Skip to content
GLP Loss
← Research
Evidence

Treating GLP-1 Constipation: What Has and Has Not Been Tested

The gastroenterology guideline gives its strong recommendations to polyethylene glycol and sodium picosulfate, conditional ones to fiber and magnesium, and no rating at all to docusate — which lost the only randomized trial it has. None of those trials enrolled anyone on a GLP-1.

Owen Castellanos9 min read
Constipation treatments, ranked by evidenceTen agents were graded. None was graded in this setting.Strong recommendationPolyethylene glycol, sodium picosulfate, prucaloprideConditional recommendationFiber, lactulose, senna, magnesium oxideReviewed by the panel: not at allDocusate, the softener most handouts name firstRandomized trials in GLP-1 patientsZero, across every agent listed aboveThe handout order and the evidence order are not the same order.

Constipation reaches the common-adverse-reaction table on both weight labels: 24% of semaglutide-treated adults against 11% on placebo, and 17%, 14% and 11% across the three tirzepatide maintenance doses against 5%.[1][2] That tirzepatide row is worth pausing on, because it runs backward: the rate falls as the dose triples and the weight loss grows. Why that happens is the subject of the mechanism article. What to do about it is a separate question, and it has a separate and much less flattering answer.

The treatment ladder has been graded, and not in the order people use it

In 2023 the American Gastroenterological Association and the American College of Gastroenterology jointly ran systematic reviews of the agents used in chronic idiopathic constipation — fiber, the osmotic laxatives polyethylene glycol, magnesium oxide and lactulose, the stimulants bisacodyl, sodium picosulfate and senna, the secretagogues lubiprostone, linaclotide and plecanatide, and the serotonin-4 agonist prucalopride — and graded each with GRADE. The panel issued ten recommendations.[3]

Five were strong: polyethylene glycol, sodium picosulfate, linaclotide, plecanatide and prucalopride. Five were conditional: fiber, lactulose, senna, magnesium oxide and lubiprostone.[3] Read that list against the advice attached to a weight-loss prescription and the ordering is close to inverted. Fiber and a stool softener lead almost every consumer page; fiber is a conditional recommendation, and the stool softener is not in the guideline at all.

Docusate is the one intervention that lost its own trial

Docusate sodium is the most-suggested and least-supported item on the list. A double-blind randomized trial gave 74 hospice patients ten days of either docusate plus sennosides or placebo plus sennosides. 35 patients received docusate and 39 received placebo, and there was no significant difference between the groups in stool frequency, stool volume, stool consistency, difficulty of evacuation or completeness of evacuation.[4] That is a palliative-care population rather than a weight-loss one, and it is the only randomized test of the question that exists.

A head-to-head trial had already put docusate against psyllium in 170 adults with chronic idiopathic constipation. Psyllium raised stool water content by 2.33% against 0.01% for docusate (P = 0.007), produced 359.9 g of weekly stool output against 271.9 g (P = 0.005), and delivered 3.5 bowel movements a week against 2.9 in the second treatment week (P = 0.02).[5] The agent marketed as a stool softener softened stool less than the fiber it is usually offered alongside.

There is a compounding footnote to this that is genuinely strange. A 2026 survey of 33 compounded semaglutide and tirzepatide products advertised as differing from the approved ones found that 48% combined the active drug with a mixture drawn from cyanocobalamin, glycine, niacinamide, docusate and ondansetron, and the authors reported little justification for adding docusate sodium to a subcutaneous product.[6] An oral stool softener with no randomized benefit by mouth is being injected. What is and is not knowable about those blends is covered in the vial article.

Among the osmotics, one has beaten the other in ten trials

A Cochrane review pooled every randomized trial comparing lactulose with polyethylene glycol — ten in total — and found polyethylene glycol better on stool frequency per week, on stool form, on relief of abdominal pain and on the need for additional products, in adults and in children alike.[7] Its stated conclusion is that polyethylene glycol should be used in preference to lactulose.

A separate systematic review of 41 randomized trials of over-the-counter products, each at least four weeks long, graded the evidence by strength rather than by popularity. Polyethylene glycol and senna earned the only grade A ratings. Psyllium, magnesium salts, bisacodyl, sodium picosulfate and fruit-based laxatives earned grade B. Nausea, bloating and abdominal pain were common across the category, and no serious adverse events were reported.[8] Bloating and nausea are already the two most reported effects of the drug these products would be added to, which is a cost the grading does not price.

Magnesium and senna: significant alone, not significant pooled

The one modern trial to put a stimulant and an osmotic in the same protocol randomized 90 patients — mean age 42, 93% women, mean symptom duration 9.9 years — to senna 1.0 g, magnesium oxide 1.5 g or placebo for 28 days. The response rate for overall improvement was 69.2% on senna, 68.3% on magnesium oxide and 11.7% on placebo (P < 0.0001).[9] Both beat placebo by roughly six-fold and neither beat the other.

