Abstract
The authors present the case of a 44-year-old American Indian woman with hematemesis, spontaneous cutaneous hemorrhages, and multiple ecchymoses. Coagulation factor analyses demonstrated both prolonged prothrombin time (PT, >40 s) and prolonged partial thromboplastin time (PTT, >120 s). Measurement of the serum level of brodifacoum (37 ng/mL), one of the superwarfarin agents commonly used in rodenticides, confirmed poisoning as the cause of the patient's symptoms. Substantial amounts of fresh frozen plasma and vitamin K were required to obtain normal coagulation parameters and maintain these parameters over a 3-week inhospital period. Oral administration of vitamin K (100 mg daily) maintained normal PT (14.1 s), PTT (33.0 s), and international normalized ratio (INR, 1.48) at 2 weeks after the patient was discharged from the hospital. By 2 months postdischarge, PT, PTT, and INR returned to elevated levels because of patient noncompliance with the prescribed tapering vitamin K regimen.
A44-year-old American Indian woman was admitted to the intensive care unit (ICU) for management of a coagulopathy requiring coagulation factor replacement therapy. The patient reported an episode of hematemesis approximately 2 months prior to admittance. During this episode, the patient underwent an esophagogastroduodenoscopy (EGD), which revealed severe esophagitis. She was prescribed a regimen of lansoprazole and sucralfate at the time. The 4 weeks following her EGD were uneventful.
At the end of this 4-week period, the patient experienced a sudden recurrence of hematemesis and noticed spontaneous cutaneous hemorrhages (1–4 mm) and ecchymoses (1–4 cm) in areas of minimal trauma on her body.
The patient then visited an outlying tertiary care center, where measurements were made of her prothrombin (coagulation factor II) time (PT) and partial thromboplastin time (PTT). To restore normal coagulation activity, the patient's therapy included fresh frozen plasma and vitamin K. The discharge summary from the hospitalization lists a diagnosis of possible Munchausen syndrome (factitious disorder with physical symptoms) and ingestion of warfarin, a synthetic anticoagulant. The hospital's treatment course indicated that the patient's serial warfarin levels were negative, while mixing studies indicated correction of her PT with coagulation factor replacement. Quantification of the amount of fresh frozen plasma and vitamin K administered to the patient during this hospitalization was not provided in the discharge summary. The patient's hospitalization ended after 10 days, when she was denied smoking privileges and left against medical advice.
One week after discharge, the patient noted a recurrence of diffuse spontaneous skin hemorrhages and ecchymoses over her upper and lower extremities and oral mucosa. She also had an incident of uncontrolled bleeding after accidentally cutting her hand with a kitchen knife while preparing a meal. In addition to developing lesions, she began to have bilateral calf pain associated with bruising and swelling. These complaints prompted her to report to the Indian Health Service clinic in Tulsa, Okla, where prolongation of her PT and PTT was once again noted. She received subcutaneous vitamin K (25 mg) and was transferred to our tertiary care hospital (Tulsa Regional Medical Center) for additional evaluation and management.
Evaluation and Management of Case
Upon admission to the Tulsa Regional Medical Center, the patient appeared to be well nourished. Physical examination revealed scattered purpuric lesions on her scalp, with some areas of excoriation that were actively bleeding. She also had a small hematoma on the left corner of her lower lip and purpuric lesions scattered on her oral mucosa. Examination of her skin revealed numerous cutaneous hemorrhages on both her upper and lower extremities, her abdomen, and her back. In addition, multiple areas of ecchymoses in several stages of healing were noted. No evidence of hemearthrosis was observed. A neurovascular examination revealed no abnormalities.
The results of our initial laboratory evaluation of the patient are reported in
Table 1. The coagulation factor analyses demonstrated both prolonged PT (>40 s; reference range, 10.4–12.6 s) and prolonged PTT (>120 s; reference range, 25–35 s). It was not necessary to initially test for the international normalized ratio (INR), which is useful for confirming PT results, because of the extremely elevated PT level.
