Isotretinoin and Bowel Disease: Separating Symptoms, Reports, and Diagnoses

By Michael Wiederkehr, M.D.

A digestive symptom during isotretinoin treatment raises two different questions. What should happen now to assess and protect the patient? And what does the research show about whether isotretinoin increases the risk of a particular bowel disease? The first question may require action before the second can be resolved.

The distinction is especially important because online discussions often combine diarrhea, abdominal pain, irritable bowel syndrome, inflammatory bowel disease, and celiac disease under a single label. These are not interchangeable. A symptom is not a diagnosis, and a diagnosis made after treatment begins does not by itself prove that treatment caused it.

The available evidence includes inflammatory bowel disease reviews, adverse-event reporting analyses, and a separate investigation of celiac disease. This article uses those documents alongside primary comparative studies and official digestive-health information. It includes a later Korean study with a positive association, rather than presenting an older reassuring meta-analysis as the final word on every population.

The practical starting point is clear: current isotretinoin labeling instructs patients to discontinue treatment and contact the prescriber for abdominal pain, rectal bleeding, or severe diarrhea. ^[1] Reassuring population evidence is not permission to ignore those symptoms. At the same time, a symptom should not be converted into a lifelong diagnosis without appropriate assessment.

A map of the terms

Inflammatory bowel disease, abbreviated IBD, includes conditions such as Crohn's disease and ulcerative colitis. Irritable bowel syndrome, abbreviated IBS, is a different condition. Celiac disease is another distinct disorder, despite overlapping digestive symptoms.

NIDDK describes Crohn's disease as chronic inflammation that can affect different parts of the digestive tract. Ulcerative colitis affects the inner lining of the large intestine. IBS involves recurring symptoms without the same pattern of visible inflammatory injury, while celiac disease involves an immune response to gluten that damages the small intestine. ^[2][3][4][5]

The table is a vocabulary guide, not a self-diagnostic tool. Its purpose is to make the research readable. A result about one row should not be transferred to another simply because both involve the bowel.

Why the abbreviation CD can create a serious error

In papers about inflammatory bowel disease, CD commonly means Crohn's disease. In papers about celiac disease, the same letters can mean celiac disease. Both meanings occur in the literature discussed here.

This is not a minor editorial issue. If a table from the celiac investigation is read as though CD means Crohn's disease, the article could attribute the wrong outcome to the study. The resulting conclusion would be medically inaccurate even if every number were copied correctly.

The reliable approach is to read the title, methods, outcome definition, and table legend before extracting results. An abbreviation should be expanded in the working notes and article whenever there is a realistic chance of confusion.

The same source discipline applies to UC, IBD, IBS, and broad categories such as gastrointestinal disorders. A report grouped under a gastrointestinal system category is not automatically an IBD case. It may describe a symptom or a different diagnosis entirely.

A practical source note should therefore include the full outcome name, how it was identified, and the population. "CD risk" is too ambiguous. "Recorded Crohn's disease identified by specified codes in an acne cohort" is much clearer.

This article keeps the conditions separate throughout. That allows an accurate discussion of reassuring evidence, positive associations, and symptom precautions without blending different diseases into a single misleading claim.

Symptoms require assessment before causal certainty

NIDDK explains that diarrhea can arise from infections, medicine effects, food-related problems, and other digestive conditions. It also identifies warning signs such as severe pain, blood or black stools, dehydration, frequent vomiting, and concerning persistence. ^[6]

For someone taking isotretinoin, the medication history is relevant, but it is not the only information needed. The clinician may need the onset, stool pattern, pain location, bleeding, fever, recent antibiotics, travel, other exposures, and previous digestive history.

The fact that several causes are possible should not become an excuse for delay. Assessment can proceed while the cause remains uncertain. A clinician can give protective instructions and investigate without claiming that the medicine has definitely caused a chronic disorder.

The reverse is also important. A patient should not be told that the drug cannot be involved solely because a large study was reassuring. Population evidence changes the background assessment; it does not examine the person currently reporting symptoms.

