Sodium Azide Poisoning at a Restaurant—Dallas County, Texas, 2010 FREE

MMWR. 2012;25:457-460.

In April 2010, Dallas County Health and Human Services (DCHHS) staff members investigated reports of acute-onset dizziness among patrons in a local restaurant. Symptoms, which included fainting resulting from low blood pressure, occurred within minutes of consuming food from the restaurant and were consistent with chemical poisoning. Toxicologic and epidemiologic investigations were begun to determine the cause of the poisonings and identify potentially exposed persons. This report summarizes the results of those investigations, including a case-control study that identified iced tea as the likely contaminated food or drink (odds ratio [OR] = 65; 95% confidence interval [CI] = 2.4-3,292). Approximately 5 months after the incident, extensive laboratory testing identified sodium azide (NaN₃) and hydrazoic acid (formed when sodium azide contacts water) as the toxic agents in the iced tea. All five ill restaurant patrons recovered from their symptoms. For rapid-onset foodborne illnesses, chemical poisons should be considered as a potential cause, regardless of negative initial toxicologic screening tests. Although unusual chemicals can be challenging to detect, a multidisciplinary approach involving public health officials and forensic and medical toxicologists can lead to appropriate testing. In the absence of an identified agent, epidemiologic tools are valuable for active case-finding and confirming suspected contaminated food vehicles.

In April 2010, city fire and police departments notified a city health department in Texas of four persons transported from a restaurant to an emergency department (ED) over a period of 4 hours, after developing similar symptoms minutes after consuming iced tea from paper cups filled from a self-serve urn. Symptoms included lightheadedness, nausea, and diaphoresis (Table). The city fire department, responding to emergency medical services calls, contacted the police and the city environmental health departments after a third call came from the same restaurant. These agencies conducted a site visit that evening and temporarily suspended further sale of the iced tea. Samples of the iced tea were collected by authorities from restaurant tea urns and from two patients who brought their paper cups of iced tea to the ED.

Laboratory and radiographic diagnostic tests performed in the ED did not reveal a cause for the illness; none of the four patients were acidotic. Toxicologic evaluation in the ED, consisting of serum ethanol concentrations and urine drug screens, was nondiagnostic. All four patients received supportive treatment and were discharged either that night or after an overnight hospitalization. Upon learning of the cases, DCHHS epidemiology and environmental health staff members obtained and reviewed medical records, interviewed the patients and restaurant staff members, and conducted a site visit. Because of a high suspicion for chemical etiology, DCHHS staff members consulted with multiple agencies, including the Dallas County Southwestern Institute of Forensic Sciences (SWIFS), Food and Drug Administration, Texas Department of State Health Services, and medical toxicologists at the University of Texas Southwestern School of Medicine (UTSW) and the North Texas Poison Center, regarding testing for possible etiologic agents. A comprehensive list of potential agents was developed by medical toxicologists; the list included metabolic inhibitors (e.g., cyanide and sodium azide), heavy metals (e.g., arsenic), antihypertensives, organophosphates, and sedatives. Toxicologists at SWIFS and UTSW assisted in identification of laboratories capable of conducting appropriate testing.

To determine whether other cases existed, the Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE) database was searched for patients with similar symptoms visiting any of 18 area EDs on the same day and during the same timeframe as the four patients. Of 1,827 ED visits recorded in ESSENCE that day, 81 patients reported symptoms of altered mental status, dizziness, fainting, tingling, or hypotension. Records of these cases were reviewed by DCHHS staff members and excluded if alternate diagnoses were well documented. Of 29 cases selected for further review, one additional case belonging to the restaurant cluster was identified. The patient initially was taken by relatives to the same hospital as the other four patients and registered a chief symptom of “tingling” at triage, but left the ED before evaluation by a physician. The patient was then examined at an outpatient clinic and referred to a different ED because of hypotension and arrhythmia.

The five patients visited the restaurant separately, over a 4.5-hour period, beginning in the afternoon. All five patients reported that iced tea was the only beverage they consumed at the restaurant; the amount of consumption ranged from two sips to 8 ounces. All of the tea came from a single self-serve urn that was in service during the period the five visited the restaurant. Regarding the food items in common, two reported eating a cookie and three reported eating bites of a sandwich. The five patients ranged in age from 32 to 52 years; three were women. Among the five, comorbidities included hypertension, anemia, hypothyroidism, hyperlipidemia, and a bicuspid aortic valve condition. Their initial heart rates at examination ranged from 62 bpm to 127 bpm; blood pressures ranged from 84/42 mmHg to 110/68 mmHg. Three of the five were discharged from the ED, and two were hospitalized (Table).

Because no contaminated vehicle was confirmed immediately, a case-control study was conducted to assess the association of the illnesses with specific food and drink. Potential controls were identified among the restaurant patrons and contacted by using records of credit card transactions. A case-patient was defined as a restaurant patron reporting dizziness or fainting within a 6-hour period that encompassed the time of symptom onset for the five known patients. Controls were patrons who purchased food or drink at the restaurant during the same timeframe and did not report dizziness or fainting after their meal. Thirteen of the 14 controls said they did not consume iced tea. The fourteenth control recalled drinking iced tea but having no symptoms; however, further investigation revealed that this person drank iced tea from a different self-serve urn, before the urn used by the five case-patients was placed in service. [Another person, who had been selected as a potential fifteenth control, reported feeling dizzy but did not drink iced tea. This person attributed the dizziness to a chronic condition. However, because the person met the case definition, for purposes of the case-control study this person was counted as a sixth case-patient.] The case-control study found that consuming iced tea was 65 times more likely among the case-patients than the controls (OR = 65.0; CI = 2.4-3,292).