Table 1
Initial Laboratory Test Results of Patient
Factor Tested
|
Test Result
|
Reference Range
|
|---|
| □ Fibrinogen, mg/dL | 339 | 200-420 |
| □ Hematocrit, % | 27.6 | 36.1-44.3 (female) |
| □ Hemoglobin, g/dL | 9.5 | 12.1-15.1 (female) |
| □ International normalized ratio* | ... | ... |
| □ Partial thromboplastin time, s | >120 | 25-35 |
| □ Platelet count, /μL | 386,000 | 150,000-400,000 |
| □ Prothrombin time, s† | >40 | 10.4-12.6 |
| □ Qualitative D-dimer | Negative | ... |
| □ Thrombin time, s | 13.8 | 14.5-18.9 |
□ White blood cell count, /μL
| 9600
| 4500-10,000
|
Table 1
Initial Laboratory Test Results of Patient
Factor Tested
|
Test Result
|
Reference Range
|
|---|
| □ Fibrinogen, mg/dL | 339 | 200-420 |
| □ Hematocrit, % | 27.6 | 36.1-44.3 (female) |
| □ Hemoglobin, g/dL | 9.5 | 12.1-15.1 (female) |
| □ International normalized ratio* | ... | ... |
| □ Partial thromboplastin time, s | >120 | 25-35 |
| □ Platelet count, /μL | 386,000 | 150,000-400,000 |
| □ Prothrombin time, s† | >40 | 10.4-12.6 |
| □ Qualitative D-dimer | Negative | ... |
| □ Thrombin time, s | 13.8 | 14.5-18.9 |
□ White blood cell count, /μL
| 9600
| 4500-10,000
|
×
Additional medical history obtained in the initial evaluation of the patient included hypertension, which had previously been managed with amlodipine, and depression, which had been managed with inpatient psychiatric hospitalization. The patient denied having a prior history of bleeding into a joint or elongated bleeding after surgery or tooth extraction. No recent history of snake bite or hyperthermia were reported. Other medications used by the patient in her previous hospitalizations included sucralfate (1 g orally before meals and at bedtime) and lansoprazole (30 mg orally four times per day). The patient denied using over-the-counter medications, alcohol, or recreational drugs. Ingestion of Coumadin (Bristol-Myers Squibb Co; Princeton, NJ) (warfarin sodium) or derivatives was also adamantly denied—a denial that was supported by negative test results for Coumadin upon admission to our hospital.
After this information was obtained in the patient's initial evaluation, a psychiatric consultation was obtained. The consultation revealed that the patient's depression had been managed with several different regimens in the past. Concerning her past inpatient psychiatric hospitalization, the patient remarked, “I was either crazy or my husband was poisoning me.” The patient further reported her belief that she probably had been poisoned numerous times by her husband, whom she had not seen in 4 weeks. (Although we were able to establish superwarfarin ingestion as the cause of the patient's symptoms, we were not able to definitively establish the cause of this ingestion.)
The patient's initial treatment course at our hospital continued for 2 weeks. During this time, she was closely observed both in the ICU and on the general medical floor for any sign of self-medication. However, no such questionable behavior was observed. Because of her prolonged PT and PTT, continual replacement of coagulation factors via plasma transfusion was necessary. In total, the patient required 56 units of fresh frozen plasma, averaging six units daily for 10 days. She also received daily doses (5 mg) of vitamin K. This regimen corrected the patient's PT to an average time of 18 seconds, and her PTT to an average of 40 seconds. Her INR tested at an average of 2.5 (reference range, 0.8–1.2)
Although substantial corrections of the patient's PT and PTT were obtained with treatment, an incomplete correction was suggested by a 50:50 mixing study on serial dilutions, which pointed to a possible circulating inhibitor as a component of the patient's anticoagulated state. The possibility of an autoimmune process behind the symptoms was considered, and the patient was prescribed the corticosteroid solumedrol for immunosuppression.
Superwarfarin toxicity was also considered. Superwarfarins are a class of rodenticides that include brodifacoum, bromadiolone, chlorophacinone, difenacoum, and difethiacone.
1 We contacted the national hotline of the American Association of Poison Control Centers (AAPCC), to learn what serologic tests are available for ingestion of superwarfarins and where these tests could be performed (
Figure). Serologic tests for two of the more common superwarfarins—brodifacoum and difenacoum—were referred out to a recommended laboratory. Laboratory evaluation for these substances is performed using reversed-phase high-performance liquid chromatography (HPLC) or radioimmunoassay/enzyme-linked immunosorbent assay (RIA/ELISA). The patient's brodifacoum level by HPLC was 37 ng/mL—a positive result indicating poisoning.
