Your Microbes Read the Label, Not the Macros | The Angry Gut, Chapter 15

Ultra-processed food · Gut health

Ultra-Processed Food: Measure the Host, Not the Additive

Additive exposure has no clinical assay, so it stays an estimate from the diet history. The variable that decides the outcome in every animal model is the host, and the host can be measured.

Ultra-processed food is now a routine item in the history, and it arrives without an assay. There is no laboratory test for additive exposure, at Measura or anywhere in ordinary practice, so the exposure stays an estimate from diet records while the host it lands in can be measured directly.

The chapter video, Your Microbes Read the Label, Not the Macros, presents chapter 15 of The Angry Gut by Dr. Gurpreet Singh Padda, MD, MBA, MHP, and Ami Michelle Grimes. Study-level numbers with the limits of each are in the book’s Deep Dive companion. Measura [Cardiometabolic and Autonomic Health Analysis] is a testing service; it performs no stool, microbiome, additive or endoscopic testing.

Is ultra-processed food linked to diabetes, cancer and IBD?

The population signal does not point at the compounds the laboratory studies. Following 104,139 adults for incident type 2 diabetes, guar gum reached 1.11 for each extra 500 mg a day and xanthan gum 1.08, with carrageenans, phosphates and citrate also significant. Carboxymethylcellulose and polysorbate 80, which carry nearly the whole rodent literature, are absent from that list.

In the cancer analysis of 92,000 adults, the hazard for any cancer was 1.15 with mono- and diglycerides of fatty acids and for breast cancer 1.32 with carrageenans, while no association was detected between any emulsifier and colorectal cancer. For whole categories, each 10% increment of intake from ultra-processed sources carried type 2 diabetes at 1.17 across 311,892 people, and those authors recommended targeting specific foods rather than the category. In 116,087 adults, five or more servings a day was associated with incident inflammatory bowel disease at 1.82, interval 1.22 to 2.72.

These are small relative differences per small increments on very large cohorts. They do not establish that guar gum causes diabetes. They indicate that the human signal lives in the mixture.

Why the exposure cannot be quantified in your patient

Two practical limits govern this workup. The exposure variable itself is unreliable: asked to classify marketed foods by processing level, expert evaluators reached a Fleiss kappa of 0.32, while trained coder pairs working from a structured database reached 0.75. The category is codable but not intuitive, which is why a diet history beats a label recalled from memory.

The second limit is compartment. The mouse work indicates the relevant microbial change sits in the mucus layer and differs markedly from what appears in feces, so stool sequencing samples the wrong room. No commercial panel measures either compartment usefully, and none is offered here.

Why do additives harm some people more than others?

This is the part that changes selection. In animals, normal mice largely shrugged the additives off, while in strains already inflamed the fecal inflammatory marker rose roughly ten-fold. In humans, the single measurement of bacterial encroachment tracked glycemic control rather than adiposity: an r-squared of 0.51 for hemoglobin A1c and 0.46 for fasting glucose against 0.16 for body mass index, with obesity alone showing nothing and the difference disappearing when subjects with diabetes were removed.

Read the null feeding trial against that background. Sixteen healthy adults, eleven days, seven receiving 15 g a day of carboxymethylcellulose, and the primary measure did not move. Its power calculation assumed a gap of 19.13 microns, the separation recorded in that colonoscopy study between subjects with diabetes and those without, set against a within-group standard deviation of 7.17 microns. In effect it looked for a difference the size of diabetes in people without diabetes, and every inflammatory readout in it was null.

Which patients should be screened?

