Secondary Logo

Journal Logo


Gussow, Leon MD

doi: 10.1097/01.EEM.0000296563.73057.88

Dr. Gussow is a voluntary attending physician at the John H. Stroger Hospital of Cook County in Chicago (formerly Cook County Hospital), an assistant professor of emergency medicine at Rush Medical College, and a consultant to the Illinois Poison Center.

Mnemonic devices are useful in remembering information that is used infrequently, and nowhere are they more valuable than when working through certain differential diagnoses. In one case, a 35-year-old woman was brought to the emergency department by her roommate because of abdominal pain, vomiting, and diarrhea.

The patient said the symptoms had been going on for five hours, and she denied fever, history of tuberculosis and other infectious or chronic diseases, and concurrent medical problems. She also denied being on any medication. She admitted alcohol abuse in the past, and said she had been previously hospitalized for alcohol withdrawal, but she denied other drug abuse. She experienced a vague syncopal episode in the waiting room that was not witnessed by medical staff. Her roommate was not experiencing any symptoms.

On examination, her vital signs were unremarkable with the exception of a respiratory rate of 24 breaths per minute. She was alert and oriented to place, person, and time. Pupils were 5 mm and reactive. Abdominal exam revealed decreased bowel sounds and marked diffuse tenderness without clear rebound or rigidity. The rectal exam was negative for occult blood. Although no focal neurological deficits were detected, the patient appeared too ill for a gait assessment.

Figure. I

Figure. I

Laboratory results included serum glucose 96 mg/dL, sodium 140 mEq/L, serum bicarbonate 5 mEq/L, BUN 24 mg/dL, creatinine 3.4 mg/dL, and anion gap 43. Serum salicylate and acetaminophen levels were undetected, and the urine drug screen was negative for opiates, amphetamines, barbiturates, benzodiazepines, cannabinoids, cocaine, methadone, phencyclidine (PCP), and propoxyphene. Methanol and ethylene glycol levels were negative. CT scan of the abdomen was unremarkable.

The toxicology service was consulted, and toxic alcohol levels were ordered along with another diagnostic test. What was that test?

Back to Top | Article Outline


The key finding in this case is a severe anion gap metabolic acidosis presenting with gastrointestinal symptoms. Certainly sepsis from an intra-abdominal source and toxic alcohol ingestion would both be high on the differential diagnosis, and in fact, this patient was initially treated with antibiotics and fomepizole to cover those possibilities. But it is important to consider all the possible causes of anion gap metabolic acidosis when deciphering cases in which the etiology is not certain. Fortunately, there is a mnemonic to help us remember the many items on the differential: A CAT PILES MUD. (See table 1.)

Table 1

Table 1

Reviewing the list, it was easy in this case to rule out severe acetaminophen toxicity, which rarely presents with a metabolic acidosis, and this was ruled out by the patient's negative serum level, as was aspirin toxicity. There was no history suggesting carbon monoxide exposure, and the patient's roommate was perfectly well. Cyanide acts and kills rapidly, and is unlikely to cause symptoms lasting many hours. Although there was no history of binge drinking that would have preceded the onset of alcoholic ketoacidosis, it would certainly be reasonable to check serum ketones and treat this patient with fluids, dextrose, and thiamine.

Theophylline rarely causes severe acidosis (pH<7.0), particularly in the absence of marked tachycardia and cardiac dysrhythmia. As for paraldehyde, it's just there so the mnemonic doesn't read A CAT ILES MUD. INH toxicity presents with severe metabolic acidosis and benzodiazepine-resistant status epilepticus, and despite the vague “syncopal” episode in the waiting room, there was no evidence of seizure activity in this case. Lactic acidosis isn't really a diagnosis; it's a secondary condition whose etiology must be determined. Renal failure doesn't usually cause acidosis this severe, nor does diabetic ketoacidosis when the serum glucose is as low as 148 mg/dL.

Metformin? There's no history of diabetes or drug ingestion, but as Sherlock Holmes remarked: “Improbable as it is, all other explanations are more improbable still.” The diagnostic test was a serum metformin level, which was 62 mcg/mL (therapeutic range 1-2 mcg/mL). When confronted with these results, the patient finally admitted that she had taken “several handfuls” of a relative's diabetes medication the day before she presented to the emergency department.

