Key Points
Why This Research Matters
How the Research Was Conducted
Key Findings: Hidden Rare Diseases in the U.K. Biobank
Multiple Sclerosis: When the Diagnosis Is Not What It Seems
Inflammatory Bowel Disease: Hidden Genetic Causes
Atopic Dermatitis: The Role of Skin-Related Rare Diseases
Noonan Syndrome: A Case Study in Underdiagnosis
Confirmation in Research and Clinical Trial Cohorts
Rare Genetic Variants and Disease Severity
Clinical Implications: What This Means for Patients
Study Limitations: What This Research Could Not Prove
Recommendations for Patients
Frequently Asked Questions
Source Information

Key Points

In a U.K. Biobank study, 1.61% of patients with multiple sclerosis, IBD, or atopic dermatitis had a hidden rare genetic disease.
Hidden rare diseases can cause standard treatments to fail and may require different medical care.
In IBD clinical trials, 4.73% of participants had a molecular diagnosis of a rare disease, higher than in the general IBD population.
Rare diseases like CADASIL, common variable immunodeficiency, and ichthyosis vulgaris can mimic common conditions.
Patients with atypical, severe, or treatment-resistant symptoms may benefit from discussing genetic testing with their doctor.

Why This Research Matters

Accurate and timely diagnosis is crucial in health care. It explains a patient's health issue and guides treatment decisions. Yet misdiagnosis has always affected the quality of medical care, and it continues to harm an unacceptable number of patients. It is estimated that 5% of U.S. adults who seek medical care each year receive a diagnostic error, sometimes with severe consequences.

The adoption of high-throughput DNA-sequencing technology has dramatically increased the rate and speed of making correct diagnoses. This is especially true in patients with cancer or rare and undiagnosed diseases, leading to notable improvements in clinical care.

Here is the core problem the researchers wanted to address. Patient recruitment for clinical trials typically depends on the clinical signs and symptoms of the disease under study. But because of misdiagnosis or underdiagnosis, a small fraction of these patients might carry a rare genetic variant capable of causing a single-gene (monogenic) disease that mimics the common disease being studied.

This matters for several reasons:

Patients affected by both common and rare diseases might have more severe or different coexisting conditions.
A therapeutic agent targeting the underlying cause of a common disease may not work in a patient whose symptoms actually stem from a different, rare disease mechanism.
This can affect clinical trial outcomes and, more importantly, patient care.

Simply put, if a patient's symptoms come from a rare genetic disease rather than the common disease they were told they had, the standard treatment for the common disease may fail — and the patient may miss out on treatments that could actually help.

How the Research Was Conducted

Researchers from AbbVie, the University of Cambridge, and other institutions designed a multi-part study. They selected four common and complex diseases — multiple sclerosis (a disease where the immune system attacks the nervous system), Crohn's disease and ulcerative colitis (the two main forms of inflammatory bowel disease, or IBD), and atopic dermatitis (a chronic inflammatory skin condition) — and tested for the co-occurrence of rare diseases that manifest with overlapping symptoms.

The Discovery Cohort: U.K. Biobank

The U.K. Biobank is a prospective, population-based cohort study involving more than 500,000 participants in the United Kingdom, recruited between the ages of 40 and 69 years from 2006 through 2010. Health outcome data were obtained through links to electronic health records, including hospital inpatient records and primary care (general practitioner) records, plus cancer and death registries.

The researchers identified disease cases using diagnosis codes from the International Classification of Diseases (ICD), which were recorded by clinical coders on the basis of patient medical notes from hospital stays. They used both the 9th and 10th revisions of the ICD to identify multiple sclerosis and inflammatory bowel disease cases. Atopic dermatitis cases were defined by a combination of ICD codes and a history of topical glucocorticoid treatment for the condition.

The final discovery cohort included 9,529 participants with sequencing data:

1,850 with a diagnosis of multiple sclerosis
6,681 with a diagnosis of inflammatory bowel disease
998 with a diagnosis of atopic dermatitis

The Follow-Up Cohorts

Because disease diagnoses in the U.K. Biobank were extracted from electronic health records, there is a risk of misdiagnosis due to administrative errors, disease misclassification, or a lack of specific ICD codes for some rare diseases. The researchers therefore treated the U.K. Biobank cohort as a discovery cohort and validated their findings in two follow-up cohorts with more rigorous phenotyping (detailed clinical characterization).

The first follow-up cohort was the Genuity Science cohort of patients with multiple sclerosis. A total of 2,025 participants in Ireland and 5,450 in the United Kingdom (7,475 total) were recruited prospectively from neurology clinics or retrospectively from existing biobanks. All were adults (18 years of age or older at diagnosis), and their diagnoses were made by trained neurologists using the revised McDonald criteria of 2013. Participants with other conditions that can cause similar neurological symptoms — such as Sjögren's syndrome, hereditary myelopathies (spinal cord diseases), brain tumors, and structural damage to the brain or spinal cord — were excluded.

The second follow-up cohort included 1,480 participants from five phase 3 clinical trials for inflammatory bowel disease treatment. The SERENE trials were two multicenter, double-blind, randomized trials evaluating adalimumab (an anti–tumor necrosis factor α [anti–TNF-α] antibody) for Crohn's disease and ulcerative colitis. Among these participants, genome sequencing was performed in 293 with Crohn's disease and 275 with ulcerative colitis, selected on the basis of variable response to treatment. Three other trials assessed upadacitinib (an oral selective Janus kinase 1 inhibitor) for ulcerative colitis and risankizumab (an interleukin-23 inhibitor) for Crohn's disease and ulcerative colitis. Among these participants, exome sequencing was performed in 287 with Crohn's disease and 625 with ulcerative colitis. All trial participants were aged 18 to 75 years and had moderate-to-severe active disease diagnosed by trained gastroenterologists.

How Genes Were Analyzed

The researchers built gene panels for each common disease using broad definitions that capture the most commonly recognized clinical symptoms. They used a gene-ranking algorithm called PhoRank implemented in VarSeq software, version 2.2.5. The algorithm ranks genes based on their relevance to specific phenotypes (observable traits) defined according to the Human Phenotype Ontology (HPO), a standardized vocabulary of phenotypic abnormalities encountered in human diseases, and Gene Ontology biomedical terminology.

For each common disease, the researchers used two HPO terms:

Multiple sclerosis: abnormality of the immune system and abnormality of the nervous system
Inflammatory bowel disease: abnormality of the immune system and gastrointestinal inflammation
Atopic dermatitis: abnormality of the immune system and abnormality of the skin

These queries initially yielded 1,973 genes for multiple sclerosis, 1,285 genes for inflammatory bowel disease, and 2,385 genes for atopic dermatitis. After manual curation with the Online Mendelian Inheritance in Man (OMIM) database — and after excluding mitochondrial genes, nondisease genes, and genes associated with multifactorial or infectious diseases or with unknown inheritance — the final panels included:

1,673 genes for multiple sclerosis
1,044 genes for inflammatory bowel disease
1,500 genes for atopic dermatitis

The researchers then queried the sequencing data for rare variants with allele frequencies of less than 1% in the entire dataset. They assigned pathogenicity (disease-causing status) to these variants in accordance with classification guidelines from the American College of Medical Genetics and Genomics (ACMG), using VarSeq software version 2.2.5.

Key Findings: Hidden Rare Diseases in the U.K. Biobank

The headline result: among the U.K. Biobank participants with a diagnosis of multiple sclerosis, inflammatory bowel disease, or atopic dermatitis, the researchers identified 153 participants (1.61%) who carried a molecular diagnosis for a mendelian (single-gene) disease. That means roughly 1 in every 62 patients with one of these common diagnoses had a hidden rare genetic condition.

The breakdown by disease:

Multiple sclerosis: 53 of 1,850 participants (2.86%) carried rare pathogenic or likely pathogenic variants
Inflammatory bowel disease: 75 of 6,681 participants (1.12%) carried such variants
Atopic dermatitis: 25 of 998 participants (2.50%) carried such variants

In total, the researchers found 52 pathogenic or likely pathogenic variants in 34 genes among the multiple sclerosis group, 54 variants in 30 genes among the IBD group, and 18 variants in 13 genes among the atopic dermatitis group. Of the multiple sclerosis-associated variants, 47 had been previously reported as pathogenic in the literature; the remaining ones were novel and classified as likely pathogenic based on ACMG criteria.