Pool senna with the rest of its evidence, though, and it stops clearing the bar. A meta-analysis of eight randomized trials in 787 adults found magnesium oxide holding up — 68% responded against 19% on control, a risk ratio of 3.32 (95% CI, 1.59 to 6.92; P = 0.001), with stool frequency up by 3.72 bowel movements a week (95% CI, 1.41 to 6.03) — and senna not: 61% against 28%, risk ratio 2.78 (95% CI, 0.93 to 8.27; P = 0.07), an interval that crosses one.[10] The same agent is a clear winner in one trial and an unproven one across the set, which is what a two-trial evidence base looks like from both sides.

The fastest agent is the one ranked last on tolerability

A network meta-analysis of 33 randomized trials covering 17,214 patients ranked the drugs against each other at two time points. At four weeks the diphenyl methane stimulants — bisacodyl and sodium picosulfate at 10 mg once daily — ranked first on failure to reach three or more complete spontaneous bowel movements a week (RR 0.55; 95% CI, 0.48 to 0.63; P-score 0.99). At twelve weeks prucalopride 2 mg ranked first (RR 0.82; 95% CI, 0.78 to 0.86; P-score 0.96).[11]

Then the same analysis ranks them on harm, and bisacodyl comes last for total adverse events and for abdominal pain, at a P-score of 0.08.[11] The agent that works fastest is the agent most likely to hurt, and abdominal pain is not a neutral side effect to add to a drug that already produces it — that symptom has its own thresholds, set out in the abdominal pain article. The authors also note that the trial durations were four to twelve weeks, so the long-term relative efficacy of every one of these agents is unknown.[11]

The interaction nobody has measured is with fluid

Polyethylene glycol works by holding water inside the bowel lumen. That water has to come from somewhere, and it comes from a person whose meals have shrunk. Both weight labels carry a warning for acute kidney injury arising from volume depletion, and both state that the reported cases mostly followed gastrointestinal reactions causing dehydration.[1][2] No trial has measured what an osmotic laxative does to fluid balance in someone taking one of these drugs, in either direction. The dehydration side of that is in the dehydration article.

The census, with its controls

Every number above was produced in chronic idiopathic constipation, not in a person taking a GLP-1. The guideline's systematic reviews covered that condition.[3] The 33 trials in the network meta-analysis covered it.[11] The 41 over-the-counter trials covered it.[8] The four systematic reviews behind the 2025 British Dietetic Association dietary guideline pooled 75 randomized trials and produced 59 statements, of which 12 rested on very low evidence, 39 on low evidence and 8 on moderate evidence — none of them in this population.[12]

The GLP-1-specific guidance says so itself. The most-cited practical management paper states in its own abstract that it draws on published evidence and the authors’ collective clinical experience.[13] A 2026 nutrition framework for incretin therapy states that direct trials of structured nutrition interventions in GLP-1 or dual-incretin populations remain limited and that several of its recommendations are extrapolated from the broader obesity, caloric restriction and gastrointestinal literature.[14] Those are honest papers. They are not trials of what to take.

One real-world record points the other way entirely

A retrospective analysis of a federated electronic health record network matched patients with irritable bowel syndrome who started a GLP-1 against those who did not — 6,665 per group in the 90-day landmark cohort. Coded chronic constipation was lower in the treated group: 19.8% against 22.0%, alongside lower chronic diarrhea (8.9% against 10.6%), abdominal pain (31.6% against 35.8%) and bloating (8.3% against 10.9%), all at P < 0.001.[15]

That is an observational study in people with irritable bowel syndrome, it counts billing codes rather than symptoms, and the authors describe it as hypothesis-generating. It cannot overturn a randomized adverse-event table. It does mean that the two instruments available — what trial participants reported when asked, and what clinicians coded when they saw someone — disagree about the direction of the effect, and the honest summary carries both.

Where that leaves a cash-pay intake

Most sellers covered here dispense compounded semaglutide or tirzepatide, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed. Every incidence figure above came from a branded product at a labeled dose, and every treatment figure came from a trial that excluded this drug entirely.

Two things do not depend on any of that. A laxative is an oral medication taken alongside another oral medication, and timing rules apply — the absorption article sets out which ones. And hard or infrequent stool with severe abdominal pain, vomiting, abdominal distension or no passage of gas is not the ordinary version of this; it belongs with a prescriber the same day rather than with another supplement, for the reasons in the obstruction article.