Within 24 hours of contacting the AAPCC hotline, empiric treatment with oral vitamin K (100 mg daily) was initiated. Laboratory tests performed within 24 hours of the initial vitamin K dose revealed corrected PT (16.3 s), PTT (29.0 s), and INR (1.3). All fresh frozen plasma support was discontinued at this point. The patient was discharged after 5 days of the 100 mg vitamin K regimen.
Diagnostic Findings
Treatment of the patient with fresh frozen plasma decreased clinical bleeding and improved INR, consistent with a decreased level of coagulation factor production. The patient's PT became repeatedly elevated in 24-hour intervals after transfusions of fresh frozen plasma were discontinued. Subsequent attempts to identify coagulation factor deficiencies revealed inconsistent results. Because both PT and PTT were elevated, the activity level of factor V (proaccelerin) was evaluated. The level of factor V activity was found to be 90% (reference range, 50%–150%). The level of factor VIII (antihemophilic factor) activity was 168% (50%–150%).
Initial evaluation of the level of factor VII (proconvertin) activity indicated 7%, but there was reversal of this level with further dilution, suggesting the presence of an inhibitor. The Bethesda unit measurement of the inhibitor was extremely low (0.2 U/mL). (One Bethesda unit is the amount of inhibitor that inactivates 50% [0.5 unit] of a coagulation factor during a given incubation period.) We believed that this degree of inhibitor activity was not sufficient to explain the severe coagulopathy that was present in the patient. If a nonspecific inhibitor was present, it would be expected to affect factor VIII as well.
2
The patient's renal function was normal. However, liver function studies revealed mild aminotransferase elevation, with an aspartate aminotransferase level of 88 U/L and an alanine aminotransferase level of 92 IU/L. These levels began to normalize on the patient's sixth day of hospitalization. The patient's serum albumin level was normal, at 4.5 g/dL. Test results for hepatitis A, B, and C were negative.
Computed tomography scans of the patient's abdomen on the fifth day of hospitalization revealed hepatomegaly, with the liver measuring 25 cm in length. The spleen, pancreas, and kidneys were unremarkable. A liver biopsy indicated hepatic parenchyma, with nonspecific inflammatory changes. No cirrhosis or fibrosis was noted.
Evaluation for autoimmune disease in the patient revealed a positive result in the test for antinuclear antibody, with a nucleolar pattern suggestive of autoimmune disease seen at titers of 1:640 (negative result, <1:40). Test results were negative for rheumatoid arthritis factor, extractable nuclear antigens, and antibodies for Sjögren syndrome A and B. The Russell Viper Venom Test, a test for lupus anticoagulant, revealed an elevated clotting time of 44.4 seconds (reference range, 0–42.8 s), with a normal viper venom ratio and detection of a factor VII inhibitor. Other study results for lupus anticoagulant are reported in
Table 2.
Table 2
Lupus Anticoagulant Study Results of Patient *
Factor Tested
|
Test Result
|
Reference Range
|
|---|
| ▪ PTT Studies | | |
| □ Heparin adsorption, U/mL | 45.8 | 20.9-37.4 |
| □ aPTT (heparin adsorption), s | 51.7 | <43.0 |
| □ aPTT (patient), s | 48.5 | 20.9-37.4 |
| □ PTT (lupus), s | 55.5 | <43.0 |
| ▪ Mixing Studies, s | | |
| □ PTT (1:1 NP) mix | 35.5 | 25-35 |
| □ PTT (1:1 NP) RT | 32.0 | 25-35 |
| ▪ Cardiolipin Antibody Studies | | |
| □ Cardiolipin IgA, APL U/mL | 18.0 | 0-12 |
| □ Cardiolipin IgG, GPL U/mL | 5.2 | 0-15 |
| □ Cardiolipin IgM, MPL U/mL | 19.0 | 0-12.5 |
| ▪ Other Studies | | |
| □ Antidouble-stranded DNA antibody, IU/mL | 83 | 0-200 |
| □ Fluorescent treponemal antibody | Nonreactive | ... |
| □ Complement activity CH50, U/mL | 442 | 400-700 |
| □ Complement protein C3, mg/dL | 134 | 73-183 |
□ Complement protein C4, mg/dL
| 17
| 15-59
|
Table 2
Lupus Anticoagulant Study Results of Patient *
