The health evidence on gardening, and the hazards it carries

A raised bed, four tomato plants and a bag of potting mix is how most people start. Almost nobody starts with a soil test. In an older neighborhood, that is the one step that changes your risk.

Gardening is unusual among the activities people file under “nature.” Most of them have a single plausible mechanism and a stack of short, self-reported outcomes. Gardening has several mechanisms running at the same time: it is physical work, it changes what you eat, it puts you outside, it is frequently done near other people, and it creates a standing obligation that gets you out the door on days you would not have gone.

That is why the case for it is stronger than the case for most of its neighbors on the shelf. It is also why the hazards deserve a real section instead of a disclaimer. Urban soil carries lead. Bagged growing media has been tied to a serious lung infection in the United States. A puncture from a dirty trowel is a tetanus question, not a band-aid question. None of that argues against gardening. It argues for doing things in a sensible order on the first weekend.

Several mechanisms, not one

The reason gardening holds up better than, say, a nature photograph on a screen is that you can point to each pathway and test it separately.

It is measurable exercise. Researchers at Cardiff University put portable gas-analysis equipment on older allotment gardeners in Wales and measured oxygen consumption across six common tasks. Published in the Journal of Aging and Physical Activity in 2014, the study put those tasks between 1.9 and 5.7 METs. A MET is a multiple of the energy your body burns sitting still, so 1.9 is barely above resting and 5.7 is brisk work. The sample was small and British, so treat the numbers as a guide rather than a national average. The point holds anyway: some garden tasks are genuine moderate-intensity activity and others are not, and the difference is mostly whether your arms and back are loaded.

It changes diet. This is the pathway nothing else on the list has. A person who grows chard eats chard.

It is outdoors and daytime. Articles 733 and 722 own what that exposure does and what it costs you in sun and heat.

It is often social, and article 734 owns group activity.

It supplies a reason. Plants die if you ignore them. That sounds like a weak mechanism and behaves like a durable one, because it converts an intention into a deadline.

What the trials found, and what they did not

Here the evidence splits into two tiers, and conflating them is how gardening articles overpromise.

The broad literature is large and weak. A 2016 meta-analysis in Preventive Medicine Reports by Soga, Gaston and Yamaura pooled 22 studies published after 2001, covering 76 comparisons between gardeners and controls, with outcomes ranging from depression and anxiety scores to body mass index and sense of community. The pooled effect was positive. The authors also ran Egger’s test, a standard check for whether a field is burying its null results, and it came back positive for publication bias. The effect survived a statistical correction for that bias, which is reassuring but not the same as the bias not existing.

The stronger tier is small and specific. The Community Activation for Prevention trial, published in The Lancet Planetary Health in 2023 by Litt and colleagues, is the best-designed gardening study in the United States. Researchers recruited 291 adults from the waiting lists of Denver Urban Gardens in Denver and Aurora, Colorado, people who had not gardened in the previous two years, and randomly assigned 145 to receive a plot and 146 to stay on the waiting list. Observers were masked to who got what. Dietary recalls, seven-day accelerometry and body measurements were taken in spring, at harvest and the following winter.

The results are worth reading precisely. Fiber intake and moderate-to-vigorous physical activity both showed significant interactions between time and the intervention, with the gardening group gaining roughly 1.4 grams of fiber per day and about 6 extra minutes of moderate-to-vigorous activity per day at harvest. Perceived stress and anxiety fell more in the gardeners than the controls between spring and harvest. And then the null results, which the trial reports openly: no significant effect on combined fruit and vegetable intake, on overall diet quality as scored by the Healthy Eating Index, on sedentary time, on body mass index, or on waist circumference.

So a federally and philanthropically funded randomized trial of community gardening moved fiber, activity and mood, and did not move weight or total produce intake. Both halves of that sentence are the result, and the second half rarely makes the coverage.

A companion analysis in Scientific Reports in 2024 explains part of why. Of the 145 people assigned a plot, 97 gardened the whole season, a median of about 90 minutes a week, with a range running from zero to fourteen hours. Nearly three-quarters of the gardens held social events, and fewer than half of the intervention participants went to any. Gardening is not a pill. Dose varies enormously, and the social benefit people assume is automatic turns out to require showing up.