Metformin (dimethylbiguanide) is commonly prescribed to treat type 2 diabetes. It is cleared unchanged through the kidney, and readily accumulates in the presence of renal insufficiency. Because the metabolic effects of the drug include increased gastrointestinal production, along with decreased hepatic utilization of lactate, increased metformin levels caused by impaired clearance or overdose can result in a significant high anion gap lactic acidosis. (Despite some controversy in the medical literature, the link between increased metformin levels and acidosis seems incontrovertible.)

Initial manifestations of metformin-associated lactic acidosis (MALA) can be subtle and nonspecific. Early symptoms typically include nausea, vomiting, and abdominal pain. It is important to consider the diagnosis in anyone on the medication who presents with gastrointestinal complaints. Although the mortality rate for patients with MALA is said to be as high as 50 percent, this estimate was based on old data and that figure is probably much lower today. Because metformin has a low volume of distribution and does not bind extensively to proteins, it is very suitable for enhanced elimination by dialysis. Dialyzing against a lactate-free solution containing bicarbonate will remove both the drug and the lactic acid, and help correct the acid-base disturbance.

Our patient improved after hemodialysis, but developed renal failure, which is often associated with metformin-induced lactic acidosis. After two months of continued outpatient dialysis, her renal function returned.

Back to Top | Article Outline


In another case, a 47-year-old woman was brought to the emergency department by her family six hours after ingesting a cola drink mixed with “rat poison” in a suicide attempt. She could not provide any further information about the rodenticide except for the fact that it had been purchased many years before in Mexico. She complained of abdominal pain, severe nausea, and vomiting. On examination she was afebrile, with a pulse rate of 112 beats per minute and a blood pressure of 155/108 mmHg. She was actively vomiting up black liquid that was described as having a “fishy” odor. Abdominal exam revealed mild diffuse tenderness, and her skin was dry. The remainder of the examination was unremarkable.

Laboratory results included a white blood count of 14.2 per mm3, hemoglobin of 17 gm/dL, and an anion gap of 16. Blood salicylate and acetaminophen levels were both negative. Electrocardiogram revealed sinus tachycardia with normal intervals. Chest x-ray was clear. Arterial blood gas showed a pH of 7.47, a bicarbonate of 34 mEq/L, and a lactate of 5.8 mEq/L.

Because of the presentation and the black-colored “fishy” emesis, the toxicology consultant suspected that the rat poison ingested was zinc phosphide. A family member was sent home to retrieve the product package, which confirmed that suspicion. The patient remained stable throughout two days of inpatient observation, and was discharged after evaluation by the psychiatry service.

Back to Top | Article Outline


In this case the toxicologist did a great job of playing Sherlock Holmes, piecing together disparate bits of the clinical history and presentation to arrive at the correct diagnosis. Several years ago in this column, I reviewed the agents to consider when a patient presents with ingestion of an otherwise unidentified rat poison. The mnemonic for this list, in slightly expanded form, appears in table 2. The original mnemonic was the apt phrase RATS PANIC. I have added a “B” to include barium and bromethalin so perhaps the new phrase should be “BRATS PANIC,” or for fans of The Simpsons, BART PANICS. By far the most common class of rodenticide involved in suicidal or unintentional ingestions are the comedian-like anticoagulants, such as brodifacoum. As in this case, however, unusual ingestions can occur.

Table 2

Table 2

The fact that the rat poison was not purchased in the United States suggested that it might be something out of the ordinary. Zinc phosphide has been used in the U.S. since the 1930s, and became quite common during World War II when shortages of strychnine and red squill developed. Since then it has grown out of favor as safer agents became available. It is, however, cheap and effective, and is still widely used in developing countries.

Zinc phosphide is a crystalline, dark grey powder that is mixed with rat bait such as wheat, oats, bread, or sugar. Like many heavy metal salts, it is corrosive and causes marked gastrointestinal irritation. In addition, when it comes into contact with water, it gives off phosphine gas. (See table 3.) Phosphine is a colorless gas with a disagreeable odor that has been compared with garlic or rotten fish. Once released, phosphine is absorbed into the system where it inhibits mitochondrial cytochrome C oxidase, blocking oxidative phosphorylation and impairing energy production at the cellular level. It also produces reactive oxide radicals, which steal electrons from lipids and damage cell membranes.

Table 3

Table 3

Symptom onset is usually fairly rapid after ingestion, but has been reported (without any good documentation I could find) to occasionally be delayed for up to 18 hours. Typical initial symptoms are caused by the corrosive properties of zinc phosphide, and include retrosternal burning, epigastric pain, profuse vomiting, and hypotension. The emesis is very dark or black, as was the case here, and frequently has a distinct odor. Hours later, shock and pulmonary edema can supervene, as the patient becomes hypovolemic from GI losses and the toxin causes myocardial depression.