Factor Tested
|
Test Result
|
Reference Range
|
|---|
| ▪ PTT Studies | | |
| □ Heparin adsorption, U/mL | 45.8 | 20.9-37.4 |
| □ aPTT (heparin adsorption), s | 51.7 | <43.0 |
| □ aPTT (patient), s | 48.5 | 20.9-37.4 |
| □ PTT (lupus), s | 55.5 | <43.0 |
| ▪ Mixing Studies, s | | |
| □ PTT (1:1 NP) mix | 35.5 | 25-35 |
| □ PTT (1:1 NP) RT | 32.0 | 25-35 |
| ▪ Cardiolipin Antibody Studies | | |
| □ Cardiolipin IgA, APL U/mL | 18.0 | 0-12 |
| □ Cardiolipin IgG, GPL U/mL | 5.2 | 0-15 |
| □ Cardiolipin IgM, MPL U/mL | 19.0 | 0-12.5 |
| ▪ Other Studies | | |
| □ Antidouble-stranded DNA antibody, IU/mL | 83 | 0-200 |
| □ Fluorescent treponemal antibody | Nonreactive | ... |
| □ Complement activity CH50, U/mL | 442 | 400-700 |
| □ Complement protein C3, mg/dL | 134 | 73-183 |
□ Complement protein C4, mg/dL
| 17
| 15-59
|
×
When all laboratory results were reviewed as a whole, elevation of cardiolipin immunoglobulin A and immunoglobulin M antibodies (
Table 2) were interpreted as being part of a reactive process. The test result for serum protein electrophoresis was negative, as was the urine drug screen. A peripheral blood smear revealed microcytic normochromic anemia.
Comment
The patient in the present case report had a clinical history and symptoms consistent with Coumadin overdose. The persistence and reoccurrence of a coagulopathy related to vitamin K deficiency, however, led to the suspicion of superwarfarin intoxication. Upon review of the literature,
1,4–9 it became apparent that coagulation parameters resulting from coumadin overdose are distinguishable from those resulting from superwarfarin intoxication. The brodifacoum level (37 ng/mL) revealed by the evaluation of the patient's serum with HPLC clearly indicated superwarfarin poisoning.
We presumed that the patient had no exposure to brodifacoum since her initial hospitalization at our tertiary care center. Therefore, at the time of discharge from our center, she was nearing completion of at least 3 weeks of half-life withdrawal from the vitamin K deficiency that was induced by brodificoum. Twenty-four hours after receiving her first dose of 100 mg oral vitamin K, tests showed complete normalization of her INR (1.3). Oral vitamin K therapy was administered without complications for the next 5 days, during which evidence of cutaneous hemorrhages and ecchymoses were resolved.
The patient was discharged from our hospital with prescriptions for a tapering dose of vitamin K over a period of 4 weeks (100 mg for 14 days, 10 mg for 7 days, and 5 mg for 7 days) and weekly INR measurements. If these follow-up measurements were to detect marked elevation of the INR, the previous dosing level would be maintained for an additional 2 weeks; then, another attempt at tapering would be made.
At outpatient follow-up 2 weeks after discharge, the patient's coagulation parameters remained within normal or near-normal ranges (PT, 14.1 s; PTT, 33.0 s; INR 1.48), with a normal platelet count of 349,000/μL. By 2 months after discharge, the patient had missed her serial INR measurements on multiple occasions. Laboratory evaluations at that time showed elevated coagulation parameters (PT, 31.8 s; PTT, 57.0 s; INR 3.22). The primary care physician reported that the patient admitted to noncompliance with her oral vitamin K regimen.
Pathologic Considerations
According to the AAPCC,
3 exposures to superwarfarins in 2004 numbered 16,054, with 14,229 of these poisonings occurring in patients younger than 6 years. All of the superwarfarins work by inhibiting the generation of the activated form of vitamin K
1 via inhibition of vitamin K
1 reductases.
4 Activation of the coagulation factors II, VII, IX, and X require the presence of vitamin K as a cofactor. When vitamin K cannot be regenerated, the coagulation factors cannot be activated, and a coagulopathy results involving both the extrinsic and intrinsic coagulation pathways.
5
Superwarfarins have a half-life of approximately 6 to 8 weeks.
