I spend my clinical life at the intersection of cancer and metabolism, which means I spend it looking at the downstream consequences of decisions made decades earlier. Insulin resistance in a 55 year old with breast cancer did not start at 50. Increasingly, the data says it might have started in childhood, and some of it started before birth. When I became pregnant, the science I had spent years thinking about clinically suddenly became personal and, in the same breath, much harder to deal with.
This series is about raising healthy children in an environment that was not designed for the task. This isn’t about your child becoming a purity project. My argument and the framing underneath all four parts is borrowed from toxicology and exercise physiology: hormesis. Biological systems do not thrive at zero stress. They thrive on the right dose of the right stress, with recovery in between. Unfortunately, the stress the modern world hands a child is the wrong kind.
I am using hormesis somewhat broadly here - not to argue that allergens, vegetables, microbes and exercise act through one pathway, but because they share a developmental principle: appropriate exposure can build capacity that avoidance cannot.
A child’s metabolism, immune system, and nervous system are all built by challenge. The modern environment fails children in both directions at once. It removes the stressors that build them, the microbes, the boredom, the bitter vegetables, the temperature swings, and replaces them with stressors that do not, the engineered food, the endocrine disruptors, the light at 9pm.
We’ll talk about food first, because it is the largest input by volume and the one with the best evidence. But rest assured, next articles will cover environmental exposures, sleep and nervous system regulation, and immunity.
The category is the problem, not the ingredient
Roughly 2/3rds of the calories American children eat are ultraprocessed. Not “processed,” which includes frozen vegetables and canned beans. Ultraprocessed: industrial formulations engineered from refined starches, extracted oils, protein isolates, and additive systems designed to make the product soft, palatable, and shelf stable.
The evidence against the category has firmed up considerably. In children followed over time, higher ultraprocessed intake tracks with greater fat mass, waist circumference, and overweight risk. The ALSPAC birth cohort found the children eating the most ultraprocessed food gaining more BMI and fat mass every year from childhood into early adulthood, and a recent meta-analysis in adolescents put the excess odds of overweight at over 60%.
The inflammation story in children is a bit messier. Some studies find higher IL-8, leptin, and CRP in the heaviest consumers. One Brazilian cohort found its highest cytokine levels in the children eating a moderate amount rather than in the children eating the most. A 2026 systematic review that pulled this literature together split the difference. It found the metabolic outcomes reasonably consistent and the inflammatory markers not, and concluded that pediatric data is still too limited to establish causality.
My attempt to explain some of the inconsistency, and why these limitations push results toward finding nothing.
Standard inflammatory markers in healthy children run low, often at the bottom of what the test can measure, which leaves little room to detect a difference between one group and another.
A single CRP is also moved by a cold from two weeks ago, by puberty, by body fat, so a diet signal has to be larger than all of that to surface. Diet is measured roughly, usually with one 24 hour recall, which captures a Tuesday rather than a childhood and misclassifies who the heavy consumers actually are.
Adiposity is probably a strong mediator, which means studies adjusting for BMI are adjusting against part of what they are trying to detect.
I believe the inflammation is real and largely downstream of fat mass accumulating over years, which would put childhood at the start of the process and adulthood at the point where we typically meet it in clinic.
The mechanism is rather straightforward and it is not any single ingredient. A good example of this is the Hall’s inpatient trial at NIH where adults who were randomized to an ultraprocessed diet ate about 500 calories a day more than the same people on a whole food diet matched for presented calories, sugar, fat, and fiber. They were not weaker willed on ultraprocessed weeks. The food was doing something, through energy density, soft texture, speed of eating, and a disrupted food matrix that outruns satiety signaling.
That finding has since been replicated. The UK UPDATE trial tested the obvious objection - whether an ultraprocessed diet that meets nutrition guidelines is fine. Fifty-five adults ate both diets (one ultraprocessed and one minimally processed, both compliant with national dietary guidance) for eight weeks each. Both met UK guidelines, but participants lost twice as much weight on the minimally processed one. Meeting the nutrient targets did not cancel out the processing.