The useful order is:

1. Identify the actual symptoms and immediate severity.
2. Follow product-specific instructions and obtain appropriate medical advice.
3. Investigate likely explanations according to the presentation.
4. Establish whether a diagnosis is present.
5. Assess the possible contribution of medicines and other factors.

These steps may overlap in clinical care. They should not be turned into a requirement that the patient prove causation before receiving help.

What the 2023 meta-analysis found

A 2023 systematic review included eight observational studies with more than 2.5 million participants. Its pooled estimate did not show increased odds of IBD among isotretinoin users: the reported odds ratio was 1.01, with a 95% confidence interval of 0.80 to 1.27. ^[7]

That is reassuring evidence about the included populations and studies. It does not mean that the risk is mathematically zero, that every individual symptom is unrelated, or that later studies cannot add information. The confidence interval and underlying designs remain relevant.

The review also addresses defined IBD outcomes, not all digestive discomfort. A patient with diarrhea cannot use it to establish the cause of the symptom, and an article cannot use it to claim isotretinoin has no gastrointestinal adverse effects of any kind.

Observational studies can be large while remaining vulnerable to confounding, incomplete exposure information, and diagnosis ascertainment. Pooling improves the synthesis but does not automatically remove limitations shared by the included studies.

The appropriate conclusion is therefore narrow enough to be useful: the meta-analysis did not find an overall increased association in the evidence it combined. That finding should inform a balanced prescribing discussion while current symptom instructions remain in force.

What the dataset review materials reveal

The 2024 dataset contains supplementary materials for another IBD review, including a table of nine studies. The table shows different countries, cohort and case-control designs, observation periods, and available exposure details. Dose and duration are not reported consistently. ^[8]

Those differences explain why a single headline should not conceal the structure of the evidence. A population cohort, an insurance case-control study, and a specialty-clinic record review do not collect the same information in the same way.

The supplementary table also helps distinguish total database size from the number actually exposed to isotretinoin. A study can include a very large comparison population while the exposed group or number of relevant events is much smaller. The headline total should not be mistaken for millions of treated patients.

The supplementary tables help compare study characteristics and definitions, but they should not be read as a new independent analysis. The numerical conclusions must remain connected to the corresponding published studies and their actual methods.

For patients, this means that several records may describe overlapping evidence rather than several independent confirmations. A supplementary table, a journal abstract, and a later review can all refer to the same underlying participants. Counting the links instead of the studies can overstate confidence.

The global isotretinoin-versus-antibiotic study

A large retrospective study compared 77,005 acne patients starting isotretinoin with 77,005 starting oral antibiotics. Over the study's overall follow-up, Crohn's disease and ulcerative colitis estimates were not statistically different between groups. A time-stratified analysis found an early ulcerative colitis increase during the first six months, with about five additional cases per 10,000 isotretinoin starters. ^[9]

This combination is not contradictory. An association can differ across time windows, and an overall estimate can conceal a short-term pattern. The result needs both the time period and the absolute scale.

The study's abstract uses the phrase "lifetime risk," but this should not be presented to patients as observation of every participant for their entire natural lifespan. It is an estimate from a healthcare database and its available follow-up. Wording should reflect that limitation.

The antibiotic comparator is clinically relevant because both groups received systemic acne treatment. It does not make them identical. Prior treatment, acne burden, healthcare use, and other factors can differ despite statistical matching.

The practical reading is therefore specific: the study was broadly reassuring over its overall comparison but identified a small absolute early ulcerative colitis difference that deserves acknowledgment. Neither "no signal at all" nor "everyone has a major bowel risk" accurately describes the result.

A later Korean study found a positive association

A 2024 nationwide Korean case-control study included 107,434 people with IBD and 393,830 controls. It reported an adjusted association between prior isotretinoin exposure and IBD, with an odds ratio of 1.20 and a 95% confidence interval of 1.10 to 1.30. Separate analyses also reported associations for Crohn's disease and ulcerative colitis. ^[10]

This study should not be omitted because its direction differs from the earlier pooled conclusion. It adds evidence from another population and design. A balanced article should acknowledge it and explain why one new observational result does not alone establish causation or automatically overturn every prior finding.

The authors examined exposure during the preceding five years and reported patterns by exposure characteristics. ^[10] The interpretation still depends on the data source, covariates, selection, and how diagnoses and prescriptions were identified.

An odds ratio of 1.20 is not a statement that 20% of isotretinoin users developed IBD. It is a relative association from the case-control analysis. The case and control counts also do not provide an incidence rate among all treated people.

The appropriate response is continued appraisal, including whether design differences, population differences, residual confounding, or a true association explain variation across studies. Declaring the question permanently settled in either direction would overstate the available evidence.

Why different studies can disagree

Study results can differ because the underlying question differs. A case-control study starts with disease status and examines previous exposure. A cohort starts with exposure groups and follows outcomes. Their strengths and biases are not identical.

The table lists possible explanations, not proven reasons that any named study is wrong. A careful comparison investigates them rather than selecting whichever study produces the preferred message.

A finding that remains after adjustment is more informative than an unadjusted comparison, but adjustment does not create randomization. A factor may be recorded imperfectly, omitted, or represented by a category that does not capture its severity.

Likewise, a dose-response pattern can strengthen a hypothesis but does not independently eliminate every alternative explanation. More cumulative prescriptions may correlate with other differences in treatment history or healthcare use. The causal argument needs the whole evidence pattern, not one attractive feature.

Acne itself and earlier treatment are part of the context

A 2023 propensity-matched study examined acne, oral antibiotic treatment, and isotretinoin in relation to incident IBD. It reported an association with acne itself rather than a clear added association from the studied medicines. ^[11]

This is relevant because treatment does not begin in a neutral background. People receiving isotretinoin have acne and often a previous treatment history. Comparing them with people without acne can mix the condition's associated factors with the medicine's possible contribution.

Earlier antibiotic exposure may also matter to study design. It can be both part of acne care and a factor requiring measurement in digestive-outcome research. A study must decide how to define and adjust for it rather than assume it is absent.

The purpose of this context is not to declare acne or antibiotics the cause of every observed association. It is to identify why the medicine cannot be evaluated accurately without considering the pathway that led to prescribing.

Patients can contribute useful history by reporting prior systemic acne treatments and the timing of digestive symptoms. The clinician does not need a perfect diary of every past dose, but the broad sequence can distinguish symptoms that predated isotretinoin from those that appeared later.

A source summary should preserve that same sequence. "IBD diagnosed after isotretinoin" is an observation. "IBD began because of isotretinoin" is a causal conclusion requiring additional evidence.

Absolute and relative risk should appear together

The global study's early ulcerative colitis result illustrates why relative and absolute measures should be reported together. A near-doubling relative estimate can coexist with a small absolute excess when the underlying event is uncommon. ^[9]

Consider a separate invented example. If an event occurs in five of 10,000 people in one group and ten of 10,000 in another, the relative risk is two and the absolute difference is five per 10,000. The arithmetic does not decide whether the difference matters clinically; it makes the scale visible.

A case-control odds ratio answers another kind of comparison and should not be translated directly into an absolute probability. To estimate a personal or population risk, the baseline frequency, time window, and appropriate assumptions are needed.

The goal is proportionate understanding. An alarming relative number should not be used without scale, and a small absolute number should not be used to dismiss someone who has developed concerning symptoms.

A patient currently experiencing bleeding needs clinical assessment regardless of how uncommon a diagnosis is in a cohort. Epidemiology informs the background; it does not replace the immediate assessment.

Adverse-event reports are not incidence studies

Pharmacovigilance analyses also include gastrointestinal reports. FDA explains that spontaneous reports do not establish causation and cannot be used to calculate event incidence. Reports can also be incomplete or duplicated. ^[12]

A broad gastrointestinal report count does not equal a count of confirmed IBD diagnoses. Even a report mentioning a diagnosis needs clinical review before it is treated as evidence of a causal medication effect.

Imagine an invented safety database with 1,000 gastrointestinal reports. If 100 mention diarrhea, that is 10% of the reports in the chosen category, not 10% of all people taking the drug. The denominator is different. The same error can occur when a signal table is summarized without explaining what is being counted.

