Digestive enzymes: when they really make sense

The tub sits in the kitchen cupboard, between the magnesium and the turmeric. On the label are five enzyme names, a number in units and the promise that things will feel lighter after eating.

It was taken for two weeks, then irregularly, then not at all. What remains is the question: was that nonsense, or did I simply do it wrong?

Many people know this pattern. Enzymes are among the most bought digestive products of all, and at the same time among the most poorly explained.

That is not because there are no data. It is because three completely different things are sold under the same word. A medicine covered by guidelines for a genuine organ weakness. A food supplement for vague complaints. And a product that suggests a safety in coeliac disease which it does not deliver.

In this text I keep these three areas strictly apart. And for each one I say how strong the evidence is.

One note that comes before everything else. Blood in the stool, black tarry stool, unintended weight loss, fever, vomiting, difficulty swallowing, anaemia, night time symptoms that wake you, or a new persistent change in bowel habit from around 45 to 50 years of age need medical assessment and must not be treated with capsules. The full list is further down in the red flags box. A recommended endoscopy, colonoscopy or laboratory workup is not replaced by any product, and a self experiment with enzymes is no reason to put it off.

Four stages, four tools, and a capsule reaches only one of them

Before you decide whether you need enzymes, you need a different answer. Which step do you actually want to replace?

That sounds like a biology lesson, but it decides the choice of product. Digestion is not a single event, it is a chain of four stations, and each works with a different tool in a different environment.

Station one is the mouth. Saliva contains amylase, which breaks starch open before the bite is even swallowed. Anyone who eats fast largely skips this step, and no capsule can make up for it.

Station two is the stomach. Pepsin takes protein apart and works only in an acidic environment. Without sufficient acid its precursor pepsinogen simply stays a precursor. Fat splitting also begins here.

Station three is the small intestine, and this is where the real turnover happens. The pancreas releases lipase, amylase and proteases, and bile breaks fat into fine droplets so that lipase can attack it at all. If a capsule can achieve anything anywhere, it is here. Because here the contents lie free in the gut lumen, and an added enzyme can mix in.

Bile is a topic of its own. Without bile acids fat stays in large droplets, and lipase reaches the surface poorly. The details are in Bile, gallbladder and TUDCA.

Station four is the gut wall itself. And here it becomes decisive for the buying decision. The last sugar splitters do not sit in the gut contents but in the membrane of the mucosal cells. Lactase belongs there, and so does sucrase isomaltase.

The case in which enzymes are beyond dispute

Now the area where the evidence is strongest. When the pancreas genuinely releases too little, supplying it from outside is not a food supplement but well documented therapy. The technical term is exocrine pancreatic insufficiency. Exocrine refers to the part of the organ that releases digestive juice into the gut, as distinct from the part that releases insulin into the blood.

This insufficiency almost always has a cause. That is precisely why, when it is suspected, I press for medical assessment rather than for a product.

The cardinal features are fatty stool with or without diarrhoea, unintended weight loss, bloating, marked flatulence and a lack of fat soluble vitamins. Fatty stool does not simply mean soft stool. It is bulky, pale, greasy looking, hard to flush and smells different.

And now the number that explains all the rest of this article.

On treatment itself I keep it short, because it does not belong in a blog text but in a consultation. For orientation, not as instruction: the active substance is called pancreatin and is obtained from porcine pancreas. High dose pancreatic enzymes are prescription only medicines in Germany and not food supplements. As an orientation value for clinicians, the AGA Clinical Practice Update of 2023 names a start in the region of 40,000 USP units of lipase per main meal in adults. How high the dose is in an individual case, whether it is increased at all and how long treatment lasts is decided by the treating doctor according to findings, weight and course. A figure from a guideline is not a dose for you.

One further sentence from the AGA is worth noting: a very low fat diet is explicitly not recommended in pancreatic insufficiency. Someone who already absorbs too little should not additionally take in less.

The test that rules out better than it rules in

Suppose you have had the test done. The report says pancreatic elastase, a number and a reference range, and the number is below the limit. The first thought is almost always the same: so it is that after all.

This is exactly where I would like to slow you down, not out of scepticism towards the laboratory but out of respect for what the test can and cannot do.

Pancreatic elastase 1 is an enzyme from the pancreas that passes through the gut largely unchanged. Its amount in stool therefore allows conclusions about output. That is elegant, non invasive and cheap, and the AGA names it the appropriate first test.

The cut off values are clearer than they are often reported. Below 100 micrograms per gram counts as good evidence for insufficiency. The range between 100 and 200 is explicitly described by the AGA as indeterminate. That is not a slip of the pen but a statement. On top of that comes a practical trap that appears on hardly any German website.

Now to the numbers that decide how you should read your own report.

And now the finding that I consider the most important in this whole article. It comes from a meta-analysis of 2018, and it explains why the same value means two different things in two different people.

