Puffs: what the air in an extruded snack actually does

A cheese puff is mostly air, and the air turns out to be the part anyone has measured. In a randomised crossover trial, 28 healthy adults were served an extruded cheese puff aerated to 0.45 kcal per ml and the same snack at 1.00 kcal per ml, identical at 5.7 kcal per gram. They ate 21 percent less of the airier one by weight and by energy, about 70 kcal per sitting, while eating 73 percent more of it by volume. Each condition was a single afternoon in a laboratory and nobody recorded what these people ate at the next meal, so the result describes one sitting rather than a week. It is still the strongest finding on this page. Puffed and extruded snacks have no single identity in the data. There is no USDA record under this name, and the studies arrive under other names: cheese puffs, extruded corn snacks, iron-fortified puffed rice and the Southeast Asian curry puff.

What a real portion looks like

• Aerating an extruded cheese puff snack to 0.45 kcal/ml instead of 1.00 kcal/ml cut energy eaten by 21%, about 70 kcal per sitting, in 28 adults. — 16 women and 12 men, healthy adults, four afternoon laboratory sessions, snack eaten ad libitum (n = 28) Evidence B source
  21% less weight and energy eaten, 70 +/- 17 kcal, while volume eaten rose by 73% (239 +/- 24 ml)
• A curry puff tested as a standard portion in 47 Southeast Asian volunteers fell into the low glycaemic index band, alongside five other local foods. — 47 healthy volunteers, 18 men and 29 women, standard portion sizes after a 10 to 12 hour fast, glucose as the reference food (n = 47) Evidence B source
  Curry puff classified low GI among fifteen foods tested; six low, three medium, six high

What your body absorbs

• Extrusion cooking of a bran product destroyed its phytase, and phytate phosphorus leaving the small intestine rose from 44% to 73% of intake in 7 ileostomy volunteers. — 7 ileostomy subjects, two 4-day periods on a constant low-fibre diet plus 54 g/day of the bran product (n = 7) Evidence B source
  Phytate phosphorus recovery 44% of intake unextruded versus 73% extruded
• Raising the extrusion temperature of iron-fortified rice from 40 C to 90 C changed the starch but left iron absorption unchanged, while adding citric acid and trisodium citrate raised absorption from 3.0% to 8.2%. — 22 young women, mean age 23 years, geometric mean plasma ferritin 29.3 mcg/L, standardised rice and vegetable test meals with stable isotope labels (n = 22) Evidence C source
  Fractional iron absorption 8.2% (95% CI 7.9, 11.0) with citrate versus 3.0% (95% CI 2.7, 3.4) without, P < 0.001; extrusion temperature had no significant effect

What cooking and storage change

• Replacing 40% of the corn in an extruded snack with pinto bean or chickpea flour lowered the pre-meal blood glucose response in 26 healthy young adults, while yellow pea and green lentil flours did not. — 26 healthy young adults, double blind repeated-measures crossover, extruded corn snacks with 40% pulse flour versus a 100% corn control (n = 26) Evidence B source
  Lower pre-meal blood glucose incremental area under the curve for pinto bean and chickpea snacks, p < 0.05; no difference in food intake or appetite

What it does to your health

• A 98 g curry puff eaten with breakfast was the medium glycaemic index arm of a 30 person trial in which sugar intake correlated negatively with Lactobacillus abundance after 24 hours. — 30 healthy volunteers aged 18 to 45, 10 per dessert group, faecal samples at baseline and 24 hours (n = 30) Evidence B source
  Significant negative correlation between sugar intake and Lactobacillus abundance in the medium and high GI groups, absent in the low GI group

Safety, limits and interactions

• Aflatoxins were measured in 27 samples of corn flour cheese puffs from three Iranian brands, with the highest single reading 1.35 ng/mL and every sample below the national limit of 5 mcg/kg. — 27 specimens of cheese puffs from three popular brands sold in Iran, analysed by HPLC with antibody columns (n = 27) Evidence C source
  Highest brand readings 1.35, 0.95 and 1.22 ng/mL; all specimens below the Iranian permissible limit of 5 mcg/kg

What people believe that the data does not support

• A scoping review of 21 human studies separated corn products by processing: fibre-rich corn bran tracked with better blood lipid and glucose profiles, while cornflakes and extruded snacks tracked with worse metabolic outcomes. — 21 peer-reviewed human studies of corn flour consumption across the lifespan, three databases searched July 2024 (n = 21 studies) Evidence C source
  Direction of association opposite for whole grain corn products versus refined corn products such as cornflakes and extruded snacks, most marked in people with metabolic disorders