Then a follow-up trial changed nothing but texture. Both diets were ultraprocessed. One was built from foods that took more chewing and more time to get through, and people ate about 370 calories a day less on it. That is most of the way to Hall’s 500 calories just from texture alone! One of the biggest costs of ultraprocessed food is simply that it is soft and fast.
Back to pediatric data, which at this point is mostly observational. Cross-sectional studies in children are frequently null; it is the cohorts with years of follow-up that show the signal, which suggests the damage is cumulative rather than immediate. I don’t think you need any formal training in medicine to understand that about food. NOVA sorts food by how it was made, not by how good it is for you. Whole grain bread with an emulsifier counts as ultraprocessed. So does soda. The line is soft enough that when researchers reclassified the ambiguous items, the national estimate moved by 6 percentage points. So I hold the category firmly and the single foods loosely. A diet that is mostly ultraprocessed is a problem. One food that lands in the category usually is not.
What this means practically is that the label reading I do is mostly a shortcut for one question: is this food or is this a formulation. Baby food pouches, including the organic ones, read as formulation. Fruit and vegetable purée as filler, the fiber matrix destroyed, sucked through a spout at the exact age a child should be chewing. Soft and fast, by design.
The same test catches most of the children's aisle. Fruit snacks made with real fruit are juice concentrate and starch with the fiber engineered out. Yogurt tubes are a sugar delivery system with stabilizers. Veggie straws are potato starch tinted with vegetable powder. Kids' protein bars are protein isolate and syrup. Toddler milks are skim milk powder, oils, and added sugar sold as an upgrade over the whole milk they replace. None of it is poison, and one of these in a lunchbox is not a crisis. And, for all you label readers, the back of the package tells you what was added to the product, it is also important to ask what was done to the ingredients.
Build the palate, don’t just feed it
The conversation about children’s food is dominated by what to remove. I am a big believer in adding and crowding out.
Fiber is the chronic shortfall that matters most, and not just nutritionally or for GI health. Fiber feeds the microbial community that trains the immune system and produces the short chain fatty acids that regulate inflammation, and American children get roughly half of what they should, around 13 grams a day against targets in the mid 20s and up. In one adolescent cohort, two participants out of 754 met the fiber recommendation. Read that again. Lower fiber intake in adolescents tracks with higher fasting insulin, HOMA-IR, and blood pressure, so this is not a “are you pooping consistently” issue, it is a metabolic one. The fix is not supplements or shakes, it is soaked-from-dry beans, berries, vegetables, intact grains, presented at every meal, and from a young age.
Bitterness deserves its own paragraph, as one of my favorite food groups and the best example of nutritional hormesis. The polyphenols and glucosinolates in vegetables are not vitamins. Many of them work as mild toxins, and the body responds by upregulating its own antioxidant and detoxification machinery through pathways like Nrf2. Sulforaphane, the compound chewing releases from broccoli, is the stimulus. The response is your own cells turning on glutathione synthesis and the phase II detoxification enzymes, and that output keeps running for days after the meal. You are not eating an antioxidant. You are building one.
This is the same logic as exercise, and not only by analogy. Exercise activates Nrf2 too, and in older adults exercise plus sulforaphane activates it more than either alone. It fails the same way too: in Ristow’s trial, high-dose vitamin C and E blocked both the insulin sensitivity gains from training and the rise in the body’s own antioxidant enzymes. Blocking the stress blocks the adaptation. That said, the human trial evidence for specific phytochemicals is thinner than the mechanism deserves.
Learning to eat bitter food is a training problem, not a nutrition problem. Flavor learning starts in utero, continues through breast milk, and heightens through the first two years. I share this one for the comical factor, in Mennella’s randomized carrot juice experiment, mothers who drank carrot juice in the third trimester or while nursing had babies who later ate carrot flavored cereal with visibly less aversion than controls. Comedy aside, amniotic fluid and breast milk carry the flavors of the maternal diet, and your fetus is the food critic in training.