That is not a dismissal of the test but a set of instructions. Where prior probability is high, that is after pancreatitis, after surgery or in cystic fibrosis, elastase is a very useful tool. Where prior probability is low, it mainly produces uncertainty.

How often does a genuinely low value actually occur in the irritable bowel group? There is a British investigation on this that is often cited as evidence for enzymes in irritable bowel syndrome. It says something different from what the quotations suggest.

The remaining stool diagnostics are covered in detail in Stool testing and dysbiosis diagnostics. Here only this applies: elastase is a function test and not a microbiome test, and the two answer different questions.

The grey zone: bloating and fullness without a proven deficiency

This is where most people stand who buy enzymes. No fatty stool, no weight loss, normal elastase or no test at all. What there is instead is a belly that feels tight after eating, and a fullness that lingers longer than it should.

For bloating as a topic in its own right, that is where the air actually comes from and what influences it, there is a separate article: Bloating, where the air comes from. Here it is only about what enzymes actually achieve in this situation.

The answer is: it depends on which enzyme, at which meal and in whom. I show you the data in both directions, because both exist.

Alpha galactosidase and the pulses

Beans, lentils and chickpeas contain galacto oligosaccharides, GOS for short. Human enzymes cannot split them, so they reach the large bowel and are fermented there. Alpha galactosidase from moulds breaks this bond in the laboratory. So far the idea, and it is a good one.

Anyone who cites only this study is doing marketing. So here is the other side, and it is larger.

And the original study from 1994, 19 people, one chilli test meal, did find fewer flatulence events per hour, but explicitly no difference in bloating and none in pain. The enzyme addressed the gas, not the sensation. This distinction runs through the entire field.

Lactase and the question of amount

With lactase things are cleaner, because here cause and tool match each other exactly. And because there is a study that shows the basic principle of all enzyme preparations in a single row of numbers.

At the same time lactase is not the only lever. A systematic review from 2020 states that symptoms in lactose malabsorption depend on several factors at once: the amount of lactose, the residual activity of your own lactase, the foods eaten alongside, gut transit time and the composition of the microbiome. Many of those affected tolerate normal amounts of milk without trouble, especially as yoghurt or cheese, because live bacteria there have already digested part of the lactose.

And a counter example that belongs here for the sake of honesty: in infant colic the idea of giving lactase seems obvious. A systematic review with five randomised trials and 391 infants found no difference in crying time and none in fussing time. A mechanistically sensible enzyme is evidently not automatically a useful one.

Pancreatin with a very fatty meal

Alongside this there is a positive study of a multi enzyme complex in functional dyspepsia, 40 participants, 60 days, a difference in all five efficacy measures. In the same breath belongs the fact that all seven authors were employed by the manufacturer or its contract research arm and that only one centre took part. That is what a large part of the positive enzyme literature looks like.

Where the evidence ends

For enzymes in irritable bowel syndrome in general, in bacterial overgrowth, in increased gut permeability or in non specific intolerances I found no robust human data. And in one case the data even point in the opposite direction: bacterial overgrowth in the small intestine can prevent fat digestion from returning to the normal range even under reliable enzyme dosing. Anyone who has overgrowth and takes enzymes is therefore working against a brake that sits elsewhere. More on this in SIBO in the small intestine and IBS, finding the causes.

One more note on a claim that circulates on several German websites. There is talk of a meta-analysis spanning sixty years of research that supposedly recommends digestive enzymes for irritable bowel type complaints. No such work could be found in this research. What exists are individual studies with contradictory results, and those are set out above.

Gluten enzymes: why they give no safety in coeliac disease

This section is the most important in the article, and I write it deliberately soberly.

And it begins with the sentence that in practice protects the most. Anyone who suspects coeliac disease should not go gluten free on their own before the diagnostic workup. Antibodies in the blood and a tissue sample from the small intestine both require that gluten continues to be eaten. Leaving it out prematurely can make the diagnosis impossible. The order of steps therefore belongs in a medical conversation before anything is changed about the diet.

On the market there are enzyme products aimed at gluten, often called glutenases. The names sound like protection. Whether this protection exists has been well studied, and the answer is clearer than the packaging suggests. First the mechanism, because it is real and instructive.

The mechanism even works measurably well. That is the point at which many people make the thinking error.

And now the study that matters. It is the largest in this field, it has hard endpoints, and it is negative.

It would be unfair to stop here, because there is one study in which an enzyme did protect the mucosa. Its conditions, however, are very narrow.

And then there is a finding that is perhaps the most uncomfortable one for everyday life. It concerns the question of whether you can read off your own gut feeling whether an enzyme has protected you.

The most recent everyday investigation comes from 2024. Forty adults with coeliac disease took AN-PEP or placebo with every meal for four weeks. The primary endpoint was missed, gluten peptides in stool did not fall significantly, while the proportion of symptomatic patients was lower than in the run in phase. The authors call their own study exploratory. The most revealing finding is a different one: in 65.6 percent of all samples no gluten was detectable at all. There was simply little that an enzyme could have broken down.

