Delayed Presentation of Nitroprusside-Induced Cyanide Toxicity

Cyanide toxicity is a rare complication of sodium nitroprusside that can be difficult to diagnose in critically ill patients. We describe a case of cyanide toxicity after cardiac surgery that presented as lactic acidosis after discontinuation of nitroprusside. Cyanide toxicity is a rare complication of sodium nitroprusside that can be difficult to diagnose in critically ill patients. We describe a case of cyanide toxicity after cardiac surgery that presented as lactic acidosis after discontinuation of nitroprusside. Sodium nitroprusside is widely used for titratable control of blood pressure. Due to concerns over cyanide toxicity, it is recommended that dosing not exceed 2 mcg·kg−1·min−1. The incidence of nitroprusside-induced cyanide toxicity is unknown but is likely less than 2.5% [1Patel C.B. Laboy V. Venus B. Mathru M. Wier D. Use of sodium nitroprusside in post-coronary bypass surgery. A plea for conservatism.Chest. 1986; 89: 663-667Crossref PubMed Scopus (21) Google Scholar, 2Lockwood A. Sodium nitroprusside-associated cyanide toxicity in adult patients - fact or fiction? A critical review of the evidence and clinical relevance.Open Access Journal of Clinical Trials. 2010; 2010: 133-148Google Scholar]. Nonetheless, the risk is serious because diagnosis can be challenging. We present a case of nitroprusside-induced cyanide toxicity after cardiac surgery, with delayed manifestation hours after discontinuation of nitroprusside infusion. It underscores the importance of a high index of suspicion when unexplained lactic acidosis occurs in the context of nitroprusside exposure and the utility of antidotes for empiric confirmation and treatment. The patient was a 67-year-old obese female, with sleep apnea, chronic kidney disease, hypertension, hyperlipidemia, and diabetes mellitus. She presented with congestive heart failure and underwent uneventful 3-vessel coronary artery bypass and mitral annuloplasty. Within 20 hours she was extubated and off all infusions except nitroprusside. Over the next 2 days she developed some typical postcardiac surgery complications including first degree heart block, junctional rhythm, and atrial fibrillation. Treatment with amiodarone caused complete heart block requiring epicardial pacing. Progressive oliguria also occurred due to acute-on-chronic kidney injury, despite fluids to attain euvolemia and later, attempted forced diuresis. We treated mild delirium with quetiapine. Meanwhile, she remained on nitroprusside for hypertension. By postoperative day 4 she was anuric, but delirium had resolved, cardiac rhythm was stable, and hypertension was controlled with oral medications. An incidental arterial blood gas (ABG) revealed lactate at 5.5 mmol/L. Her vital signs were normal and, aside from mild nausea, she was comfortable and had no complaints. Lactate was higher (7.2 mmol/L) on repeat ABG with worsening base deficit. Supplemental bicarbonate was given and evaluation initiated to ensure adequate oxygen delivery and detect increased metabolic demand. Echocardiogram showed normal biventricular function. Thermodilution cardiac index exceeded 2.4 L/m2. Mixed venous oxygen saturation (SVO2) was 78%. Filling pressures indicated euvolemia. Hemoglobin was 9 g/dL and had been stable postoperatively; abdomen was nontender with vigorous bowel sounds. She remained asymptomatic over the next 18 hours, although lactate steadily increased to 18 mmol/L by midnight of postoperative day 4. She then deteriorated, exhibiting restlessness, tachypnea, tachycardia, and hypotension. Inotropes, vasopressors, and bicarbonate infusions were started. She was reintubated for increasing work of breathing and uncompensated acidosis. Continuous hemodialysis was started yet lactate levels kept rising, peaking at 22 mmol/L by 0900 the next day. With clinical exam now unreliable, computerized tomography scans were obtained but found no ischemia, inflammation, or infection. Lipase, amylase, and liver function tests were normal. Further review clarified that she had received 319 mg of nitroprusside (0.12 to 2.125 mcg/kg/min) for 70 hours, ending about 12 hours prior to the rise in lactate. This period coincided with her worsening renal function. Therefore, having excluded oxygen supply and demand problems we considered cyanide and thiocyanate toxicity. Accordingly we measured their levels using fresh and residual specimens from the preceding 96 hours. Thiocyanate levels were low and cyanide levels, reported days later, were normal. Nonetheless we treated her with hydroxocobalamin at 1,100 on postoperative day 5. Subsequent ABGs showed rapid decline in lactate within a few hours (Fig 1). Concurrently, SVO2 dropped to about 61% (a level that better reflected her postoperative context). She was extubated and off vasoactive medications within 24 hours of receiving hydroxocobalamin. Dysoxia had persisted for almost 48 hours before treatment. By postoperative day 6, aspartate and alanine aminotransferase levels were elevated but normalized within 2 days. She was transitioned to intermittent hemodialysis and renal function recovered to baseline by postoperative day 14. She went home on postoperative day 20. Sodium nitroprusside interacts with oxyhemoglobin releasing nitric oxide, cyanide, and methemoglobin [2Lockwood A. Sodium nitroprusside-associated cyanide toxicity in adult patients - fact or fiction? A critical review of the evidence and clinical relevance.Open Access Journal of Clinical Trials. 2010; 2010: 133-148Google Scholar]. Cyanide is metabolized by hepatic rhodanase, using thiosulfate to produce renally excreted thiocyanate. Thiosulfate stores are limited, with an estimated maximum metabolic capacity for 50 mg of nitroprusside, although this does not account for other sulfate donors [2Lockwood A. Sodium nitroprusside-associated cyanide toxicity in adult patients - fact or fiction? A critical review of the evidence and clinical relevance.Open Access Journal of Clinical Trials. 