Spontaneous reporting remains useful. It can identify unusual patterns and questions that need further investigation. Its limitations define the conclusions it can support, not whether patients' experiences deserve attention.

A sound article uses reporting data to describe suspected concerns and uses appropriately designed comparative studies to discuss population associations. It does not force one source type to do the job of another.

For an individual, submitting a report and receiving medical care are also separate actions. A report does not replace diagnosis or treatment, and uncertainty about whether to report should not delay assessment of concerning symptoms.

A diagnosis date may not be a disease-onset date

Administrative data often identify the date a diagnosis was recorded. That can differ from when symptoms began or when the underlying disease process started. The distinction matters when researchers examine a short interval after a prescription.

A hypothetical patient may have had intermittent bowel symptoms for months before beginning acne treatment but receive an IBD diagnosis only afterward. The sequence in the database can look like new disease after exposure even when earlier symptoms were not coded.

The opposite problem is possible too. A study may miss earlier diagnoses made outside the available system and classify a later record as incident disease. Methods that use lookback periods and repeated definitions can reduce some misclassification, but the available history remains a constraint.

This does not invalidate every study using records. It identifies a limitation to inspect. A well-designed analysis explains how prior disease was excluded, how outcomes were confirmed, and what sensitivity analyses were performed.

Patients should provide the actual symptom timeline during assessment. "Diagnosed last month" and "symptoms began last month" are different statements. The clinician needs both when considering the possible contribution of treatment.

For readers, the practical rule is to avoid treating a recorded sequence as a complete causal story. Timing is necessary to investigate a suspected effect, but it does not by itself establish the effect.

How Crohn's disease and ulcerative colitis are evaluated

NIDDK describes diagnosis as a combination of history, examination, and appropriate tests. Blood tests, stool tests, endoscopy, biopsies, and imaging may contribute according to the suspected condition. No single symptom establishes Crohn's disease or ulcerative colitis. ^[13][14]

The evaluation serves several purposes: determine whether inflammation is present, identify its location and pattern, assess severity, and consider other explanations. The clinician chooses tests according to the presentation rather than ordering every available investigation for every episode.

A person should understand what each test is intended to answer. A stool test for infection and an endoscopic biopsy are not interchangeable. A normal result in one category may not answer a different question.

The table explains the logic, not a personal workup. A reader should not demand a specific invasive test solely because isotretinoin appears on a medicine list. Equally, normal routine acne monitoring tests should not be treated as a complete digestive assessment.

The best outcome is a clear explanation of what is suspected, what remains uncertain, and how the next step will reduce that uncertainty.

IBS is a real condition, but it is not IBD

NIDDK describes IBS as a disorder involving recurring abdominal pain and changes in bowel movements without the same visible digestive-tract damage characteristic of inflammatory diseases. ^[4] The distinction does not mean IBS symptoms are imaginary or unimportant.

An IBS label should not be used casually to dismiss a new symptom that has not been assessed, especially when bleeding, significant weight loss, or other concerning features are present. The clinician needs to evaluate the actual presentation.

The global isotretinoin study examined IBS separately from Crohn's disease and ulcerative colitis. ^[9] That separation should remain in any summary. A lower recorded IBS association cannot be used to claim protection against IBD, and an IBD finding cannot automatically be transferred to IBS.

Patients sometimes hear "functional" as meaning no help is available. A better explanation identifies what diagnosis has been considered, what has been excluded or still needs assessment, and what symptom-management plan applies.

For the isotretinoin decision, the relevant questions are whether symptoms predated treatment, whether they have changed, what assessment has been completed, and whether the current plan remains appropriate. The answer depends on the individual's situation, not only on the diagnostic label.

Clear terminology helps preserve both seriousness and proportion. It avoids turning every bowel symptom into inflammatory disease while also avoiding the dismissal of distress when inflammatory disease is not established.

Celiac disease is a separate question in the search results

The search included a 2015 investigation of isotretinoin exposure and celiac disease, not Crohn's disease. ^[15] Its inclusion in a general search does not make it interchangeable with the IBD literature.