Frequently asked

What does the guideline actually recommend first?
The joint American Gastroenterological Association and American College of Gastroenterology guideline issued strong recommendations for polyethylene glycol, sodium picosulfate, linaclotide, plecanatide and prucalopride, and conditional recommendations for fiber, lactulose, senna, magnesium oxide and lubiprostone. That grading was done in chronic idiopathic constipation, not in people taking a GLP-1. Docusate was not among the agents the panel reviewed.
Does a stool softener help?
The randomized evidence says no. In a double-blind trial of 74 hospice patients, docusate added to sennosides produced no significant difference from placebo added to sennosides on stool frequency, volume, consistency, or difficulty of evacuation. In a separate 170-patient trial in chronic idiopathic constipation, psyllium beat docusate on stool water content, weekly stool output and bowel movement frequency.
Is magnesium better than senna?
In the one trial that compared them directly, no: 68.3% responded to magnesium oxide and 69.2% to senna, both against 11.7% on placebo across 90 patients over 28 days. Pooled across the wider literature the two separate, with magnesium oxide holding a risk ratio of 3.32 against control and senna at 2.78 with a confidence interval that crosses one.
Why does the tirzepatide label show less constipation at higher doses?
The label reports constipation in 17% at 5 mg, 14% at 10 mg and 11% at 15 mg, against 5% on placebo. The rate falls as the dose rises, which is the opposite of what a simple dose-response would predict and is not explained anywhere on the label. It is one reason to treat any confident mechanistic story about this effect with caution.
Has anything been tested in people actually taking these drugs?
No laxative, fiber supplement or fluid protocol has been tested in a randomized trial of people taking a GLP-1 receptor agonist. The most-cited practical management paper states that it draws on published evidence and its authors' collective clinical experience, and a 2026 nutrition framework states that its recommendations are extrapolated from the broader obesity and gastrointestinal literature because direct trials remain limited.
When does constipation stop being routine?
Hard or infrequent stool that responds to an osmotic laxative is the ordinary presentation. Severe abdominal pain, vomiting, abdominal distension, or no passage of stool or gas for several days is a different problem and warrants same-day contact with a prescriber rather than another over-the-counter product. Both weight labels also warn about acute kidney injury from volume depletion following gastrointestinal reactions.

Sources

  1. [1] Novo Nordisk Pharmaceutical Industries, LP (2026). WEGOVY (semaglutide) injection, solution — Adverse Reactions 6.1, Table 3, and Warnings and Precautions 5.6 Acute Kidney Injury DailyMed, U.S. National Library of Medicine. Source
  2. [2] Eli Lilly and Company (2026). ZEPBOUND (tirzepatide) injection, solution — Adverse Reactions 6.1, Table 1, and Warnings and Precautions 5.5 Acute Kidney Injury DailyMed, U.S. National Library of Medicine. Source
  3. [3] Chang L, Chey WD, Imdad A, et al. (2023). American Gastroenterological Association-American College of Gastroenterology Clinical Practice Guideline: Pharmacological Management of Chronic Idiopathic Constipation. Am J Gastroenterol. PMID 37204227
  4. [4] Tarumi Y, Wilson MP, Szafran O, Spooner GR (2013). Randomized, double-blind, placebo-controlled trial of oral docusate in the management of constipation in hospice patients. J Pain Symptom Manage. PMID 22889861
  5. [5] McRorie JW, Daggy BP, Morel JG, et al. (1998). Psyllium is superior to docusate sodium for treatment of chronic constipation. Aliment Pharmacol Ther. PMID 9663731
  6. [6] Belcourt J, Sapowadia A, White CM (2026). Compounded Semaglutide and Tirzepatide Products Use Unique Formulations but Efficacy and Safety Largely Unknown. Ann Pharmacother. PMID 41689811
  7. [7] Lee-Robichaud H, Thomas K, Morgan J, Nelson RL (2010). Lactulose versus Polyethylene Glycol for Chronic Constipation. Cochrane Database Syst Rev. PMID 20614462
  8. [8] Rao SSC, Brenner DM (2021). Efficacy and Safety of Over-the-Counter Therapies for Chronic Constipation: An Updated Systematic Review. Am J Gastroenterol. PMID 33767108
  9. [9] Morishita D, Tomita T, Mori S, et al. (2021). Senna Versus Magnesium Oxide for the Treatment of Chronic Constipation: A Randomized, Placebo-Controlled Trial. Am J Gastroenterol. PMID 32969946
  10. [10] van der Schoot A, Creedon A, Whelan K, Dimidi E (2023). The effect of food, vitamin, or mineral supplements on chronic constipation in adults: A systematic review and meta-analysis of randomized controlled trials. Neurogastroenterol Motil. PMID 37243443
  11. [11] Luthra P, Camilleri M, Burr NE, et al. (2019). Efficacy of drugs in chronic idiopathic constipation: a systematic review and network meta-analysis. Lancet Gastroenterol Hepatol. PMID 31474542
  12. [12] Dimidi E, van der Schoot A, Barrett K, et al. (2025). British Dietetic Association Guidelines for the Dietary Management of Chronic Constipation in Adults. J Hum Nutr Diet. PMID 41081513
  13. [13] Wharton S, Davies M, Dicker D, et al. (2022). Managing the gastrointestinal side effects of GLP-1 receptor agonists in obesity: recommendations for clinical practice. Postgrad Med. PMID 34775881
  14. [14] Zambrano-Villacres R, Campuzano-Donoso M, Reytor-González C, et al. (2026). Nutrition-First Support for GLP-1 and Dual Incretin Therapy in Obesity: A Practical Framework for Dietary Management, Symptom Tolerability, and Long-Term Weight Maintenance. Nutrients. PMID 42280393
  15. [15] Khan SA, Marasini A, Shrestha A, et al. (2026). Impact of GLP-1 Analogue Therapy on Gastrointestinal Outcomes in Patients with Irritable Bowel Syndrome: A Real-World TriNetX Analysis. Dig Dis Sci. PMID 42570075

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence