Across five Australian cities over eleven years, one simple thing held: the average temperature of a month predicted the salmonellosis notifications of the next one. Seasonal patterns in the notifications were fully explained by changes in temperature. The most useful part of that study is not the correlation, it is the lag: one month.
If heat acted in the kitchen, the cases would follow it the same week. A one-month lag says the temperature does its work earlier, back in the production chain. The authors say exactly that.
The summer peak is real and measured
A 2012 systematic review pulled together 86 studies spanning half a century across five enteric zoonoses. Every bacterial disease showed a distinct summer peak. The unevenness across the year was measured with the Gini coefficient: campylobacteriosis and salmonellosis came out identical at 0.22, and vero-cytotoxigenic E. coli higher at 0.36.
Parasites behave differently. Cryptosporidiosis produced two peaks, spring and summer, and the most marked variation of all at 0.39. Giardiasis rises only slightly in summer, at 0.18. Cyclospora is not in this review.
Why the seasonality of vero-cytotoxigenic E. coli is sharper than that of Salmonella, the review does not explain. It records the spread and notes that the pattern holds across national borders with regional variation. What matters in practice is that every bacterial pathogen peaks on the same side of the year.
The authors name two groups of causes: environmental effects on the pathogen and on its association with the host, and population characteristics and behavior. What follows takes them in turn.
German data give the same lag
In Munich, Berlin and three rural regions between 2001 and 2004, a rise in ambient temperature correlated positively with a rise in Salmonella and Campylobacter cases. The delay came out at roughly five weeks, the same figure as in Australia.
Two independent datasets, different hemispheres, different surveillance systems, and the same time shift. For epidemiology that is a strong argument that the outcome is decided before the store.
What heat does on a leaf
The clearest experiment was done on cilantro in 2002, using a strain of Salmonella Thompson taken from a real outbreak. At 72°F (22°C) the ordinary residents of the leaf, Pantoea agglomerans and Pseudomonas chlororaphis, held populations ten times larger than the pathogen. The leaf was already occupied, and Salmonella played a minor part on it.
At 86°F (30°C) the balance shifted. Salmonella reached significantly higher population levels and made up a noticeable share of the total culturable flora on the leaf. What the heat changed was the winner of the competition rather than the growth rate alone.
One more detail from the same work: Salmonella tolerated dryness at 60% relative humidity at least as well as the epiphytes did, and once the humidity came back it recovered to its maximum population. A hot dry afternoon does not clear the pathogen off, it holds it until nightfall.
Then the arithmetic of storage takes over: cut greens in a warm room add up to a log a day, and with a dose measured in single cells that is enough.
For Cyclospora, heat means something else
Cyclospora does not multiply on produce at all. An infected person sheds unsporulated oocysts and those are not infectious. To become infectious, an oocyst has to spend 7 to 15 days in the environment at 73 to 81°F (23 to 27°C). At 39°F (4°C) maturation drags out.
The window is narrow at both ends. In the laboratory, oocysts sporulated in 7 to 14 days at 72 and 86°F (22 and 30°C), while two weeks of storage at either 39°F (4°C) or 99°F (37°C) retarded sporulation. An oocyst turns infectious in mild warmth rather than in the peak of a heatwave.
So summer means growth for a bacterium and a maturation clock for a parasite. The mechanisms differ, the calendar coincides, and that is why both stories land in the same summer reports. Why chemistry does not touch this one is covered in our piece on chlorine and Cyclospora.
The inside of the leaf has a season too
In a two-year survey of iceberg lettuce, the share of microscope fields with Salmonella beneath the epidermis ranged from 0 to 100%, and the high values fell mainly in summer. How the bacteria get there, and why no rinse takes them out, is covered in our piece on internalization.
The second group of causes: us
The review’s second group is population characteristics and behavior. Summer changes what people eat, where they eat it, and how long the food stands in the warmth before it is served. The review gives no numbers to separate the contribution of the environment from the contribution of behavior, so saying which weighs more would be dishonest.
The one-month lag in two independent datasets does tilt it toward the first group, though. A behavioral cause would land faster: a Saturday barbecue produces cases that week, not four weeks later.
An honest correction: association is not yet cause
The Australian paper words its conclusion carefully: the strength of the association, its consistency across five cities and a plausible biological pathway together suggest that higher ambient temperatures are a cause. That is a statement about a weight of evidence rather than causation demonstrated in one experiment.
The authors see two practical uses. One: the month-long shift gives a basis for an early warning system, because this month’s temperature is already visible. Two, less cheerful: a warming climate means the summer regime lasts longer.
What it means for a shopper
The main point: whatever decided your risk happened a month before you bought the bag and a thousand kilometers from your kitchen. No washing routine compensates for that, and nothing at home reaches back into it.
Winter does not make the risk vanish either. A Gini of 0.22 describes unevenness across the year, not a season with zero outside the peak: outbreaks happen all year, there are simply more of them in summer.
What summer does justify is keeping your own stretch of the chain short. A bag of salad riding home in the heat sits in the warmest part of the car, and an hour in the boot at 86°F (30°C) costs more than it looks. After that it is the fridge and a clean knife, the part of the chain you control completely.
For a grower the same data point elsewhere: the decisive stretch is the hours between cutting and cold. The cooling step itself is not innocent either, and what it does to bacteria on the leaf is in the piece on internalization.
Key points
- The summer peak in bacterial enteric infections is measured: Gini 0.22 for salmonellosis and campylobacteriosis, 0.36 for vero-cytotoxigenic E. coli.
- Cases follow temperature with a one-month lag in Australia and about five weeks in Germany. The lag points at production rather than the kitchen.
- At 72°F (22°C) Salmonella on cilantro was outnumbered tenfold by ordinary leaf residents; at 86°F (30°C) it reached significantly higher levels.
- Dryness does not kill the pathogen: after 60% humidity it recovers to its maximum as soon as moisture returns.
- Cyclospora does not multiply in summer, it matures: 7 to 15 days at 73 to 81°F (23 to 27°C), and only then is it infectious.
- Internalization in iceberg lettuce also runs higher in summer.
- Your stretch of the chain is the drive home and the fridge. The rest was settled a month before you bought it.
How long thawed food can sit warm before refreezing is off the table is covered in our piece on refreezing. The wider risk picture is in our leafy greens food safety hub.
The last link in that chain sits in the kitchen, and it is the only one a shopper can measure directly: why the right number on the refrigerator dial guarantees nothing.