Celiac disease involves a particular immune response to gluten, and its diagnostic pathway differs. NIDDK advises testing before starting a gluten-free diet because dietary restriction can affect results. ^[16]

A patient who develops digestive symptoms should therefore avoid using a search result as a reason to diagnose themselves or make major dietary changes before assessment. If celiac disease is suspected, tell the clinician what has already changed in the diet.

The article should also avoid presenting the celiac paper's supplementary tables as proof about Crohn's disease. This is exactly where the abbreviation CD can cause an error. The full disease name should accompany the citation.

This section does not suggest that every isotretinoin user needs celiac screening. Testing should follow the clinical situation and relevant history. It explains why a distinct document appeared in the source set and how to use it accurately.

A research corpus can contain related but separate questions. Good synthesis does not combine them simply to make the article longer. It identifies what each document can contribute and where it should remain outside a particular causal claim.

A pre-existing bowel condition needs coordinated care

Evidence about new diagnoses in a population does not automatically answer the question of treatment in someone who already has IBD. The relevant outcome might be disease activity or a flare rather than incident diagnosis, and the evidence base may differ.

A person with an established condition should ensure that the acne prescriber knows the diagnosis, current treatment, recent disease activity, and relevant specialist advice. The gastroenterology team may also need the acne treatment plan.

This is not an automatic recommendation to exclude every patient with bowel disease. It is a reason for an individualized benefit-risk assessment. The clinician should consider acne burden, available alternatives, current stability, and the response plan if symptoms change.

The same applies to an unresolved digestive investigation. A patient should not omit it because no diagnosis has yet been confirmed. The uncertainty is relevant information, particularly if symptoms are active.

A clear shared plan should identify who assesses new symptoms, whether any additional follow-up is appropriate, and what medication instructions apply if a concern arises. The patient should not be left to negotiate contradictory advice between services.

The outcome should be a coordinated decision with a documented rationale. A broad reassuring meta-analysis or a single positive study cannot alone settle every pre-existing disease scenario because those studies may have asked a different question.

What to report when symptoms begin

A concise factual account helps the clinician assess urgency and possible causes. It should describe what changed without assuming the diagnosis. The patient does not need to provide a research-quality diary.

Useful information includes:

• The date symptoms began and whether anything similar occurred before treatment.
• The current medicine and any recent prescription changes.
• Stool frequency and the nature of the change from usual.
• Blood, black stools, fever, vomiting, or significant pain.
• Reduced fluid intake, dizziness, or other dehydration concerns.
• Recent antibiotics, travel, illness, supplements, or relevant exposures.
• Previous digestive diagnoses and ongoing investigations.

The exact list should not delay urgent care. A person with a serious presentation should seek appropriate assessment and provide what information is available.

The timing of isotretinoin matters, but so does the timing of other changes. A newly added supplement, antibiotic, or illness can complicate attribution. Reporting the whole list is more useful than selecting the exposure the patient already suspects.

Avoid minimizing language such as "it is probably just the medicine" when symptoms are significant. A medication explanation would still require an appropriate response. Equally, avoid concluding "I definitely have IBD" before evaluation. Concrete observations give the clinician a better starting point than either assumption.

Why symptom relief does not necessarily settle the diagnosis

A symptom can fluctuate or improve temporarily for several reasons. Improvement after changing a food, pausing an exposure under medical advice, or using a symptom treatment does not independently prove the original cause.

This matters when a patient feels better and cancels a recommended investigation without discussing it. The clinician may have requested follow-up to answer a question not resolved by temporary relief. The appropriate next step is to ask whether the plan should change.

The reverse is also true. Persistent symptoms after stopping a suspected exposure do not automatically prove that the exposure played no role. The clinical course and possible diagnoses need assessment rather than a simplistic on-off test.

Patients should not deliberately restart isotretinoin to see whether symptoms return. Rechallenge is a consequential medical decision, not a home experiment. The prescriber should decide whether further exposure is appropriate after evaluation.