Reasonable candidates for cardiometabolic measurement when the diet history is additive-heavy:

Patients with dysglycemia or known type 2 diabetes, in whom the human barrier measurement tracked A1c and fasting glucose.
Patients with obesity, where body mass index conflates fat with lean mass and was the weakest correlate of the three.
Patients with inflammatory bowel disease considering dietary restriction, where a blinded add-back trial in 154 adults with active Crohn’s disease reported response in 49.4% on restriction against 30.7% with emulsifiers returned, adjusted relative risk 3.1, interval 1.5 to 6.6. That is a congress abstract and weighs as one, and a four-week trial in 24 patients found no difference between diets.
Patients whose reported intake is heavily engineered, where measured markers document the baseline before any dietary change is attempted.

What should be measured instead of the additive?

Two measurements carry this pathway, with results returned to the ordering physician:

Laboratory panels for glycemic, lipid and inflammatory markers, which establish whether the patient resembles the inflamed host the mechanism describes or the healthy volunteer the trial enrolled.
Bioimpedance body composition for the fat and lean compartments behind a body mass index, and for tracking change when the diet actually shifts.

Neither measures additives, mucus thickness or barrier penetrability, and no test in the library does. What they document is metaflammation as a set of current numbers, which is the variable the animal work says decides the outcome. Prevalence supports a low threshold: metabolic health is uncommon, at under 12.2% of US adults on NHANES 2009-2016 and under 7% under the tighter criteria after 2021.

Standing orders and documentation

A standing order attaching a laboratory panel and body composition measurement when the diet history is additive-heavy makes the baseline reproducible rather than dependent on which clinician takes the history. Filed as discrete data in the record, those values are still there when the patient returns having changed the food, which is the only way a dietary intervention gets evaluated. Those values also feed the quality measures practices already report through MIPS and quality reporting, which is a documentation benefit rather than an indication.

What the evidence does not settle

The strongest argument against acting on the category comes from inside it. Stratify by diet quality and the category stops predicting mortality, while diet quality keeps predicting within every stratum of intake. The field’s own umbrella review grades incident Crohn’s disease as class IV, weak evidence, and ulcerative colitis as class V, no evidence. Recent review language is explicit that current evidence supports biological plausibility rather than causality, and no human trial has demonstrated the whole pathway from additive to clinical disease.

The regulatory files answer a different question again. Their endpoints were acute toxicity, genotoxicity, carcinogenicity and reproductive toxicity, not microbiota, mucus or low-grade inflammation. The polysorbate group acceptable daily intake is 25 mg per kilogram of body weight per day, and the agency’s own exposure assessment placed toddlers at the highest level at 24.5. When data were formally requested on carboxymethylcellulose, business operators did not provide them, and one infant category could not be assessed at all. An unanswered question is not a safety finding in either direction.

What survives adjudication is practical. Restriction is feasible, with emulsifier-containing food frequency falling 94.6% in patients with stable Crohn’s disease, and the measurable variable in front of you is the host rather than the label. The patient-facing version of this argument is what no test can measure about emulsifier exposure.

Frequently asked questions

Is there a test for food additive exposure?

Not in clinical practice, and not at Measura. Cohort studies estimate intake from repeated diet records mapped to additive composition databases, which is a research method rather than a patient-level measurement. The practical substitute is a diet history plus objective metabolic markers documented before and after a dietary change. Selection criteria sets out who to test.

Should I advise patients to avoid emulsifiers?

Scope it to the patient. A four-week trial in 24 patients found no difference and its authors said avoidance is not supported in the context of a healthy diet; a larger blinded add-back trial in active Crohn’s disease reported 49.4% response against 30.7%. Advising fewer items with ingredient lists is defensible and feasible, and it changes fiber intake at the same time. Clinical rationale.

Does a stool or microbiome panel help here?

The mechanism argues against it. In a cultured human community, cellulose gum raised inflammatory potential without significantly changing which species were present, so a composition report would read as normal. The mouse data place the relevant change in the mucus compartment, which stool does not sample. Measura offers no such testing. What a test result can and cannot tell you.

What does the Measura report add to a dietary discussion?

Objective before-and-after values: glycemic, lipid and inflammatory markers plus fat and lean mass. That converts a dietary trial from a symptom impression into a measured comparison, and it identifies the patients whose baseline resembles the inflamed host in whom the mechanism matters most. Interpreting the report.

Which quality measures does this documentation support?

Cardiometabolic findings and the counseling they prompt fit the diabetes, blood pressure, body mass index and tobacco measures practices already report, and the annual wellness visit is the natural place to capture the baseline. The measurement is ordered for the clinical question; the documentation follows it. Quality measures that cardiometabolic testing supports.

Are ultra-processed foods bad for you?

Very large cohorts link higher intake to disease: each 10% increment of intake from ultra-processed sources carried type 2 diabetes at 1.17 across 311,892 people, and five or more servings a day tracked with inflammatory bowel disease at 1.82 in 116,087 adults. How much harm lands depends on the host. In animals, normal mice largely shrugged the additives off, while in strains already inflamed the fecal inflammatory marker rose roughly ten-fold.

Which additives in ultra-processed food show the strongest human signal?

Not the ones the laboratory studies. Following 104,139 adults for type 2 diabetes, guar gum and xanthan gum stood out, with carrageenans, phosphates and citrate also significant. Carboxymethylcellulose and polysorbate 80, which carry nearly the whole rodent literature, were absent from that list. In the cancer analysis, mono- and diglycerides and carrageenans carried the signals, which points to the mixture rather than a single compound.

How much ultra-processed food is too much?

The cohorts report risk per increment rather than a cutoff. Type 2 diabetes rose with each extra 10% of intake from ultra-processed sources, and inflammatory bowel disease was associated with five or more servings a day. Those are small relative differences on very large populations. For an individual patient, the useful numbers are the host’s own glycemic, lipid and inflammatory markers plus body composition, measured before the diet changes.

Document the host before the diet changes

See how the Measura protocol adds laboratory and body composition measurement to patients whose diet history is additive-heavy, with results returned to the ordering physician.

Send a message

4477 Woodson Rd, Suite 201, St. Louis, MO 63134. Monday to Friday, 9:00 a.m. to 5:00 p.m. Please do not send symptoms, diagnoses or images through a web form — a website form is not a secure medical channel. Send your name and number and we will call you back.