Who this works differently for

• Two thirds of Australian commercial toddler foods audited in 2022 were snacks, and only 19% of the audited snack foods, infant and toddler together, met all the WHO Europe compositional requirements for children under 36 months. — cross-sectional audit of commercial infant and toddler foods in three major Australian supermarkets, 2022 (n = not stated in abstract) Evidence C source
  23% of infant and 67% of toddler products were snack foods; 19% met all NPPM requirements; energy density compliance 18% for toddler versus 72% for infant snacks; total sugar compliance 42% versus 88%
• Seven in ten commercial snacks and finger foods for children under three in the same Australian audit dissolved in the mouth, and 63% were ultra-processed by the NOVA system. — commercial snack foods for infants under 12 months and toddlers 12 to under 36 months, Australian supermarket audit 2022 (n = not stated in abstract) Evidence C source
  71% dissolvable with or without an initial crispy or crunchy texture; 63% ultra-processed
• In a 2023 audit of 189 commercial infant and toddler products in Turkey, 40% exceeded the WHO Europe sodium standards and 9.5% carried added salt. — 189 commercial infant and toddler food products in Turkish in-store and online supermarkets, March 2023, composition taken from label panels (n = 189) Evidence C source
  76 of 189 products (40%) above WHO Europe sodium standards; 18 products (9.5%) with added salt; 89 products (49%) with added sugar or sweeteners

What is still unknown

• The same review names the gap directly: long term health effects of corn based products are not settled and larger studies are still needed. — 21 human studies of corn flour consumption identified in a systematic database search (n = 21 studies) Evidence C source
  No long term outcome estimate available; authors call for more rigorous larger-scale studies

The bottom line

Extrusion does chemistry to the grain. Mild extrusion of a bran product destroyed its phytase. In 7 ileostomy volunteers eating 54 g of that product a day on a fixed low-fibre diet, phytate phosphorus leaving the small intestine rose from 44 percent of intake to 73 percent. That crossover weighed what reached the end of the small intestine rather than any mineral absorbed. Temperature is a weaker lever than people assume. Raising the extrusion temperature of iron-fortified rice from 40 C to 90 C changed the starch and left iron absorption unchanged in 22 young women, in a non-randomised single-meal isotope study in women who were already iron-replete. Adding citric acid and trisodium citrate to the same rice raised fractional iron absorption from 3.0 percent (95 percent CI 2.7 to 3.4) to 8.2 percent (95 percent CI 7.9 to 11.0). The recipe matters too. Replacing 40 percent of the corn with pinto bean or chickpea flour lowered the pre-meal blood glucose area under the curve in 26 healthy young adults (p < 0.05) in a double blind crossover. Yellow pea and green lentil flours did not, and appetite and later food intake did not differ. That is one acute measurement at one substitution level. The curry puff in this evidence is a filled pastry rather than a bagged puffed snack. In 47 healthy Southeast Asian volunteers eating standard portions after a 10 to 12 hour fast, it landed in the low glycaemic index band among fifteen foods tested over 120 minutes. In a separate trial of 30 volunteers with 10 per arm, a 98 g curry puff was the medium glycaemic index breakfast. The relationship reported after 24 hours was a negative correlation between sugar intake and Lactobacillus abundance in the medium and high index groups, which the authors call a temporary shift. Corn snacks and corn get spoken about as one food. A scoping review of 21 human studies found the associations pointing opposite ways: fibre-rich corn bran with better blood lipid and glucose profiles, cornflakes and extruded snacks with less favourable metabolic outcomes, most marked in people with metabolic disorders. A scoping review pools no estimate and grades no bias, so it maps the literature instead of weighing it. The same authors say the long term health effects of corn based products are unsettled and that larger studies are needed. The cautions we have are narrow and specific. Aflatoxins were measured in corn flour cheese puffs from three Iranian brands, with highest brand readings of 1.35, 0.95 and 1.22 ng/mL, all below the Iranian permissible limit of 5 mcg/kg. That survey covered 27 retail specimens from one country, reported concentrations in ng/mL against a limit written in mcg/kg, and modelled nobody's intake. The rest concerns small children. In a 2022 audit of three Australian supermarket chains, 67 percent of toddler products were snack foods and only 19 percent of the audited snacks, infant and toddler together, met every WHO Europe compositional requirement. Toddler snacks were 18 percent compliant on energy density against 72 percent for infant snacks, and 42 percent on total sugar against 88 percent, all read off label panels rather than analysed samples. Seven in ten of those snacks dissolved in the mouth and 63 percent were ultra-processed by the NOVA system, with texture judged from product descriptions. A 2023 Turkish audit of 189 infant and toddler products found 40 percent above WHO Europe sodium standards, 9.5 percent with added salt and 49 percent with added sugar or sweeteners, again from labels in one month. We found no study on allergy, pregnancy or interactions with medicines for this food, and that silence is a hole in the evidence we gathered rather than an all-clear.