6 Of the vitamin K–dependent factors, factor VII has the shortest half-life, at 5 hours.
6 Therefore, factor VII levels are used to monitor synthesis of vitamin K–dependent coagulation factors.
7 Patients are able to tolerate a reduction in coagulation factor synthesis of as much as 70% before manifesting coagulation disorders.
7
Chronic exposures and overdoses of superwarfarins may result in bleeding from any organ system. However, more than one system is usually involved. Other possible results of chronic exposure include ecchymoses, epistaxis, friable gums, hematemesis, hematuria, hemoptysis, melena, petechiae, and vaginal bleeding.
8 Measurements of specific coagulation factors support the diagnosis of superwarfarin toxicity if all vitamin K–dependent factors (II, VII, IX, and X) are decreased while others remain normal in the absence of Coumadin therapy, liver disease, or the use of an inhibitor.
4 In addition to using HPLC and RIA/ELISA to make diagnoses of toxicity, the vitamin K
1(2,3-epoxide):K
1 ratio can be used, with a ratio of 2:3 (normal, 0.1:0.2) indicating ingestion of superwarfarins.
9
There is no specific therapeutic regimen for patients diagnosed with superwarfarin poisoning other than restoration of normal PT and coagulation factor levels. In patients with active bleeding, treatment involves administering fresh frozen plasma and/or coagulation factor concentrates, in addition to packed red blood cells and oral vitamin K.
5 In patients who no longer demonstrate active bleeding or life-threatening hemorrhage but continue to have prolongation of PT, supratherapeutic doses of oral vitamin K (>100 mg daily) have been demonstrated to normalize PT.
5 Monitoring of broudifacoum concentrations are useful for predicting when to terminate therapy.
7
Conclusion
Of the 16,054 reported cases of superwarfarin exposure in 2004, 15,339 exposures were unintentional, according to the AAPCC.
3 Patients with unexplained anticoagulation and evidence of abnormal bleeding should be evaluated using both an individual history and a physical examination focusing on the consideration of possible superwarfarin exposure. Prolongation of both the PT and PTT with requirement for continual coagulation factor replacement—despite treatment with fresh frozen plasma—is also a clue to possible superwarfarin exposure. Reference laboratories are available to perform diagnostic tests for superwarfarin toxicity.
Kuijpers EA, den Hartigh J, Savelkoul TJ, de Wolff FA. A method for the simultaneous identification and quantitation of five superwarfarin rodenticides in human serum. J Anal Toxicol. 1995;19:557 –562.
Handin RL. Disorders of coagulation and thrombosis. In: Kasper DL, Braunwald E, Fauci AS, Hauser SL, Longo DL, Jameson JL, eds. Harrison's Principles of Internal Medicine. 16th ed. New York, NY: McGraw-Hill; 2005:685 .
Watson WA, Litovitz TL, Rodgers GC Jr, Klein-Schwartz W, Reid N, Youniss J, et al. 2004 annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System.
Am J Emerg Med. 2005;23:589–666. Available at:
http://www.aapcc.org/2004.htm. Accessed April 5, 2006.
Chua JD, Friedenberg WR. Superwarfarin poisoning. Arch Intern Med. 1998;158:1929 –1932.
Burkhart KK. Anticoagulant rodenticides. In: Ford MD, ed. Clinical Toxicology. 1st ed. Philadelphia, Pa: WB Saunders; 2000:848 –853.
Weitzel JN, Sadowski JA, Furie BC, Moroose R, Kim H, Mount ME, et al. Surreptitious ingestion of a long-acting vitamin K antagonist/rodenticide, brodifacoum: clinical and metabolic studies of three cases.
Blood. 1990;76:2555–2559. Available at:
http://www.bloodjournal.org/cgi/reprint/76/12/2555. Accessed April 5, 2006.
Bruno GR, Howland MA, McMeeking A, Hoffman RS. Long-acting anticoagulant overdose: brodifacoum kinetics and optimal vitamin K dosing. Ann Emerg Med. 2000;36:262 –267.
Burucoa C, Mura P, Robert R, Boinot C, Bouquet S, Piriou A. Chlorophacinone intoxication. A biological and toxicological study. J Toxicol Clin Toxicol. 1989;27:79 –89.
Ross GS, Zacharski LR, Robert D, Rabin DL. An acquired hemorrhagic disorder form long-acting rodenticide ingestion. Arch Intern Med. 1992;152:410 –412.