References

Chazelas, E., Druesne-Pecollo, N., Esseddik, Y., de Edelenyi, F. S., Agaesse, C., De Sa, A., Lutchia, R., Rebouillat, P., Srour, B., Debras, C., Wendeu-Foyet, G., Huybrechts, I., Pierre, F., Coumoul, X., Julia, C., Kesse-Guyot, E., Alles, B., Galan, P., Hercberg, S., … Touvier, M. (2021). Exposure to food additive mixtures in 106,000 French adults from the NutriNet-Sante cohort. Scientific Reports, 11(1), 19680. https://doi.org/10.1038/s41598-021-98496-6
Salame, C., Javaux, G., Sellem, L., Viennois, E., de Edelenyi, F. S., Agaesse, C., De Sa, A., Huybrechts, I., Pierre, F., Coumoul, X., Julia, C., Kesse-Guyot, E., Alles, B., Fezeu, L. K., Hercberg, S., Deschasaux-Tanguy, M., Cosson, E., Tatulashvili, S., Chassaing, B., … Touvier, M. (2024). Food additive emulsifiers and the risk of type 2 diabetes: analysis of data from the NutriNet-Sante prospective cohort study. The Lancet Diabetes & Endocrinology, 12(5), 339-349. https://doi.org/10.1016/S2213-8587(24)00086-X
Sellem, L., Srour, B., Javaux, G., Chazelas, E., Chassaing, B., Viennois, E., Debras, C., Druesne-Pecollo, N., Esseddik, Y., Szabo de Edelenyi, F., Arnault, N., Agaesse, C., De Sa, A., Lutchia, R., Huybrechts, I., Scalbert, A., Pierre, F., Coumoul, X., Julia, C., … Touvier, M. (2024). Food additive emulsifiers and cancer risk: Results from the French prospective NutriNet-Sante cohort. PLoS Medicine, 21(2), e1004338. https://doi.org/10.1371/journal.pmed.1004338
Dicken, S. J., Dahm, C. C., Ibsen, D. B., Olsen, A., Tjonneland, A., Louati-Hajji, M., Cadeau, C., Marques, C., Schulze, M. B., Jannasch, F., Baldassari, I., Manfredi, L., Santucci de Magistris, M., Sanchez, M.-J., Castro-Espin, C., Rodriguez Palacios, D., Amiano, P., Guevara, M., van der Schouw, Y. T., … Batterham, R. L. (2024). Food consumption by degree of food processing and risk of type 2 diabetes mellitus: A prospective cohort analysis of the European Prospective Investigation into Cancer and Nutrition (EPIC). The Lancet Regional Health – Europe, 46, 101043. https://doi.org/10.1016/j.lanepe.2024.101043
Narula, N., Wong, E. C. L., Dehghan, M., Mente, A., Rangarajan, S., Lanas, F., Lopez-Jaramillo, P., Rohatgi, P., Lakshmi, P. V. M., Varma, R. P., Orlandini, A., Avezum, A., Wielgosz, A., Poirier, P., Almadi, M. A., Altuntas, Y., Ng, K. K., Chifamba, J., Yeates, K., … Yusuf, S. (2021). Association of ultra-processed food intake with risk of inflammatory bowel disease: Prospective cohort study. BMJ, 374, n1554. https://doi.org/10.1136/bmj.n1554
Chassaing, B., Compher, C., Bonhomme, B., Liu, Q., Tian, Y., Walters, W., Nessel, L., Delaroque, C., Hao, F., Gershuni, V., Chau, L., Ni, J., Bewtra, M., Albenberg, L., Bretin, A., McKeever, L., Ley, R. E., Patterson, A. D., Wu, G. D., … Lewis, J. D. (2022). Randomized controlled-feeding study of dietary emulsifier carboxymethylcellulose reveals detrimental impacts on the gut microbiota and metabolome. Gastroenterology, 162(3), 743-756. https://doi.org/10.1053/j.gastro.2021.11.006
Chassaing, B., Raja, S. M., Lewis, J. D., Srinivasan, S., & Gewirtz, A. T. (2017a). Colonic microbiota encroachment correlates with dysglycemia in humans. Cellular and Molecular Gastroenterology and Hepatology, 4(2), 205-221. https://doi.org/10.1016/j.jcmgh.2017.04.001
Chassaing, B., Van de Wiele, T., De Bodt, J., Marzorati, M., & Gewirtz, A. T. (2017b). Dietary emulsifiers directly alter human microbiota composition and gene expression ex vivo potentiating intestinal inflammation. Gut, 66(8), 1414-1427. https://doi.org/10.1136/gutjnl-2016-313099
Chassaing, B., Koren, O., Goodrich, J. K., Poole, A. C., Srinivasan, S., Ley, R. E., & Gewirtz, A. T. (2015). Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome. Nature, 519(7541), 92-96. https://doi.org/10.1038/nature14232
EFSA Panel on Food Additives and Flavourings (FAF), Younes, M., Aquilina, G., Castle, L., Degen, G., Engel, K.-H., Fowler, P. J., Frutos Fernandez, M. J., Furst, P., Gurtler, R., Husoy, T., Manco, M., Mennes, W., Moldeus, P., Passamonti, S., Shah, R., Waalkens-Berendsen, I., Wright, M., Dusemund, B., … Gundert-Remy, U. (2022). Opinion on the re-evaluation of sodium carboxy methyl cellulose (E 466) as a food additive in foods for infants below 16 weeks of age and follow-up of its re-evaluation as food additive for uses in foods for all population groups. EFSA Journal, 20(12), e07665. https://doi.org/10.2903/j.efsa.2022.7665