Multiple Sclerosis: When the Diagnosis Is Not What It Seems

Multiple sclerosis is typically diagnosed through clinical, radiologic (MRI), and laboratory assessments. However, several rare diseases or syndromes can resemble multiple sclerosis, and occasional misdiagnoses have long been a concern for patient care.

The most striking finding involved NOTCH3, a gene linked to a rare condition called CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy). There have been numerous documented instances of CADASIL being misdiagnosed as multiple sclerosis based on MRI findings of confluent white-matter changes. In this study, NOTCH3 harbored the highest number of pathogenic variants: 11 variants in 13 unrelated participants.

Other notable findings in the multiple sclerosis group included:

One participant carried the pathogenic BICD2 p.Thr703Met variant, associated with congenital, autosomal-dominant spinal muscular atrophy (a rare nerve and muscle disorder).
Two participants carried variants causative of autosomal-dominant familial amyotrophic lateral sclerosis (ALS): p.Gly38Arg in ANXA11 and p.Thr622Ala in MATR3.

The researchers found that certain rare diseases were underdiagnosed in participants with a multiple sclerosis diagnosis. These misdiagnoses and underdiagnoses may have occurred because of overlapping symptoms or the phenotypic variability (variable ways the disease can present) of multiple sclerosis itself. When the researchers compared the complete electronic health records of participants with rare disease–causing variants against those without such variants, they uncovered cases with severe coexisting conditions not typically expected in patients with multiple sclerosis.

Two coexisting conditions stood out dramatically:

Dementia: odds ratio of 14.79 (P = 0.002), meaning dementia was nearly 15 times more common in variant carriers
Convulsion (seizure): odds ratio of 3.54 (P = 0.002), meaning seizures were about 3.5 times more common

These findings show that in addition to misdiagnosis, rare diseases can also exacerbate common disease symptoms when they occur together (co-occur).

Inflammatory Bowel Disease: Hidden Genetic Causes

Sporadic inflammatory bowel disease is a complex condition influenced by both genetic and environmental risk factors, and it typically shows polygenic inheritance (meaning many genes each contribute a small amount of risk). However, there are also inherited forms of IBD, with more than 100 genes linked to monogenic (single-gene) disease. Many of these monogenic conditions do not respond to standard therapies and are associated with high rates of illness and death.

Among 6,681 U.K. Biobank participants with diagnosis codes for inflammatory bowel disease, the researchers identified 54 rare pathogenic or likely pathogenic variants in 30 genes in 75 participants (1.12%).

Several genes stood out as having an excess of pathogenic variants:

TNFRSF1A — a gene associated with a rare autoinflammatory condition called TRAPS (tumor necrosis factor receptor-associated periodic syndrome).
TNFRSF13B — strikingly, one third of these 75 participants (25 of 75) carried pathogenic or likely pathogenic variants in this gene, which is associated with common variable immunodeficiency (a disorder characterized by low levels of protective antibodies and increased susceptibility to infections).

An accurate diagnosis for these patients is crucial for providing effective medical care and for determining their eligibility for clinical trials, since many monogenic IBD-like conditions do not respond to standard IBD therapies.

Atopic Dermatitis: The Role of Skin-Related Rare Diseases

Among 998 U.K. Biobank participants with a diagnosis of atopic dermatitis, the researchers identified 18 pathogenic or likely pathogenic variants in 13 genes in 25 participants (2.50%).

As reported previously, FLG — the gene encoding the filaggrin protein, which is crucial for skin barrier function — was frequently mutated in people with atopic dermatitis. Loss-of-function variants in FLG cause a severe, rare skin condition called ichthyosis vulgaris (a disorder characterized by dry, scaly skin). Specifically:

8 participants carried biallelic (two-copy) loss-of-function variants in FLG
Of these, 4 were homozygous (both copies of the gene affected) and 4 had two different loss-of-function variants

Directly determining the chromosomal arrangement (phase) in these 4 participants was not possible because of the physical distance between the variants and the limitations of short-read sequencing technology. However, based on maximum-likelihood estimates, the variants are predicted to be on different chromosomes (in trans), meaning both gene copies were affected.

