In the fall of 1996, Americans started getting sick from fresh apple juice. Not from a farm stand barrel, but from a bottle in a supermarket cooler, made by a company that sold freshness as its whole proposition and refused on principle to pasteurize.
The final count was 70 people infected with E. coli O157:H7, 25 of them hospitalized, 14 with hemolytic uremic syndrome, and one death, a 16-month-old girl in Colorado.
What grew out of that outbreak was the first criminal conviction of a company in a large-scale American foodborne outbreak, and two rules that still define what safe juice means on a store shelf.
From sick people to one production date
On October 30, 1996, the Seattle-King County Department of Public Health and the Washington State Department of Health reported an outbreak of E. coli O157:H7 linked to unpasteurized Odwalla brand apple juice. A case was defined as HUS, or a stool culture yielding O157:H7, in a person who got sick after September 30 and had drunk the juice within ten days of onset.
Washington state investigators used DNA fingerprinting to assemble a cluster of 15 related cases in which every patient named the same juice. On October 31 the company began a voluntary recall of every product containing apple juice. Distribution at that point covered ten jurisdictions: British Columbia, California, Colorado, Idaho, Montana, Nevada, New Mexico, Oregon, Texas and Washington.
The hard evidence arrived on November 5. The FDA laboratory in Seattle tested samples pulled from the company’s warehouse in Washington state, and a sample of juice produced around October 7 grew E. coli O157:H7. Separately, the pathogen was cultured from a previously unopened bottle, and its genetic fingerprint was indistinguishable from the patient isolates.
The numbers moved as the investigation went on. The interim report of November 6, 1996 listed 45 cases across four states and one province, a median patient age of five years, 12 cases of HUS and no deaths. The peer-reviewed accounting came out two and a half years later with the final figures: 70 cases, 25 hospitalizations, 14 cases of HUS, one death. Every outbreak investigation works this way, and it is usually the first, incomplete snapshot that survives in retellings.
Where the E. coli came from
Almost every case traced back to a single production date, October 7, which narrowed the search to the fruit that came in that day. The authors of the final report identified incoming apples as the suspected source, and three lots could account for the contamination: two came from an orchard frequented by deer that were later shown to carry O157:H7, and one contained decayed apples that had been waxed.
The FDA never pinned down the exact mechanism either, and left three theories standing: apples contaminated with animal feces, contaminated wooden shipping crates, and employees who failed to wash their hands properly.
The conclusion of the report is the uncomfortable part for anyone who trusts washing: standard procedures at a state-of-the-art plant producing unpasteurized juice were not enough to eliminate E. coli O157:H7. Washing, brushing and a sanitizer reduce the cell count on the surface of the fruit, and none of them works as a kill step. For this organism that distinction decides everything, because the infectious dose is measured in tens of cells, not millions.
A decade later the same logic played out in leafy greens: wild pigs near the field, the outbreak strain in the soil, and no broken rule anywhere. That case is in our account of the 2006 spinach outbreak.
What investigators found in the company’s own records
Then came the part that turned this into a criminal case. The FDA ran a 14-month investigation, and its Office of Criminal Investigations interviewed former employees, suppliers and others familiar with company operations. Four things surfaced.
First: in June 1996 the US Army walked away from a contract with the company after its own analysis found a high bacterial count in one of the products. That was four months before the outbreak.
Second: a consultant the company hired ran weekly microbiological tests from May 1994 through December 1995 and kept finding Listeria and other bacterial contamination in the plant. The company started out trying to find the source, then shifted its attention to extending shelf life and never resolved the contamination.
Third: accepted industry practice called for washing and sanitizing fruit with a chlorine solution. The company had dropped chlorine in favor of phosphoric acid, which may not be effective as a fruit wash.
Fourth, recorded by the FDA investigator who inspected the plant: product testing was built around shelf life rather than bacterial contamination. Records also showed the company had been buying an inferior grade of apples.
The mechanism here is the one that shows up in every later prosecution: the charges grew out of the company’s own paperwork. Three American cases where that ended in prison terms for executives are covered separately in our piece on food safety prison sentences.
The sentence: $1.5 million and five years of supervision
In July 1998 the company pleaded guilty to 16 misdemeanor counts of delivering adulterated food products and agreed to pay $1.5 million. The plea agreement was filed in the US District Court for the Eastern District of California, and the case was litigated out of Fresno.
The assistant US attorney called it the first criminal prosecution of an epidemic of this size and the largest criminal settlement of a food injury case in FDA history. The company drew five years of court-supervised probation and an obligation to implement a HACCP plan: sanitary conditions, a written sanitation control program run by a qualified manager, and employee training.
The money split this way: $250,000 went to the victims’ organization Safe Tables Our Priority and to the food safety research centers at two universities, and the rest went to the US Treasury. Two months before the plea, in a separate civil case, the company had agreed to pay up to $15 million to the families of five children with severe kidney damage. Installing flash pasteurization cost it more than $1 million.
A follow-up inspection in June 1997 came back clean: juice samples and equipment swabs were negative.
The two rules that came out of it
On July 8, 1998, two weeks before the plea, the FDA issued its juice labeling rule. Unpasteurized packaged juice has to carry this statement: “WARNING: This product has not been pasteurized and, therefore, may contain harmful bacteria that can cause serious illness in children, the elderly, and persons with weakened immune systems.”
On January 19, 2001, the juice HACCP rule followed. It requires juice processors to build in controls that consistently deliver a 5-log reduction of the pertinent pathogen, a hundred-thousand-fold cut, and to hold that result for at least the shelf life of the product. The rule took effect January 22, 2002, became binding on small businesses January 21, 2003, and on very small businesses January 20, 2004.
Pasteurization was never made mandatory. The agency wrote a performance standard instead of a technology and named the alternatives directly: UV irradiation, pulsed light, high pressure processing. For apple juice the FDA knew of no way at the time to reach 5 logs without a kill step applied to the extracted juice. The underlying logic is the same one described in our HACCP guide for microgreens: identify the hazard, put a control on it, then prove the control works.
One gap was left open deliberately. Operating a retail establishment does not count as processing, so juice squeezed and sold on the premises falls outside juice HACCP. And juice sold by the glass, at an orchard, a farmers market, a juice bar or a restaurant, needs no warning statement at all.
What this means if you drink or make fresh juice
First: fresh-squeezed juice has no kill step by definition. Whatever was on the fruit, and whatever was inside a damaged piece of fruit, rides into the glass. A juice bar, a home juicer and a jug of cider at a fall market all sit in the same position.
Second: the risk is not distributed evenly. The FDA names four groups for whom untreated juice is a separate matter: children, older adults, pregnant women, and people with weakened immune systems. In the 1996 outbreak the median patient was five years old, and the person who died was a toddler.
Third, and most practical: the fruit contributes more risk than the kitchen does. Windfalls, bruised fruit and softening apples sit next to a contaminated lot in this case for a reason. Cut away damaged areas, throw out anything rotten, wash under running water and dry the fruit afterward. That lowers the cell count without becoming a kill step.
Fourth: if the juice is for a child or for anyone in those risk groups, buy it pasteurized or treated by an approved nonthermal process. In a store the label tells you, since unpasteurized packaged juice must carry the warning. Juice sold by the glass carries nothing, so there the question has to be asked out loud.
The brand itself outlived the story. Coca-Cola bought the company in 2001 and shut it down in the summer of 2020, and in early 2025 the Mexican beverage company Grupo Jumex brought the label back to the US market. The 5-log rule has its own life and applies to everyone.
How the same principles play out for greens, sprouts and microgreens is collected in our food safety guide.
