The FDA number is 40°F (4.4°C) or below. Most people set the dial once, when the appliance is delivered, and never touch it again. In 2024 a group at Wageningen University published a study in the International Journal of Food Microbiology that measured what household refrigerators actually do, and the result says the dial setting is close to irrelevant.
They surveyed 1,020 consumers. Of those, 534 actually measured the temperature inside their own refrigerator. The average came out at 42.3°F (5.7°C), with a standard deviation of 4.0°F (2.2°C), and the warmest reading in the set was 62.6°F (17°C).
The average is the least interesting number in that paper. The important part came from a separate set of fifty refrigerators that were logged continuously for 24 hours.
A single refrigerator is two different refrigerators
The 24-hour profiles found a gap that no thermostat setting will reveal. The top shelf averaged 45.9°F (7.7°C). The bottom shelf averaged 42.3°F (5.7°C). That is a statistically significant spread of roughly 3.6°F (2°C) inside the same cabinet, on the same setting, at the same moment.
A 2023 Korean study mapped the same problem by zone instead of by shelf. Door storage averaged 41.1°F (5.06°C), the area along the back wall 39.5°F (4.18°C), and the interior storage bin 38.1°F (3.41°C). The maximum spread between locations within one refrigerator reached 3.9°F (2.18°C), and the extremes across the whole data set ran from 17.2°F (minus 8.2°C) to 60.4°F (15.8°C).
The two studies disagree about where the warm pocket sits. The Dutch data put it on the top shelf, the Korean data on the door. What they agree on matters more than the location: the internal spread is around 2 degrees Celsius, which is the entire distance between the recommended limit and the temperature at which risk climbs sharply.
The Korean work also supplies the mechanism. The more heavily a refrigerator is loaded, the colder the back-wall shelves become and the warmer the door shelves get, because the food itself obstructs cold air on its way to the door side. The warm pocket is partly built by whoever put the groceries away.
Where the 40°F line comes from
That threshold exists because of one organism. Listeria monocytogenes grows across a range of 31.3°F to 113°F (minus 0.4°C to 45°C), with an optimum at 98.6°F (37°C). Refrigeration does not stop it. Refrigeration slows the clock.
This is where household intuition and microbiology part company. The bacteria that make food smell wrong and look slimy mostly stall in the cold, which is why chilled food seems fresh for a long time. Listeria keeps growing quietly, with no odor, no discoloration, and no visible signal of any kind. The CDC estimates about 1,600 illnesses, more than 1,500 hospitalizations and 260 deaths in the United States each year, which makes listeriosis the third leading cause of death from foodborne illness. The ratio of hospitalizations to illnesses is the detail worth noticing: this infection does not come in a mild version.
Listeria is not alone in this. The same logic shows up in fish botulism, where type E grows from 37.9°F (3.3°C), so refrigeration fails there too and the hazard turns out to live in the construction of the product rather than in its storage temperature. That case is covered separately in the piece on dried fish and botulism. Cold buys time. It does not build a wall.
Food that nobody will heat again
The risk concentrates in one category: ready-to-eat products that go from the package to the plate. Deli meat, hot dogs, pate, prepared salads, soft cheeses, smoked fish. They come out of the refrigerator and get eaten, and there is no step in between that would kill Listeria.
European Regulation 2073/2005 sets a limit of 100 colony forming units per gram for ready-to-eat foods capable of supporting Listeria growth, plus a stricter “not detected in 25 g” requirement before the product leaves the producer’s control. Amendment 2024/2895, published in November 2024, moves that “not detected in 25 g” requirement onto the entire shelf life of the product on the market as of July 1, 2026. Until then it applied at the factory gate, and whatever happened to the package afterward fell outside the criterion.
The Dutch paper modeled exactly what home storage does to that limit. Its quantitative microbiological risk assessment identified the dominant risk factors as the initial concentration in the product plus the time and temperature of storage at home. Then comes the number that justifies the whole exercise.