A useful follow-up account describes the actual sequence: what instruction was given, what changed, when symptoms improved or persisted, and what other treatments occurred. The clinician can then interpret the course without relying on an incomplete causal story.

The purpose of follow-up is not merely to assign blame to a medicine. It is to establish the diagnosis where possible, treat the patient, and make a safe decision about future acne care.

Five illustrative situations

Brief diarrhea during a household illness

A hypothetical patient develops symptoms while other household members are also ill. Infection is relevant context, but the medication history and product instructions still need attention. The clinician assesses the presentation rather than assuming one explanation settles everything.

New rectal bleeding during treatment

Another patient reads a reassuring meta-analysis and considers waiting. The current label's instruction and the symptom itself take priority over reassurance from a population average. Prompt contact and appropriate assessment are needed.

Symptoms that began before isotretinoin

A third person had intermittent symptoms before the first prescription but receives a diagnosis afterward. The earlier timeline matters for both care and causal interpretation. The diagnosis date should not erase the pre-exposure history.

Established IBD with stable acne-treatment planning

A fourth patient already receives specialist bowel care. The decision concerns current disease status, treatment options, and coordination, not simply the risk of a first diagnosis in a general cohort.

A mistaken abbreviation in an article

A fifth person reads that a study examined CD and assumes Crohn's disease, although the paper concerns celiac disease. Expanding the abbreviation prevents the wrong study from shaping a treatment decision.

These examples show why one broad question about "gut problems" needs several more precise questions. The purpose of the evidence is to inform those distinctions, not force every presentation into the same narrative.

Frequently asked questions

Reading the next study without losing the practical question

A new paper may appear after a review or guideline. Evaluate what it adds: another population, a better comparator, longer follow-up, more reliable outcome definitions, or a different exposure window. A newer date alone does not make the design superior, but an older review cannot include evidence published afterward.

Compare the estimate and uncertainty, not just the direction of the headline. A small positive association with a narrow interval and a large imprecise association raise different interpretive questions. Also ask whether the endpoint is overall IBD, Crohn's disease, ulcerative colitis, IBS, or a broad gastrointestinal category.

Check whether the number of participants refers to the entire database, the matched sample, the exposed group, or the people with events. This is particularly important for rare outcomes. A very large database can still produce few exposed cases relevant to a specific comparison.

Then return to the practical question. Does the study change who should be considered for treatment, how symptoms should be assessed, or what follow-up is appropriate? A paper can be scientifically important without changing the immediate instruction for a person with rectal bleeding: obtain appropriate care rather than wait for an epidemiological debate to end.

This approach allows the evidence to develop without making the patient choose between certainty and paralysis. Research can be mixed, limitations can be real, and a clinical plan can still be clear.

A worked example of discussing an uncertain association

Consider a fictional patient who has read both the reassuring meta-analysis and the Korean case-control study before an acne consultation. They have no current digestive symptoms but have a close relative with IBD. The patient asks whether the two papers mean that prescribing is either completely safe or categorically inappropriate.

A useful answer begins by rejecting the forced choice without dismissing the concern. The papers describe associations in populations under different methods. Neither supplies an exact personal probability, and neither replaces the patient's history. The clinician should consider the acne indication, relevant history, alternatives, and the patient's preferences.

The family history belongs in that assessment, but the article should not manufacture a numerical multiplier or a mandatory testing protocol from it. The clinician can explain whether any targeted evaluation is warranted and what information would change the treatment decision.

If treatment is prescribed, the patient should leave with a clear symptom plan. They should know which symptoms require contact, which medication instructions apply, and how to access care. This is useful even when the expected risk is low because the consequences of uncertainty are easier to manage when the next step is known.

Now change the example: the same patient already has persistent diarrhea and intermittent bleeding. The decision is no longer only a discussion of future population risk. The current symptoms need assessment, and reassuring research should not be used to bypass that work. The difference is the patient's actual presentation, not which paper they prefer.

Finally, imagine symptoms begin during the course. The patient follows the product instructions and obtains medical advice. The clinician investigates without requiring proof that isotretinoin caused the problem. If a diagnosis is eventually made, future prescribing can be reconsidered with the relevant specialists. The immediate care and the causal investigation remain connected but distinct.

