Background. On October 20, 1992, >40 children from one elementary scho
ol visited the school nurse due to the acute onset of blue lips and ha
nds, vomiting, and headache during and after the school lunch periods.
Forty-nine children were seen by physicians that day and 14 were hosp
italized. Laboratory analysis revealed methemoglobinemia in many of th
e children. All recovered in 36 hours. Objective. A case-control study
was supplemented by environmental and laboratory investigations to de
termine the outbreak source. Methods. Cases were selected based on the
laboratory diagnosis of methemoglobinemia (methemoglobin level >2%).
Children whose methemoglobin levels were missing or <2% were excluded
from analysis. Controls were obtained by selecting every third name fr
om a school roster. The parents of 29 students who met the case defini
tion and 52 controls were interviewed. Results. All 29 cases and 33% (
17/52) of the controls ate soup during the school lunch (odds ratio un
defined, lower 95% confidence limit 16.1). Two pots of soup were prepa
red from ready-to-serve cans, which were diluted with water and enrich
ed with a commercially prepared flavor enhancer. The school's boiler,
dormant during the previous 5 months, was restarted on the morning of
the outbreak. The boiler also served as a tankless hot water heater. L
aboratory analysis of the soup identified abnormally high quantities o
f nitrite (459 ppm) and sodium metaborate, major components of the boi
ler water treatment solution. Undiluted soup from the same lot had 2.0
ppm nitrites; the flavor enhancer had 2.2 ppm nitrites. Nitrites were
present in the hot potable water system (4 to 10 ppm) and absent in t
he cold potable water system. Conclusions. This outbreak of methemoglo
binemia due to nitrite poisoning was traced to soup contaminated by ni
trites in a boiler additive. Nitrites are ubiquitous and potentially h
azardous inorganic ions. Extreme caution should be used when the possi
bility for toxic human exposure to nitrites exists.