a disrupted microtubule-dependent cytoskeletal network.¹⁸ In one large case series of colchicine myopathy in humans (n = 82), toxicity was most commonly associated with a colchicine-ketoconazole drug interaction, as in the dog in this case report.¹⁸

Although most side effects of colchicine are unequivocally dose related, increasingly recognized drug-drug interactions explain colchicine toxicity developing on “standard” daily low-dose regimens, as suspected in the case described herein.^5,12 In humans, ketoconazole simultaneously imposes P450 cytochrome inhibition of colchicine metabolism and MDR1 competitive inhibition, which impair drug metabolism and its cellular efflux.^12,18 Thus, concurrent ketoconazole administration increases risk for colchicine toxicity even with conventional dose administration.^12,18 In people, pharmacogenetic differences among individuals (e.g., MDR1 polymorphisms) independently influence colchicine disposition and metabolism heightening risk for toxicity with conventional dosing.⁵ Although a functional polymorphism of MDR1 is documented in some dog breeds, the dog reported herein tested negative for the canine MDR1 loss of function mutation.¹⁹ Investigation into the influence of ketoconazole on colchicine pharmacokinetics in healthy humans demonstrated a 100% increase in mean maximal plasma colchicine concentration, 210% increase in total colchicine exposure, 70% reduction in total apparent oral colchicine clearance, and a four-fold prolongation of the terminal elimination half-life (6.3–26 hr).¹² Thus, it is quite plausible that colchicine toxicity manifested within days of once daily colchicine administration in this dog because of concurrent ketoconazole administration.

Although the possibility exists that idiosyncratic ketoconazole toxicity directly contributed to liver injury in this case, ketoconazole is rarely associated with hepatic injury in human beings and only anecdotally in dogs.^20–22 Rare ketoconazole induced hepatotoxicity in humans may cause an acute increase in ALT and AST associated with necrosis either with or without cholestasis and rare case fatalities (fulminate hepatic failure) have been reported. Asymptomatic transient increases in serum transaminase activity are also reported with a similar scenario recognized in dogs.^20,22 However, hepatic histological changes associated with ketoconazole hepatotoxicity have not been characterized in the dog to the authors’ knowledge. Because ketoconazole was administered within the conventional dosing range, the dog tested negative for the canine MDR1 loss-of-function mutation and signs of illness only surfaced after increasing the colchicine dose frequency, the study authors believe that colchicine toxicity was the most likely cause of the multiorgan illness involving gastrointestinal, hepatic, and skeletal muscle toxicity. The persistent red blood cell microcytosis despite resolution of clinical illness might represent a unique breed-related finding because microcytosis has been commonly observed in shar peis in one of the author’s hospitals (S.C.).

There is no effective antidote for colchicine poisoning; therefore, early recognition of risk for an adverse drug-drug interaction requires judicious reduction of colchicine dose or dosing frequency and immediate drug discontinuation at the earliest suspicion of toxicity. Heightened risk for toxicity has been shown in humans with preexistent hepatic or renal functional impairment.³ There was no evidence of antecedent renal or hepatic dysfunction in the dog reported herein, and liver biopsy ruled out an antecedent chronic hepatopathy. Optimal management of acute large-dose colchicine toxicity includes immediate drug discontinuation, gastrointestinal decontamination (gastric lavage), repeat administration of activated charcoal to thwart enterohepatic drug recirculation, and supportive care addressing organ system toxicity and fluid and electrolyte imbalances. In the case described herein, rapid resolution of gastrointestinal and myopathic signs occurred within a few days of treatment discontinuation. Hepatic biopsy, pursued because of escalating liver enzyme activity and development of hyperbilirubinemia despite resolution of other clinical signs, revealed classic features of colchicine toxicity (i.e., hepatocyte metaphase arrest and ring mitoses). Resolution of all clinicopathological abnormalities within 3 mo of drug withdrawal further implicated an adverse drug-drug interaction resulting in colchicine toxicity. Because the dog described in this report has not been rechallenged with either colchicine or ketoconazole alone, the role of an idiosyncratic drug toxicity cannot be discounted.

Conclusion

Overlooked iatrogenic drug interactions are a recognized cause of illness and death in human beings and are most commonly encountered when multiple medications are used to treat diverse disease processes.²³ The dog reported herein demonstrated clinical features of an avoidable adverse colchicine-ketoconazole interaction previously described in humans. This case illustrates the importance of investigating potential for adverse drug interactions when unusual multidrug treatment regimens are instituted.