Start with the soil, because almost nobody does

Lead does not break down. It sits in the top few inches of soil for decades, and the two big deposits in American cities are exterior lead paint shed from houses built before 1978 and exhaust from leaded gasoline settling along roads. The Centers for Disease Control and Prevention lists both, plus industrial sources, and notes that children are exposed by swallowing or inhaling soil, tracking it indoors as dust, and eating produce grown in or near it.

Penn State Extension, in guidance updated in October 2026, puts soil near older lead-painted foundations in a range from roughly 150 to 10,000 milligrams per kilogram. That is a span of nearly two orders of magnitude between houses on the same street, which is exactly why guessing is useless.

For a reference point, the Environmental Protection Agency issued a residential lead directive on 16 October 2025 covering Superfund and RCRA cleanup sites, setting a regional screening level of 200 parts per million for lead in residential soil and a removal management level of 600. Those are decision thresholds inside federal cleanup programs, not a legal standard for your backyard, and that distinction matters. But they are the numbers extension services hang their practical advice on.

Here is how a test happens, because the mechanism is less intimidating than it sounds. The Cooperative Extension System was created by the Smith-Lever Act of 1914 and runs through more than 100 land-grant colleges and universities, with offices in or near almost all of the roughly 3,000 US counties, according to USDA’s National Institute of Food and Agriculture. Those universities run the soil labs. Penn State’s Agricultural Analytical Services Laboratory sells its standard soil test kit for $10, through county extension offices or direct by mail. The University of Massachusetts lab runs a separate total sorbed metals panel with a turnaround of ten to twelve business days. Penn State’s guidance tells you what to request: total soil lead, pH, lime requirement and phosphorus. Collect eight to twelve small scoops across the area, mix them, and send about a cup. Sample the top inch or two in undisturbed ground and about six inches where you intend to dig.

What you do with the result scales. Penn State’s thresholds run from ordinary hygiene and keeping bare ground covered at the low end, through covering soil and avoiding root crops and leafy greens in the middle range, to not growing food in native soil at all above roughly 1,000 parts per million. Two details are worth carrying: lead concentrates in root crops and leafy greens far more than in fruiting crops like tomatoes, peppers and beans, and keeping soil pH between 6.5 and 7.0 with added organic matter reduces how available the lead is to plants.

Raised beds with clean imported fill are the practical answer, because they replace the growing medium rather than treating it. Ask where the fill came from. “Topsoil” is not a regulated term.

What can come out of a bag

Two infections are documented in people who handle bagged growing media, and both are rare enough that the right response is a habit, not alarm.

Legionella longbeachae causes Legionnaires’ disease, the same pneumonia people associate with cooling towers, except that this species lives in soil and compost rather than plumbing. The CDC’s Morbidity and Mortality Weekly Report laid out the American evidence on 1 September 2000. Three patients were described: a 46-year-old woman in Washington and a 77-year-old woman in Oregon, both of whom had potted plants in the ten days before symptoms started, and a 45-year-old man in California who died before his home could be sampled. Investigators recovered the organism from potting soil at both of the first two homes. CDC’s conclusion was that transmission from potting soil may have occurred in the United States for the first time, and that surveillance was warranted to find out.

Twenty-five years later, the count outside Australia and New Zealand is still small. A 2021 case report and literature review in IDCases, written by clinicians at the University of Maryland, counted 38 reported cases in Europe, 33 in Asia and 8 in North America across two decades in countries where the organism is not endemic. Average age was 65 and about two-thirds were men. Both CDC and the Maryland team flag the same reason the count is probably understated: the routine urine antigen test is not sensitive for species other than Legionella pneumophila serogroup 1, so this one slips past unless somebody orders a culture.

A case-control study in the CDC journal Emerging Infectious Diseases in 2017 is the best look at behavior. Researchers in Canterbury, New Zealand, compared 31 hospitalized patients with 172 randomly sampled controls. The risk factors were chronic obstructive pulmonary disease, more than ten years of smoking, and exposure to compost or potting mix. Two specific habits stood out: tipping and troweling the material, and putting unwashed hands near the face afterward. The finding people skip is that wearing a mask or gloves was not protective in that sample. That does not make protective equipment pointless, but it does mean the reliable moves are the dull ones. Dampen the medium before you open and tip it, open it away from your face, work outdoors or with real air movement, and wash your hands before you touch your face.