Treatment is supportive. Because the patient is generally vomiting profusely, gastric lavage is not recommended. Some authors suggest that activated charcoal would be beneficial, but vomiting will most likely interfere with its administration, and there is no evidence that charcoal is effective or even feasible in this situation. Various antioxidants such as glutathione, vitamin C, and beta-carotene also have been mentioned as possible antidotes, again without any evidence.

Three other points about zinc phosphide and phosphine gas:

  • ▪Zinc phosphide becomes less potent as a rat killer if it is old or has been exposed to moisture in the air. That may explain why this patient presented with GI irritation but no more serious manifestations of zinc phosphide and phosphine toxicity.
  • ▪Phosphine can be released in clandestine laboratories that make methamphetamine using the ephedrine/hydriotic acid/red phosphorous method. Drug enforcement officials who enter these makeshift labs should take care and use appropriate personal protective equipment, which includes either a self-contained breathing apparatus or phosphine-specific gas mask.
  • ▪Patients who have ingested zinc phosphide (or the closely related rodenticide aluminum phosphide) are hazmat incidents waiting to happen. They should at minimum be decontaminated by having all clothing and jewelry removed before being brought into the emergency department, and should be treated in a well-ventilated area.
Back to Top | Article Outline


An emergency medicine resident doing her month of rotation on the toxicology service was contacted by a classmate who was treating an animal handler at the local zoo. He had been bitten on the right wrist by a slow loris (Nycticebus coucang), a small prosimian mammal native to Southeast Asia. The zoo worker knew that northern Thai folklore held that the slow loris was venomous, but the resident believed that the only mammal that produced venom was the platypus. The service was asked: What is the real story? Should the treating physician anticipate any toxic signs or symptoms? How should these wounds be treated?

The toxicology service surveyed the entire world's medical literature concerning the bite of the slow loris. This was a fairly simple procedure because the literature consists of exactly one paper. (Amer J Trop Med Hyg 1972;21:592.) The paper described the case of a 43-year-old physician who developed anaphylaxis with acute circulatory collapse within minutes after having been bitten by his pet slow loris.

Initial symptoms were burning in the hands and feet followed by severe throbbing backache and a feeling of being very cold. He became hypotensive but had no bronchospasm. He improved after treatment with fluids, epinephrine, diphenhydramine, and opiate analgesics. Steroids were not given. Laboratory examination revealed hematuria with red blood cell casts. The author later contacted an expert at the Chiang Mai Zoo in northern Thailand, who noted that although the bite of the slow loris is often extremely painful, he knew of just one anecdotal case of a death apparently after such a bite.

Whether the slow loris was actually venomous remained a point of controversy until 2003, when researchers at Columbia University in New York isolated a protein from fluid secreted by sebaceous glands located at a hairless area on the inside of the animal's elbows. This protein consists of two amino-acid chains linked by disulfide bonds. Analysis revealed that the protein was remarkably similar to the major allergen of the domestic cat. When attacked, the loris licks its elbow area, mixing the venom with saliva. Contact with saliva may actually activate the venom, which is delivered into the wound by a grove on incisor-like front teeth. Interestingly, the loris also will lick the venom onto the skin of their young before leaving them alone to search for food.

Because anaphylactic manifestations did not occur after the zoo worker was observed for several hours, the toxicology service recommended that the wound be handled in a similar fashion to cat bite, with local care, tetanus prophylaxis if indicated, analgesics, and careful monitoring for infection. So there are venomous mammals other than the platypus. Aside from the loris, several species of shrews and Solenodons (rat-like creatures) also produce distinct toxins.

Back to Top | Article Outline


Because this article started with two mnemonics, I will relate my favorite mnemonic of all time. It was devised by an unheralded first-year Chicago medical student in the mid-1970s, who had obviously spent too many late nights studying anatomy. It is: Frank Sinatra Takes Four-Fifths Seagrams Seven Each Night To Ease Tensions. This mnemonic helps remember the ribs, in order, from cephalad to caudad: First, Second, Third, Fourth, Fifth, Sixth, Seventh, Eighth, Ninth, Tenth, Eleventh, Twelfth.

© 2007 Lippincott Williams & Wilkins, Inc.