And in gluten sensitivity without coeliac disease? There too the measurable reduction changes little. In a cross-over study with bread whose gluten had been about 40 percent pre digested, the digestive symptoms triggered did not differ from the control bread. That fits a thought set out in detail in the article Gluten without coeliac disease: in gluten sensitivity, gluten is often not the only and not the most important trigger.

Why for me stomach acid comes before the enzymes

Now I lay open my own practice rule. Because this is a rule and not a body of evidence, I mark it as such.

When someone comes to me with fullness and has already tried a tub of enzymes, I do not ask about the pancreas first. I ask about what is happening further up. The thinking behind it has two parts, and both can be justified physiologically.

First: pepsin needs acid in order to come into being at all. The stomach releases pepsinogen, an inactive precursor. Only the acidic pH cuts active pepsin out of it. Without sufficient acid, protein digestion therefore begins more weakly, and what is left undone up there arrives further down in larger pieces.

Second: the acid that reaches the small intestine is a control signal. The duodenum in effect gauges how acidic the arriving chyme is and calls up bicarbonate, pancreatic enzymes and bile in response. There is an old, small but clean human investigation on this.

There is a further piece of work pointing in the same direction. I name it explicitly with its limits, because it is not a human study.

And now the paragraph that in my view makes this article stronger than any assertion.

On stomach acid itself I deliberately write nothing further here. How a lack of acid arises in the first place, how it is appraised and what to make of betaine HCl is set out in full in Low stomach acid and betaine HCl. We do not need two explanations of the same thing in this cluster.

Animal, plant, microbial: what the pH gradient makes of it

On the shelf two worlds stand side by side. On one side pancreatin with units for lipase, amylase and protease. On the other side enzyme complexes with bromelain, papain, fungal and bacterial enzymes, sold as food supplements. The difference is bigger than the packaging suggests, and it has to do with the pH gradient.

Pancreatin comes from porcine pancreas. The AGA Clinical Practice Update states explicitly that all preparations come from the same source and are equally potent at the same dose. Differences between brands therefore concern the formulation, not the potency.

And the formulation is precisely the interesting part.

Timing is the real art. Enzyme and chyme have to arrive in the small intestine at the same time. That is why the AGA names intake during the meal. Swallowed too early, the capsule leaves the stomach before the food. Swallowed too late, it lags behind. In both cases the enzyme passes the work by.

Now to an apparent contradiction that I would rather name openly than dissolve out of sight.

And the plant and microbial enzymes? Here I have to stay honest, even though a clearer answer would be more comfortable. Bromelain from pineapple and papain from papaya are cysteine proteases with their own pH optima. Fungal and bacterial enzymes cover partly different ranges, and some of them are more acid stable than pancreatin.

The pH argument can be explained. What remains open is whether these enzymes arrive active in the small intestine in relevant amounts in humans. Robust human studies on this I did not find in this research. Nothing is thereby refuted, but nothing is established either.

The same applies to the notion that pineapple and papaya as foods are a noteworthy enzyme source for your own digestion. That is a claim without a human study.

When enzymes are the wrong tool, and what would then need clarifying

That leaves the question that is practically never asked online in German. Not whether enzymes are of use, but when they are the wrong tool for your problem. The most common objection to this runs: but I do notice that it gets better. I take that seriously, and it still does not hold.

The AGA explicitly names the therapeutic enzyme trial as an unreliable diagnostic criterion. And how strongly expectation carries this field was shown impressively by the latiglutenase study with 494 people: all groups improved markedly, including on placebo, measured by symptoms and by tissue samples.

There is a second, harder reason why wellbeing is not a good yardstick here.

A review of 257 papers and guidelines from 2024 fits with this. Dosing recommendations differ considerably, within individual disease groups as much as between world regions, and most studies investigate safety and benefit rather than optimal titration. That is precisely why the dose question in a genuine insufficiency belongs in medical care.

What would then need clarifying instead? I put it as questions and not as a protocol, because the answers are individual.

On the fifth question one more number that gives food for thought but should be read carefully. A Turkish evaluation examined 980 people who had undergone exome sequencing for other reasons, and in addition 148 children with disorders of gut brain interaction. In the first group symptomatic genetic sucrase isomaltase deficiency was found in 0.3 percent, in the second in 10 percent. Those are children from a single centre, and these figures are not transferable to German adults. As food for thought they still serve: in a group with irritable bowel type complaints there occasionally sits a genuine but very specific enzyme deficiency.

And now the finding with which I would like to end this article. It is the strongest evidence that enzyme weakness can also be a consequence and not a cause.

Frequently asked questions about digestive enzymes

Where digestion connects to the rest of the body

An enzyme question rarely stands alone. It hangs on what is eaten, on the uptake of nutrients, on inflammatory load and on how strongly the belly is perceived at all. How these building blocks fit into an overall picture is set out in The gut reset, the whole picture. These articles open the neighbouring rooms.