The second organism is Sporothrix, a fungus the CDC says lives in soil and on plants including sphagnum moss, rose bushes and hay. It causes sporotrichosis, which the agency calls rose gardener’s disease, and it gets in through cuts and thorn punctures rather than through the air. Symptoms start one to twelve weeks later as a small painless bump that can grow and open. The long lag is the catch: by the time a sore appears, nobody connects it to a rose bush two months ago. Tell your clinician you garden.

Punctures, tetanus and a red line that is moving

Clostridium tetani lives in the environment and enters through broken skin, typically a wound that is deep or contaminated with dirt, according to the CDC. Symptoms usually appear between three and twenty-one days later, averaging about eight days, and often start with jaw spasm. The agency states that one in ten cases in the United States is fatal.

The defense is a vaccine you may already be overdue for. The CDC recommends a tetanus booster every ten years for adults, as Td or Tdap. Most people cannot remember their last one, which is the whole problem, and a garden is a place where you will eventually drive something into your hand. Ask at your next appointment rather than at the urgent care.

Separately, watch any garden wound for cellulitis, a bacterial infection of the skin and tissue underneath. The CDC describes it as painful, red, swollen, tender and warm skin, most often on the feet and legs, sometimes with fever and chills. On when to act, the agency is direct: “Seek medical attention immediately if the red area of the skin spreads quickly or you develop a fever or chills.” Spreading redness is the sign to stop waiting. A deep puncture from a stake, a pruner or a buried piece of metal is also a reason to be seen rather than to clean it yourself and hope.

Knees, backs and the machine in the shed

Two of the garden’s physical costs are documented well enough to name.

Kneeling has a dose-response relationship with knee osteoarthritis, but the evidence comes from work, not weekends. A 2017 systematic review in Safety and Health at Work pooled 15 studies and estimated an odds ratio of 1.26 for knee osteoarthritis per additional 5,000 hours of cumulative kneeling, with a confidence interval from 1.17 to 1.35. The authors rated the evidence moderate quality and judged all but one of the included studies at high risk of bias. Five thousand hours is a career of flooring or mining, not a season of weeding. The honest reading is that kneeling load accumulates, that the research measured a population no recreational gardener resembles, and that a kneeling pad and standing up every twenty minutes cost nothing. Article 671 owns strength training, which is what protects you while lifting bags of soil.

Powered tools are a different scale of risk entirely. A 2018 analysis in the American Journal of Emergency Medicine used the Consumer Product Safety Commission’s National Electronic Injury Surveillance System to count lawn mower injuries treated in US emergency departments from 2005 to 2015: an estimated 934,394 over eleven years, averaging about 84,944 a year. Hands and fingers accounted for 22.3 percent and lacerations were the most common injury type at 23.1 percent. Mean age was 46.5, men were injured more than three times as often as women, and roughly one in twelve patients was admitted. The incidence showed no decline across the eleven years.

Ask a clinician first if this is you

Two groups should raise soil before starting rather than after.

Anyone with a weakened immune system, from medication, a transplant, cancer treatment, advanced HIV or another cause, faces a different version of these risks. The CDC notes that disseminated sporotrichosis, which spreads to bones, joints or the central nervous system, typically occurs in people with weakened immunity. The Legionnaires’ risk factors in the New Zealand study were chronic lung disease and long-term smoking. These are not reasons to give up gardening. They are reasons to have the conversation before you buy the soil.

Anyone pregnant should also ask. Toxoplasma can be present in soil contaminated by cat feces, and the CDC’s prevention guidance says plainly: “Wear gloves when gardening or when touching any soil or sand that may be contaminated with cat feces containing Toxoplasma.” The agency also advises washing with soap and water afterward and tells people at higher risk to talk with their healthcare provider. That last instruction is the one to follow rather than improvising from an article.

The order that makes sense

If you do five things in this sequence, you will have handled most of what is above before you have spent much money.

1. Test the soil where you plan to grow food, through your county extension office, and ask for total lead along with pH. Do this before you build anything, because the result decides whether you are amending ground or filling a box.
2. Check your tetanus date at your next medical appointment, not at the emergency room.
3. Decide in-ground or raised based on the number that comes back, and if you are importing fill, ask the supplier what it is and where it came from.
4. Set up for dust, which means wetting bagged media before opening it, tipping it away from your face, and keeping hands away from your face until they are washed.
5. Put a kneeling pad and a pair of real gloves where you will reach them, because equipment you have to go find is equipment you skip.