Testing of iced tea samples by SWIFS was negative for approximately 100 different chemicals and drugs, except for caffeine. The following analyses were performed: gas chromatography/mass spectrometry to screen for drugs (including gamma-hydroxybutyric acid, diltiazem, labetalol, metoprolol, propranolol, and verapamil), atomic absorption spectrometry with hydride generation for arsenic analysis, chemical tests for cyanide, a Geiger counter scan for radioactivity, and gas chromatography to identify alcohols and other volatiles. Samples also were sent to a private toxicology laboratory to test for nitrites and nitrates; all of those samples were positive because nitrites and nitrates were components of the leaves used to brew the tea. SWIFS then referred tea samples to the Federal Bureau of Investigation (FBI).

Nearly 5 months after the restaurant incident, headspace gas chromatography/mass spectrometry analysis by the FBI laboratory detected hydrazoic acid, which is formed when sodium azide contacts water. The presence of hydrazoic acid led to the use of infrared spectroscopy to detect sodium azide. Both chemicals were found in three samples: two from patients' paper cups and one from the tea urn in use at the time of the incident. No sodium azide or hydrazoic acid was found in a control sample from a second tea urn used earlier that day. Neither chemical compound was used in the FBI laboratory or at the restaurant.

The mechanism by which the tea became contaminated with sodium azide remains unknown. Investigators learned that water for the tea and soda served in the restaurant came from a common line. The self-serve iced tea urns were kept in an open location in the restaurant, accessible by customers, and out of the direct line-of-sight of employees.

After receiving notification of sodium azide as the etiologic agent, all five patients were evaluated in the medical toxicology clinic at UTSW and were confirmed to have fully recovered, without clinical sequelae. The police investigation is now closed.

Reported by: Evan S. Schwarz, MD, Paul M. Wax, MD, Kurt C. Kleinschmidt, MD, Kapil Sharma, MD, Univ of Texas Southwestern School of Medicine and North Texas Poison Center; Elizabeth Todd, PhD, Erin Spargo, PhD, Dallas County Southwestern Institute of Forensic Sciences; Wendy M. Chung, MD, Gabriela Cantu, MPH, Mariama Janneh, MPH, Juan Rodriguez, MPH, Taye Derse, MD, Dallas County Dept of Health and Human Svcs. Joshua G. Schier, MD, Div of Environmental Hazards and Health Effects, National Center for Environmental Health, CDC. Corresponding contributor: Evan S. Schwarz, schwarze@wusm.wustl.edu, 314-747-3690.

Editorial Note: Sodium azide is an odorless, tasteless, water-soluble crystalline powder that has been used in the manufacture of automobile airbags and explosives, and as a laboratory preservative.^1- 2 In the 1950s, sodium azide was used to treat hypertension because of its profound vasodilatory effects, possibly resulting from the production of nitric oxide.^2- 3 Systemic absorption of sodium azide can result in hypotension, loss of consciousness, headache, dyspnea, nausea, vomiting, metabolic acidosis, dysrhythmias, and death. The mechanism of poisoning with sodium azide is similar to cyanide, which suggests that use of traditional cyanide antidotes such as nitrite therapy and sodium thiosulfate might be beneficial. However, experience with antidotes in humans is limited mainly to case reports, and none have shown a conclusive, convincing, or consistent benefit. Therefore, good supportive care remains the cornerstone of treatment.

Previous reports of sodium azide poisoning primarily are limited to case reports involving ingestion of sodium azide either by accident or during attempted suicide. Patients at a dialysis center became symptomatic when treatment filters preserved with sodium azide were not rinsed properly.⁴ Two foodborne sodium azide mass poisonings have been described briefly, but neither report included clinical or epidemiologic details regarding the victims or the investigations.^3,5 In 2009, six workers at Harvard Medical School became ill after drinking from a communal coffee pot that was contaminated with sodium azide; whether or not the poisoning was intentional is unknown.³ Within minutes of ingestion, the laboratory workers had palpitations and diaphoresis; two fainted. All of their symptoms rapidly resolved. In 1998, sodium azide was deliberately added to a teapot in Niigata, Japan, poisoning nine persons.⁵ No clinical data or other details about these two poisoning incidents have been reported.^3,5

The potential for future incidents and the challenges in identification of unusual chemicals reinforce the need to continue educating health-care providers regarding the possibility of such poisonings. Public health agencies should consider such incidents and the complexity of such investigations in their emergency response planning.^6- 7

John T. Carlo, MD, Joey Stringer, Tony Jenkins, Dallas County Dept of Health and Human Svcs; Richard T. Briley, Garland Health Dept; David Klein PhD, Linda Gaul PhD, Texas Dept of State Health Svcs. William H. Keene, PhD, Oregon Health Svcs. Larry Enmon, Food and Drug Administration. Marc A. LeBeau, PhD, Eileen M. Waninger, Federal Bureau of Investigation Laboratory.

Sodium azide is a toxic chemical used widely in industry that has been added to beverages, either intentionally or unintentionally, in foodborne poisoning incidents in the United States and Japan. However, few details of these cases and their investigations have been reported previously.

What is added by this report?

This is the first detailed report of sodium azide poisonings at a public venue and the subsequent investigation. All five patients experienced rapid-onset of similar symptoms, requiring emergency department evaluation; all recovered quickly after supportive care, without sequelae. A multi-agency investigation led to conclusive identification of the poison.

What are the implications for public health practice?

Chemical poisoning should be considered as a potential cause of any rapid-onset foodborne illness. However, unusual chemicals can be challenging to detect in food, drink, and biologic specimens with traditionally available screening tests. A multidisciplinary approach, including consultations with public health officials, forensic toxicologists, and medical toxicologists, can lead to successful referral for appropriate testing. In the absence of an identified etiologic agent, epidemiologic tools can be of value in active case-finding and in confirming contaminated food vehicles.