People who eat French fries three or more times a week face a 20 percent higher risk of developing type 2 diabetes compared with those who rarely touch them, according to a study that followed roughly 205,000 U.S. health professionals across several decades. The finding, drawn from three of the longest-running dietary cohorts in American epidemiology, adds fresh weight to a growing body of evidence that how a potato is prepared matters as much as how often it is eaten.
Why the fry-diabetes link demands attention right now
Type 2 diabetes already affects tens of millions of Americans, and diet remains one of the few modifiable risk factors within an individual’s control. The new research, published in a recent BMJ analysis, pooled individual-level data from the Nurses’ Health Study (1984 to 2020), the Nurses’ Health Study II (1991 to 2021), and the Health Professionals Follow-up Study (1986 to 2018). Together, these three cohorts covered more than 205,000 participants whose diets were reassessed every four years through validated food-frequency questionnaires.
The 20 percent increase in diabetes risk tied to three weekly servings of fries persisted even after researchers accounted for overall diet quality, body mass index, and other lifestyle factors. That adjustment is significant because it suggests the risk is not simply a byproduct of eating too many calories or carrying extra weight. One plausible mechanism involves the cooking process itself. Deep frying at high temperatures generates advanced glycation end-products, or AGEs, compounds formed when sugars react with proteins or fats under intense heat. Repeated exposure to dietary AGEs has been linked in laboratory and clinical studies to insulin resistance and chronic inflammation, both of which precede type 2 diabetes. In other words, the oil bath that turns a raw potato into a crispy fry may introduce biochemical hazards that a baked or boiled potato does not.
Decades of cohort data behind the 20 percent risk increase
The strength of this finding rests on the scale and duration of the underlying cohorts. Participants reported their potato consumption, broken down by preparation method, on questionnaires administered every four years. That repeated measurement allowed researchers to capture changes in eating habits over time rather than relying on a single dietary snapshot. An earlier cohort analysis, published in Diabetes Care, had already flagged a connection between frequent fried-potato intake and higher diabetes incidence. The 2025 BMJ paper extended the follow-up period and accumulated substantially more confirmed diabetes cases, sharpening the statistical precision of the estimate and confirming that the association persisted over decades.
Baked, boiled, and mashed potatoes showed weaker associations with diabetes risk in the new analysis. That gap between fries and other preparations reinforces the idea that cooking method, not the potato itself, drives much of the observed risk. When researchers modeled dietary substitutions, replacing a weekly serving of fries with whole grains or non-starchy vegetables was associated with lower diabetes incidence, offering a concrete trade-off rather than a blanket warning against all potatoes. In contrast, swapping fries for other refined starches, such as white rice or pasta, produced smaller gains, underscoring that multiple elements of a Western-style diet can nudge risk in the wrong direction.
The pattern also fits within a broader evidence base. A peer-reviewed evidence map that synthesized observational studies on potato consumption and metabolic health found consistent signals across multiple research groups and geographies: fried preparations carried higher metabolic risk than non-fried ones. That consistency across independent datasets reduces the chance that the BMJ result is a statistical fluke specific to American health professionals and instead points to a more generalizable concern about deep-fried starchy foods.
Open questions about fry frequency and real-world diets
Several gaps in the evidence prevent a clean causal conclusion. The three cohorts consist predominantly of health professionals, a group whose baseline health behaviors, income levels, and access to medical care differ from the general population. Whether the same 20 percent risk increase would hold in a more socioeconomically diverse sample is an open question. Observational studies, no matter how large, also cannot rule out residual confounding. People who eat fries frequently may share other habits, from higher fast-food consumption to lower physical activity, that the statistical models did not fully capture.
The dose-response curve above three servings per week is less clearly defined in the primary data tables. Secondary coverage has cited higher risk estimates for five or more weekly servings, but those figures lack the same level of granularity in the published study’s main results. Readers should treat escalating risk numbers beyond the three-serving threshold with some caution until additional primary reporting clarifies the shape of the curve at higher intake levels. It is plausible that risk continues to rise with heavier intake, but the exact slope of that increase remains uncertain.
There is also the question of portion size. A serving of fries at a sit-down restaurant can be twice the volume of a serving used in a food-frequency questionnaire, and fast-food portions can be larger still. Self-reported dietary data, even when collected with validated instruments, tends to underestimate actual intake. If participants systematically underreported their fry consumption, the true risk associated with a given real-world portion could be either larger or smaller than the headline figure, depending on the direction and consistency of the measurement error. For example, if heavy fry eaters underreport more than light fry eaters, the study may be underestimating risk at the highest consumption levels.
Another limitation is that the analysis did not distinguish between different types of frying oils or cooking technologies. Fries prepared in partially hydrogenated oils, which contain industrial trans fats, may pose different metabolic risks than fries cooked in newer oils or air-fried at lower temperatures. Likewise, toppings and accompaniments-from sugary sodas to processed meats-travel with fries in typical meals and may contribute additional calories, sodium, and saturated fat. While statistical models attempt to adjust for overall dietary patterns, they cannot fully recreate the complexity of real-world eating contexts.
What this means for your plate
For people looking to act on this research, the practical takeaway is straightforward rather than alarmist. The data do not demand that anyone eliminate potatoes entirely. They do suggest that swapping fries for less processed options most of the time is a prudent move, especially for those already at elevated risk of diabetes because of family history, excess weight, or prediabetes. Choosing baked, boiled, or roasted potatoes, ideally with the skin on and paired with vegetables and lean protein, preserves the potato’s fiber and micronutrients without the added burden of deep-frying oils and high-temperature byproducts.
Portion awareness also matters. Opting for a small order instead of a large, sharing fries at the table, or treating them as an occasional side rather than a default staple can meaningfully cut weekly exposure without demanding a complete lifestyle overhaul. At home, oven-roasting potato wedges with a modest amount of oil or using an air fryer can deliver much of the familiar texture with fewer calories and potentially fewer harmful compounds than traditional deep frying.
For clinicians and public health practitioners, the study offers a concrete message to weave into counseling: not all starches, and not all potatoes, are metabolically equivalent. When discussing dietary changes with patients, emphasizing specific swaps-fries for non-starchy vegetables, sugary beverages for water, refined grains for whole grains-may feel more actionable than abstract advice about “eating healthier.” Over time, those small, repeated choices can shift an individual’s overall risk profile.
Ultimately, the emerging science around French fries and diabetes risk does not single out a single villain so much as it illustrates how cooking methods, food processing, and eating patterns intersect. For now, the most defensible interpretation is that frequent fry consumption is one marker-and likely one contributor-to a dietary pattern that nudges blood sugar regulation in the wrong direction. Treating fries as an occasional indulgence rather than a thrice-weekly habit aligns both with the new cohort data and with broader guidance on preventing type 2 diabetes.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