If you keep one thing, keep the first. The health benefits of gardening are modest, partly real and spread across several ordinary pathways, and none of them depends on your technique. The soil lead question is the one variable that differs by orders of magnitude between two houses on the same block, that you cannot see or smell, and that a ten-dollar test answers in a couple of weeks.

Sources

1. Centers for Disease Control and Prevention. “About Tetanus.” cdc.gov
2. Centers for Disease Control and Prevention. “Tetanus Vaccination.” cdc.gov
3. Centers for Disease Control and Prevention. “About Sporotrichosis.” cdc.gov
4. Centers for Disease Control and Prevention. “Lead in Soil.” cdc.gov
5. Centers for Disease Control and Prevention. “Preventing Toxoplasmosis.” cdc.gov
6. Centers for Disease Control and Prevention. “Cellulitis: All You Need to Know (Group A Strep).” cdc.gov
7. Centers for Disease Control and Prevention. “Legionnaires' Disease Associated With Potting Soil: California, Oregon, and Washington, May-June 2000. MMWR Morbidity and Mortality Weekly Report, 1 September 2000, volume 49, number 34, page 777.” cdc.gov
8. US Environmental Protection Agency. “Residential Soil Lead Directive for CERCLA Sites and RCRA Hazardous Waste Cleanup Facilities (16 October 2025).” epa.gov
9. Penn State Extension. “Lead in Residential Soils: Sources, Testing, and Reducing Exposure.” extension.psu.edu
10. Penn State Agricultural Analytical Services Laboratory. “Soil Fertility Testing.” agsci.psu.edu
11. University of Massachusetts Amherst, Center for Agriculture, Food and the Environment. “Soil and Plant Nutrient Testing Laboratory.” umass.edu
12. USDA National Institute of Food and Agriculture. “Cooperative Extension System.” nifa.usda.gov
13. Litt JS, Alaimo K, Harrall KK, Hamman RF, Hebert JR, Hurley TG, Leiferman JA, Li K, Villalobos A, Coringrato E, Courtney JB, Payton M, Glueck DH.. “Effects of a community gardening intervention on diet, physical activity, and anthropometry outcomes in the USA (CAPS): an observer-blind, randomised controlled trial. The Lancet Planetary Health, 2023. PubMed. DOI: 10.1016/S2542-5196(22)00303-5.” doi.org
14. Coringrato E, Alaimo K, Leiferman JA, Villalobos A, Buchenau H, Decker E, Fahnestock L, Quist P, Litt JS.. “A process evaluation of a randomized-controlled trial of community gardening to improve health behaviors and reduce stress and anxiety. Scientific Reports, 2024. PubMed. DOI: 10.1038/s41598-024-63889-w.” doi.org
15. Soga M, Gaston KJ, Yamaura Y.. “Gardening is beneficial for health: A meta-analysis. Preventive Medicine Reports, 2016. PubMed. DOI: 10.1016/j.pmedr.2016.11.007.” doi.org
16. Hawkins JL, Smith A, Backx K, Clayton DA.. “Exercise intensities of gardening tasks within older adult allotment gardeners in Wales. Journal of Aging and Physical Activity, 2014. PubMed. DOI: 10.1123/japa.2013-0171.” doi.org
17. Kenagy E, Priest PC, Cameron CM, Smith D, Scott P, Cho V, Mitchell P, Murdoch DR.. “Risk Factors for Legionella longbeachae Legionnaires' Disease, New Zealand. Emerging Infectious Diseases, 2017. PubMed. DOI: 10.3201/eid2307.161429.” doi.org
18. Bell H, Chintalapati S, Patel P, Halim A, Kithas A, Schmalzle SA.. “Legionella longbeachae pneumonia: Case report and review of reported cases in non-endemic countries. IDCases, 2021. PubMed. DOI: 10.1016/j.idcr.2021.e01050.” doi.org
19. Verbeek J, Mischke C, Robinson R, Ijaz S, Kuijer P, Kievit A, Ojajarvi A, Neuvonen K.. “Occupational Exposure to Knee Loading and the Risk of Osteoarthritis of the Knee: A Systematic Review and a Dose-Response Meta-Analysis. Safety and Health at Work, 2017. PubMed. DOI: 10.1016/j.shaw.2017.02.001.” doi.org
20. Harris C, Madonick J, Hartka TR.. “Lawn mower injuries presenting to the emergency department: 2005 to 2015. The American Journal of Emergency Medicine, 2018. PubMed. DOI: 10.1016/j.ajem.2018.01.031.” doi.org