In 2021 the FDA investigated something it had never investigated before: a multistate Salmonella outbreak traced to leafy greens that had never touched soil, a field, or an irrigation canal. They were grown indoors, hydroponically, in a controlled environment operation in Illinois. Thirty one people fell ill across four states, four of them badly enough to be hospitalized.
Investigators took twenty five samples at the farm. The outbreak strain turned up in exactly one of them: a stormwater drainage basin on the property, outside, next to the building.
That detail is why this article exists. A countertop hydroponic garden is sold as an escape from the lettuce supply chain, and it does escape a real part of it. The useful question is which part, and what comes indoors along with the device.
The same failure, twice, at two addresses
Three years before the Illinois case, romaine from Yuma, Arizona caused the worst E. coli O157:H7 outbreak in the United States in over a decade. FDA’s environmental assessment found the outbreak strain, matched by whole genome sequencing, in three water samples drawn from a 3.5 mile stretch of an irrigation canal near Wellton. A large concentrated animal feeding operation sits next to that stretch. FDA looked for a route from the feedlot to the canal and did not find one, which is worth stating plainly, because the story is usually told as though they did.
What FDA did conclude was narrower and more interesting: the most likely route into the lettuce was the canal water itself, because the outbreak strain was found in the canal and nowhere else that was sampled.
| Yuma 2018, open field | Illinois 2021, indoor hydroponic | |
|---|---|---|
| Pathogen | E. coli O157:H7 | Salmonella Typhimurium |
| Illnesses | 210 | 31 |
| States | 36 | 4 |
| Hospitalized | 96 | 4 |
| Deaths | 5 | 0 |
| Where the strain was found | Irrigation canal beside the fields | Stormwater basin beside the building |
| Route proven | No | No |
Moving the farm indoors and taking away the soil did not remove the failure mode. It changed the address. A canal became a drainage basin, and a field became a growing pond inside a building which, as FDA noted at the time, the firm did not treat as agricultural water at all. Investigators also recorded condensate dripping from a chiller line inside the building and packaged soilless growing media stored outdoors, unprotected. The farm’s own root cause analysis never established how the pathogen got in.
What a countertop garden genuinely removes
Look at the table again, at the two rows that differ most: 210 against 31, and 36 states against 4. Both farms had a pathogen. Only one of them had a mechanism for turning that pathogen into a national event.
An outbreak is contamination multiplied. One contaminated lot enters a shared wash tank, and the water carries it onto everything else in the tank, which is the whole argument of wash water sanitizers. Then it is bagged, dated and shipped to thirty six states. Your kitchen has none of that. No lot, no shared tank, no distribution, no gap of two weeks between harvest and the fork.
There are no documented outbreaks linked to consumer countertop gardens. Read that sentence exactly as written. Single household illnesses are almost never investigated and almost never traced, so the absence of case reports is not evidence of safety. It does mean the specific thing frightening you out of the salad aisle, the multistate recall, structurally cannot happen on your counter.
The fear that does not survive the literature
The standard objection to growing lettuce in water goes like this: the roots sit in a nutrient solution around the clock, so anything living in that solution gets pulled up into the leaf, where no amount of washing can reach it.
The experimental record does not support it. Three studies, eighteen years apart, point the same way:
- A 2025 study in Microorganisms compared five E. coli O157:H7 strains in soil and in hydroponics, and reported soil cultivation to be more susceptible to internalization than hydroponics.
- A 2022 study in Food Microbiology kept live Salmonella Enteritidis in the nutrient solution and found no leaf internalization at any growth stage of the lettuce.
- A 2007 study using fluorescent strains got internal colonization in the hydroponic system only at very high cell densities and only with one strain, while the soil system produced internal counts of 3.95, 2.57 and 2.37 log CFU/g with all three.
So water culture is not the weak point here. Soil was equal or worse every time it was compared directly.
The 2007 authors did leave a warning, and it is aimed at everyone, not at hydroponics: once a pathogen is inside leaf tissue, it sits “in locations where they are unlikely to be removed by the actions of consumer washing”. That is the ceiling on every soak, spray and gadget in this cluster, and the reason triple washed claims promise less than they sound like they do.
The dirtiest part of the machine is the pod
A 2020 survey of a commercial closed hydroponic system measured every component it could reach. The highest microbial load in the entire system was not the water and not the roots. It was the peat moss plug: aerobic plate count 6.8 log CFU per gram, coliforms 4.5 log MPN per gram, yeast and mold above 4.8. Roots embedded in the plugs came in nearly as high. The other hot spot in any hydroponic setup is the slimy film on reservoir walls and tubing, which chlorine clears for a few weeks at best; the wiki explains why in its article on biofilms in hydroponic systems.
The number that matters most is the last one. Harvested lettuce leaves carried 4.1 log CFU per gram against 1.7 for leaves sampled before harvest. The leaves picked up microbes from the root ball at the moment of cutting.
Two honest limits on that study before anyone panics. These are indicator organisms, general microbial load, not pathogens. And the authors specifically looked for Listeria and found none in any sample. High counts in peat are ordinary. Peat is decomposed plant matter, not a clean room. Who actually lives in a hydroponic system, and why sterility is not the goal there, is explained in the wiki on microbiology in hydroponics.
Still, this is the one finding that transfers directly to the appliance on your counter, because a countertop garden is a plug sitting in warm water under a lamp for five weeks. FDA noticed the same category of risk at the Illinois farm, where the growing media was stored outside without protection. Which gives three practical habits worth more than any accessory:
- Cut leaves with clean scissors above the pod. Do not uproot the plant and rinse the root ball over your salad bowl.
- Keep the unit away from the zone where you handle raw meat and poultry, and wash your hands before topping up the reservoir. Your hands are the most likely route in.
- Empty and wipe the reservoir between crops rather than topping it up forever. This is ordinary hygiene, not disinfection theater.
What a head of lettuce costs on your counter
Click & Grow sells the Smart Garden 3 at $124.95 and the nine pod model at $249.95, rising to $949.95 for the twenty five pod unit. Lettuce pods are $14.95 for three, which is close to five dollars per plant in consumables alone, before the machine and before the lamp that runs sixteen hours a day.
The same company describes its produce as “even better than organic produce”. On food safety grounds that claim is unsupported. The plant is grown in a non sterile plug, in an open vessel, in a kitchen, handled by hands that also handle everything else in that kitchen. What it is, defensibly, is lettuce with a very short supply chain. When the plug in the device is made of sponge or foam, biofilm builds up in it even more readily, and the wiki on hydroponic foam and sponge substrates explains where that material makes sense and where it does not.
There is also a continuity problem specific to pod based hardware. AeroGarden, the brand most people name first, was owned by Scotts Miracle-Gro from 2020 until the fall of 2024, when the company announced it was winding down a twenty year old operation. Manufacturing stopped at the end of that year. A relaunch was announced for spring 2025. A device that only eats one supplier’s cartridges inherits that supplier’s luck.
Bottom line
If you enjoy the thing, buy it. Fresh leaves cut ten seconds before eating are genuinely good, and you have removed the canal, the wash tank and the two week journey in one purchase. That is a real reduction in a real risk.
What you have not bought is sterility, and roughly five dollars a head is a steep price for a risk reduction you can also get from a pot of leaf lettuce on a windowsill and a packet of seed.
And if the reason you are reading this is pregnancy, chemotherapy, an immunosuppressant or an elderly parent, the answer this whole food safety cluster keeps arriving at has not changed. No device in it, this one included, gets you to the certainty that heat gets you to. Cooked greens are the answer for the people with the most to lose, and the countertop garden is a pleasant thing to own, not a medical control.