Food neophobia arrives on schedule too: it climbs steeply from around 18 months, stays high through the preschool years, and is developmentally normal. Acceptance of a new food takes 8 to 10 exposures, and the acceptance you build generalizes to similar foods within the same category, another reason vegetables are worth the repetitions. Most parents stop offering after three to five tries and conclude the child dislikes it. The infant window is also the easy one, babies under a year accept new foods after remarkably few exposures, so the cheapest time to build the palate is before the year mark and before the neophobia arrives.
My son eats from our plates, at home and in restaurants. He gets the spicy food, the fermented things, the dishes I know he will reject again and again, until one day he doesn’t. My job is repeated exposure without pressure. His job is deciding whether to eat.
The kids’ menu - beige, fried, and sweet, served at precisely the age when food preferences are taking shape - is the food environment working against our children. It is a construct of modern food culture that makes my blood boil. We are feeding the population with some of the highest nutritional needs some of the least nutritious food.
The case against avoidance
For years, official guidance told parents to delay allergenic foods. The LEAP trial - thankfully - ended that. It showed that high risk infants who ate peanut early and regularly had 81% less peanut allergy than infants who avoided it. The follow-up work gets even better! In LEAP-On, a year of deliberate avoidance after age five did not undo the protection. In LEAP-Trio, the same children at age 12 still carried a 75% reduction in peanut allergy, eating peanut or not as they pleased in the intervening years. The tolerance was built once, early, and it held. Avoidance has a cost. Sometimes that cost is the disease it was meant to prevent.
I keep coming back to LEAP because it is the best documented case of a larger pattern, and it is the pattern this whole series is about. The developing system needed the exposure. The protective instinct, remove the challenge, produced the harm.
What does organic actually buy you?
We buy mostly organic. But, what does it actually get us? Intervention studies show an organic diet drops children’s urinary pesticide metabolites quickly and substantially. In Lu’s studies, malathion and chlorpyrifos metabolites fell to undetectable within days of the switch and returned when the conventional diet did. Two caveats though.
In children from agricultural communities, the reductions are partial, closer to 40 or 50%, because their exposure is not only dietary.
And a few small trials have found shifts in oxidative stress biomarkers on organic diets, which is interesting and not the same thing as a health outcome.
What no study has yet shown is that the reduction in exposure improves a measured clinical outcome in those children. That trial has not been done and would be hard to do. The observational literature does not settle it either. Organic consumers are wealthier, leaner, and eat more plants, and the studies split accordingly: one large French cohort found 25% less cancer, a UK cohort of 623,000 women found no reduction, and a 2025 meta-analysis pooling them found no association at all. So organic buys a documented reduction in exposure with a plausible but unproven benefit, at real cost. I pay it, prioritizing the produce that carries the highest residues. EWG publishes the “dirty dozen” list annually.
That said, I will always tell anyone on a budget that conventional produce beats eating fewer vegetables, every time, and that the ultraprocessed question matters far more than the organic one.
So what are you to do?
Feed the family one meal from whole ingredients and let the child eat it. Count additions, protein, fiber, and plants, before subtractions. Make food that takes chewing, because soft and fast is most of what ultraprocessing costs you. Offer the bitter thing fifteen times without commentary. Introduce allergens early and keep them in rotation. Spend organic dollars on the highest residue produce or do not spend them at all, but do not let that decision distract from the one that matters, which is whether the food is food.
If you are pregnant or nursing, eat the foods you want your child to eat, because they are already tasting them. That window is cheap, it is short, and it does not come back.
Then hold all of it loosely. Childhood sets a trajectory, not a verdict. What I am actually asking for is not a clean, perfectly fed child. It is a child who got handed enough of the right stress, early enough, to build something that holds.
Next: A Companion Guide on Food and Cancer landing in your mailboxes early next week. A discussion on whether anything in your child’s lunchbox causes childhood leukemia, what the epigenetics data really shows, + meat and soy.
Then Part 2: the chemical environment, what the evidence actually supports, and where I think the fear is miscalibrated, including plastics, pesticides, and EMFs.
Ksenia Malarkey, ND is a naturopathic physician specializing in integrative oncology, hormones, and metabolic health.
She provides medical care for Washington and Oregon states residents and offers global health coaching.
To schedule a free 15-minute consult, visit drkseniamalarkey.com
This is an excellent article. Thank you.