Sources

1. puff-r3-01 · randomised crossover trial
   We tested two versions of an extruded snack (cheese puffs) that were equal in energy density (5.7 kcal/g), but differed in energy per volume (less-aerated snack: 1.00 kcal/ml; more-aerated snack: 0.45 kcal/ml)...subjects consumed 21% less weight and energy (70+/-17 kcal) of the more-aerated snack than the less-aerated snack, although they consumed a 73% greater volume of the more-aerated snack (239+/-24 ml)
   Appetite, 2007 · checked 2026-09-24
2. puff-r3-05 · randomised crossover trial
   Among the fifteen foods tested, six belonged to low-GI foods (Ice Green Tea, Beehoon, Pandan Waffle, Curry Puff, Youtiao and Kaya Butter Toast), three belonged to medium-GI foods (Barley Drink, Char Siew Pau and Nasi Lemak)
   Br J Nutr, 2015 · checked 2026-09-24
3. puff-r3-02 · randomised crossover ileostomy study
   Extrusion cooking, using mild conditions, did not change the content of starch, dietary fibre components or phytate of the bran product, but the phytase (EC 3.1.3.26) activity was lost. During the period using the extruded bran product, there was a significant increase in recovery of phytate-phosphorus (period A, 44% of intake; period B, 73% of intake).
   Br J Nutr, 1986 · checked 2026-09-24
4. puff-r3-04 · non-randomised stable isotope trial
   Warm and hot extrusion altered starch morphology from native type A to type V and increased retrograded starch. However, extrusion temperature did not significantly affect iron solubility or iron bioavailability. The geometric mean fractional iron absorption of iron from fortified rice extruded with CA/TSC (8.2%; 95% CI: 7.9%, 11.0%) was more than twice that from extruded rice without CA/TSC (3.0%; 95% CI: 2.7%, 3.4%; P < 0.001).
   J Nutr, 2022 · checked 2026-09-24
5. puff-r3-03 · randomised crossover trial
   In a randomized, repeated-measures crossover study, adults (n = 26) consumed extruded corn snacks made with the addition of 40% pulse flour from either whole yellow pea, split yellow pea, green lentil, chickpea, or pinto bean...Pinto bean and chickpea snacks led to lower (p < 0.05) pre-meal BG incremental area under the curve (iAUC), compared with control, whole yellow pea and green lentil snacks.
   Appl Physiol Nutr Metab, 2021 · checked 2026-09-24
6. puff-r3-06 · randomised parallel trial
   Participants were randomly assigned to one of three groups: Phetchaburi's Custard Cake (192 g, low-GI group, n = 10), Saraburi's Curry Puff (98 g, medium-GI group, n = 10), and Lampang's Crispy Rice Cracker (68 g, high-GI group, n = 10), each consumed alongside their standard breakfast...Correlation analysis revealed a significant negative relationship between sugar intake and Lactobacillus abundance in the medium- and high-GI groups, but not in the low-GI group.
   Nutrients, 2024 · checked 2026-09-24
7. puff-r3-07 · cross-sectional product survey
   The highest concentration of aflatoxins in brands A, B, and C was observed in specimens A1 (1.35 ng/mL), B2 (0.95 ng/mL), and C2 (1.22 ng/mL), respectively...The results indicated that the concentration of aflatoxins was lower than the permissible limit in Iran (5 µg/kg) in all understudied specimens.
   Sci Rep, 2025 · checked 2026-09-24
8. puff-r3-11 · scoping review
   Findings indicated that whole grain corn products, particularly those high in fiber-rich corn bran, were associated with favorable effects on blood lipid and glycemic profiles. Conversely, refined corn products, such as cornflakes and extruded snacks, were linked to less favorable metabolic outcomes, particularly in individuals with metabolic disorders.
   J Nutr, 2025 · checked 2026-09-24
9. puff-r3-08 · cross-sectional market audit
   Around a quarter (23%) of infant foods and two-thirds (67%) of toddler foods were CSFs, and most (77%) were assessed as being sweet rather than savoury. Only 19% of CSFs met all applicable NPPM compositional requirements, with toddler CSFs performing comparatively worse than infant CSFs, particularly in terms of their energy density (18% of products compliant vs. 72%) and total sugar content (42% vs. 88%).
   Health Promot Int, 2025 · checked 2026-09-24
10. puff-r3-09 · cross-sectional market audit
    Overall, 71% of CSFs were categorized as dissolvable (either with or without an initial crispy/crunchy texture), while 63% were ultra-processed based on the NOVA classification system.
    Health Promot Int, 2025 · checked 2026-09-24
11. puff-r3-10 · cross-sectional label survey
    Almost half of the products (n = 89, 49%) contained added sugar or sweeteners, 41 (22%) had added sugar, and 1 in 3 products (n = 68, 36%) had sugar from fruit-based sources. One in 10 products (n = 18, 9.5%) contained added salt while 40% of the products (n = 76) were above the WHO Europe sodium standards.
    Nutr Health, 2025 · checked 2026-09-24
12. puff-r3-12 · scoping review
    There is a clear need for more rigorous, larger-scale studies to better understand the long-term health impacts of corn flour consumption and its role in health.
    J Nutr, 2025 · checked 2026-09-24