Scenario analysis found that storing opened ready-to-eat cooked meat at home for fewer than 7 days or below 44.6°F (7°C) cut predicted illness cases by more than 80%. The conjunction is “or”. One lever is enough to get most of that reduction, which is useful for anyone whose refrigerator will not go colder.
Who this is actually addressed to
Two findings from the same study belong side by side. Refrigerators belonging to people aged 65 and over ran an average of 1.1°F (0.6°C) warmer than those of people 35 and under. And adults over 65 account for nearly 90% of all listeriosis cases in the model, and 59% of the years of life lost.
The warmest refrigerators and the most vulnerable population are the same households. The authors draw a narrow conclusion from that: targeted messaging aimed at older adults, storage of ready-to-eat meat on the bottom or middle shelf, and consumption within two to three days of opening the package. Pregnancy belongs in the same bracket for the same reason.
For leafy greens, 40°F is already a loss
The same dial has a second job, and that job wants a different number. Lettuce holds best at 32°F (0°C): iceberg keeps for 21 to 28 days there. Raise it to 41°F (5°C) and the window drops to 14 days. Romaine behaves the same way, 21 days against 14. At 68°F (20°C) the respiration rate of lettuce runs 4 to 6 times higher than at freezing. Those figures are worked through in the article on lettuce shelf life after harvest.
Stack the numbers up. Greens want 32°F. The safety ceiling sits at 40°F. The average measured refrigerator holds 42.3°F. Its top shelf holds 45.9°F. So 40°F is a ceiling rather than a target, and for leafy greens that ceiling already costs half the shelf life.
Greens also carry a seasonal layer. In summer, field temperature, transport temperature and refrigerator temperature all drift upward together, and the resulting summer peak in produce-linked illness is a measured phenomenon rather than an impression. The reasons behind that peak are examined in the piece on why summer is outbreak season. The home refrigerator is the last link in a long chain, and the only one a shopper controls.
What to do about it
- Measure it instead of trusting the dial. The cheapest appliance thermometer, left inside for a full day, tells you more than the setting does. Put it on the shelf where ready-to-eat food actually sits.
- Measure more than one spot. Top shelf, bottom shelf, door shelf. The 2 degree spread becomes visible immediately.
- Move ready-to-eat food to the coldest zone. Deli meat, soft cheese and prepared salads go low or middle, away from the door. The top shelf and the door are for things that will be cooked later anyway.
- Finish an opened package in two to three days. That is the study authors’ own recommendation, not a general suggestion.
- Do not pack the cabinet solid. A dense load blocks cold air from reaching the door shelves and manufactures the warm pocket.
- Keep greens separate and lower. They want 32°F, and they are the one item that benefits from the coldest zone.
- If the household includes someone over 65 or pregnant, none of this is general hygiene any more. It is a specific task with a known payoff.
The short version
The thermostat describes one point in the cabinet while the food occupies seven. Logged profiles show a spread of roughly 2 degrees Celsius: 45.9°F on the top shelf against 42.3°F on the bottom in the Dutch data, up to 3.9°F between zones in the Korean data. A correctly set dial therefore guarantees nothing to the package sitting in the door.
The 40°F line exists because Listeria grows from 31.3°F upward and gives no sensory warning at all. Cold slows it rather than killing it, which is why temperature always travels with time. Both levers work: fewer than 7 days after opening, or below 44.6°F, delivers over 80% of the available risk reduction.
In practice it is three actions. Put a thermometer inside and read the real number. Move ready-to-eat food off the door and off the top shelf. Finish opened deli meat in two or three days rather than in a week.
The worst case for that two degree spread is food that reached the kitchen with no cold chain at all: raw milk and the test that is not performed.
A separate question is the list of foods that supposedly must never go in the fridge. One entry was retired by the agency that wrote it, and the rest charge different prices: flavor, texture or safety.