This example shows why a balanced article needs both research appraisal and practical clinical reasoning. A list of odds ratios without a care pathway leaves the reader uncertain what to do. A care pathway without accurate evidence can exaggerate or minimize the background concern. Combining them supports a decision that is informed, proportionate, and responsive to change.

The record should preserve the final assessment

After an investigation, the patient should understand what was diagnosed, what was not established, and what follow-up remains. If the cause is uncertain, record that uncertainty rather than replacing it with a definitive medication allergy or a definitive dismissal without supporting assessment.

The record should also state the future acne plan. A patient whose course was interrupted may need an alternative, a later review, or specialist coordination. They should not be left with untreated acne and an unexplained instruction to avoid every related medicine forever.

A concise summary can include the symptom timeline, relevant test findings, specialist opinion, medication instructions, and the next review. This helps subsequent clinicians interpret the event accurately. It also reduces the chance that a tentative suspicion becomes an unquestioned causal label as it passes between services.

Good documentation does not require certainty that medicine cannot always provide. It requires separating observations, diagnoses, and causal judgments clearly enough that the next decision can build on what is actually known.

References

1. Isotretinoin capsules prescribing information, revised August 2026. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1cf11710-f966-4529-8e08-02175f588bca
2. NIDDK. Definition and Facts for Crohn's Disease. https://www.niddk.nih.gov/health-information/digestive-diseases/crohns-disease/definition-facts
3. NIDDK. Definition and Facts of Ulcerative Colitis. https://www.niddk.nih.gov/health-information/digestive-diseases/ulcerative-colitis/definition-facts
4. NIDDK. Definition and Facts for Irritable Bowel Syndrome. https://www.niddk.nih.gov/health-information/digestive-diseases/irritable-bowel-syndrome/definition-facts
5. NIDDK. Definition and Facts for Celiac Disease. https://www.niddk.nih.gov/health-information/digestive-diseases/celiac-disease/definition-facts
6. NIDDK. Symptoms and Causes of Diarrhea. https://www.niddk.nih.gov/health-information/digestive-diseases/diarrhea/symptoms-causes
7. Isotretinoin Exposure and Risk of Inflammatory Bowel Disease: A Systematic Review with Meta-Analysis and Trial Sequential Analysis. 2023. https://pubmed.ncbi.nlm.nih.gov/37010780/
8. Association of Inflammatory Bowel Disease Incidence with Isotretinoin Usage: supplementary dataset. 2024. https://data.mendeley.com/datasets/74nvfb9wy6/1
9. Kridin K, Ludwig RJ. Isotretinoin and the risk of inflammatory bowel disease and irritable bowel syndrome: A large-scale global study. 2023. https://pubmed.ncbi.nlm.nih.gov/36529376/
10. Hyeong JH, et al. Association between Isotretinoin Exposure and Inflammatory Bowel Disease: A Nationwide Case-Control Study in South Korea. 2024. https://pubmed.ncbi.nlm.nih.gov/39296682/
11. Taylor MT, et al. A propensity score matched cohort study identifying an association of acne, but not oral antibiotic or isotretinoin use, with risk of incident inflammatory bowel disease. 2023. https://pubmed.ncbi.nlm.nih.gov/36682724/
12. FDA. Adverse event reporting dashboard: limitations of reported data. https://www.fda.gov/drugs/fdas-adverse-event-reporting-system-faers/fda-adverse-event-reporting-system-faers-public-dashboard
13. NIDDK. Diagnosis of Crohn's Disease. https://www.niddk.nih.gov/health-information/digestive-diseases/crohns-disease/diagnosis
14. NIDDK. Diagnosis of Ulcerative Colitis. https://www.niddk.nih.gov/health-information/digestive-diseases/ulcerative-colitis/diagnosis
15. Isotretinoin Exposure and Risk of Celiac Disease. PLOS One. 2015. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0135881
16. NIDDK. Diagnosis of Celiac Disease. https://www.niddk.nih.gov/health-information/digestive-diseases/celiac-disease/diagnosis