2010; 2010: 133-148Google Scholar]. Toxicity results when unmetabolized cyanide binds to cytochrome C oxidase, inhibiting oxidative phosphorylation and resulting in anaerobic metabolism with lactate production. With sustained administration at greater than 2 mcg · kg −1·min−1, cyanide production will exceed the endogenous metabolic capacity which may be augmented by co-administration of thiosulfate. Our patient’s total nitroprusside dose over a relatively short period likely exceeded her metabolic capacity. Features of cyanide toxicity are nonspecific, including altered sensorium, seizures, coma, hypotension, tachycardia, arrhythmias, and gastrointestinal disturbances. In emergency rooms, circumstances such as fire and smoke inhalation and poisoning will typically raise suspicion, whereas in intensive care units presentation of nitroprusside-induced cyanide toxicity is atypical and insidious [3Hall A.H. Rumack B.H. Clinical toxicology of cyanide.Ann Emerg Med. 1986; 15: 1067-1074Abstract Full Text PDF PubMed Scopus (185) Google Scholar]. Tachyphylaxis may occur but other features could be masked by underlying critical illness. Lactic acidosis was the presenting feature in our patient. Although some believe that lactic acidosis only occurs as a terminal event related to circulatory failure (rather than to cyanide toxicity), others view lactic acidosis as an indicator correlating with severity of cyanide poisoning [4Robin E.D. McCauley R. Nitroprusside-related cyanide poisoning. Time (long past due) for urgent, effective interventions.Chest. 1992; 102: 1842-1845Crossref PubMed Scopus (70) Google Scholar, 5Baud F.J. Borron S.W. Mégarbane B. et al.Value of lactic acidosis in the assessment of the severity of acute cyanide poisoning.Crit Care Med. 2002; 30: 2044-2050Crossref PubMed Scopus (167) Google Scholar]. Our patient’s presentation was also unusual in that lactate elevation occurred 12 hours after discontinuation of nitroprusside, and nearly 24 hours elapsed before other features manifested. We are unaware of any other reports of delayed manifestation of toxicity after discontinuation of nitroprusside. We postulate that her cyanide accumulation was insufficient to cause a catastrophic cessation of oxidative metabolism but enough to impair oxidative phosphorylation. Then, steady compounding of the resultant oxygen debt caused delayed but persistent lactate production. We also considered thiocyanate toxicity because our patient developed renal failure during nitroprusside therapy. Thiocyanate accumulates in renal dysfunction, truncating metabolism and increasing the cyanide pool. Moreover, thiocyanate toxicity resembles cyanide toxicity with features including anxiety, confusion, hyperreflexia, miosis, muscle cramps, seizures, tinnitus, and coma. However, lactic acidosis is absent and thiocyanate levels correlate with severity of toxicity. Our patient’s thiocyanate levels were unremarkable, likely due to thiosulfate depletion from metabolic consumption, increased perioperative excretion, and malnutrition [1Patel C.B. Laboy V. Venus B. Mathru M. Wier D. Use of sodium nitroprusside in post-coronary bypass surgery. A plea for conservatism.Chest. 1986; 89: 663-667Crossref PubMed Scopus (21) Google Scholar]. Cyanide assays are not useful because levels do not correlate with the severity of toxicity due to technical issues with specimen selection and handling. Optimal specimen handling involves anticoagulation, addition of sodium nitrite or silver sulfate, and prompt storage at −20°C [2Lockwood A. Sodium nitroprusside-associated cyanide toxicity in adult patients - fact or fiction? A critical review of the evidence and clinical relevance.Open Access Journal of Clinical Trials. 2010; 2010: 133-148Google Scholar, 6Lundquist P. Rosling H. Sörbo B. Determination of cyanide in whole blood, erythrocytes, and plasma.Clin Chem. 1985; 31: 591-595PubMed Google Scholar, 7Vesey C.J. Langford R.M. Stabilization of blood cyanide.J Anal Toxicol. 1998; 22: 176-178Crossref PubMed Scopus (5) Google Scholar]. These prevent cyanide evaporation or conversion of thiocyanate to cyanide ex vivo. Even with proper steps, degradation still occurs in frozen specimens. Moreover, cyanide assays differ in sensitivity, specificity, technical difficulty, turnaround time, cost, and availability. These may explain why our retrospectively measured cyanide levels were normal; the specimens were not specifically processed for cyanide testing. Management includes cardiopulmonary supportive measures, buffering of acidosis, and antidotes to decouple cyanide from mitochondria and detoxify it [8Barillo D.J. Diagnosis and treatment of cyanide toxicity.J Burn Care Res. 2009; 30: 148-152Crossref PubMed Scopus (52) Google Scholar]. Amyl nitrite (or sodium nitrite) is administered to form methemoglobin, which binds cyanide to yield cyanmethemoglobin. Thiosulfate is then given and reacts with cyanmethemoglobin to form thiocyanate, which is dialyzable. Alternatively hydroxocobalamin administered alone rapidly chelates cyanide forming nontoxic cyanocobalamin. After administering hydroxocobalamin, the rapid clearance of lactate confirmed our diagnosis of nitroprusside-induced cyanide toxicity. Decoupling of cyanide from the mitochondria allowed unimpaired oxidative phosphorylation and normalization of SVO2. Nitroprusside-induced cyanide toxicity is uncommon but can be severe and fatal. After cardiac surgery its nonspecific clinical features may contribute to delayed recognition. Unexplained lactic acidosis could be the only indicator and may not manifest until hours after discontinuation of nitroprusside. In the context of recent or ongoing nitroprusside exposure, the risk of cyanide toxicity should not be overlooked. Clinicians should monitor infusion rates and cumulative doses and maintain vigilance with a high index of suspicion for cyanide toxicity.