Additionally, in 2 participants, the researchers identified two ultrarare, pathogenic PTPN11 variants (p.Gly60Ala and p.Gln256Arg), previously associated with Noonan's syndrome, a rare genetic disorder that can cause dry skin and other features.

Noonan Syndrome: A Case Study in Underdiagnosis

At least 12 genes have been implicated in Noonan's syndrome. To understand how often this rare condition goes undiagnosed in the general population, the researchers screened for these 12 genes among all U.K. Biobank participants and conducted a comprehensive analysis of coexisting conditions using electronic health records from carriers of pathogenic or likely pathogenic variants.

The results were striking:

The researchers identified 164 people who were probably affected by Noonan's syndrome and clinically related disorders.
However, only 5 of 73 participants (about 7%) for whom general practitioner records were available had a formal diagnosis of Noonan's syndrome.
Most of the participants had instead received a diagnosis of either related symptoms or different diseases entirely.

It is important to note that while Hospital Episode Statistics records are accessible for the entire U.K. Biobank cohort, general practitioner records are currently available for only 45% of the cohort. For the remaining 91 participants without a Noonan's syndrome diagnosis in hospital records, it remains unclear whether a diagnosis exists in their general practitioner records.

Confirmation in Research and Clinical Trial Cohorts

The researchers then tested whether their discovery-cohort findings held up in the more rigorously phenotyped follow-up cohorts.

Genuity Science Multiple Sclerosis Cohort

In the Genuity Science research cohort of 7,475 participants with multiple sclerosis, 88 participants (1.18%) had a potential molecular diagnosis for a rare disease. Although the Genuity Science cohort had four times as many participants as the U.K. Biobank multiple sclerosis group (7,475 vs. 1,850), the proportion carrying pathogenic variants was less than half (1.18% vs. 2.86%).

This difference likely reflects how diagnoses were made. The U.K. Biobank relied on electronic health record–based diagnosis codes, while the Genuity Science cohort underwent targeted phenotyping by trained neurologists, and participants with suspected rare diseases were mostly excluded from recruitment. This comparison highlights how often rare diseases hide behind common disease labels in real-world health records.

Clinical Trial Inflammatory Bowel Disease Cohort

Among 1,480 participants with sequencing data in the inflammatory bowel disease clinical trials, the researchers identified 70 (4.73%) who had a molecular diagnosis of a rare disease. This was more than four times higher than the rate in the U.K. Biobank IBD group (1.12%).

Why the difference? The clinical trial cohort specifically recruited patients with moderate-to-severe inflammatory bowel disease. In these patients, the disease may have developed due to highly penetrant (strongly expressed) rare variants, making rare genetic causes more common among the most severely affected patients.

Rare Genetic Variants and Disease Severity

The researchers also asked a crucial question: do hidden rare diseases affect how severely a common disease progresses?

To answer this, they calculated the Age-Related Multiple Sclerosis Severity (ARMSS) score in 31 carriers of pathogenic variants and 3,158 noncarriers with available data. The ARMSS score is a validated measure of disability progression in multiple sclerosis, adjusted for age.

The results: participants carrying pathogenic or likely pathogenic variants had higher ARMSS scores (indicating more severe disease) than those without such variants, with a P value of 0.02. This is a statistically significant difference, meaning there is only about a 2% probability that this finding was due to random chance.

In plain terms: people whose "multiple sclerosis" was actually caused by, or complicated by, a rare genetic disease tended to have more severe disability over time.

Clinical Implications: What This Means for Patients

This study has major implications for how common diseases are diagnosed and treated. Here is what the findings mean in practical terms:

Misdiagnosis is measurable and not rare. The 1.61% rate of hidden rare diseases in the U.K. Biobank, and the 4.73% rate in IBD clinical trials, means that among large groups of patients with common diagnoses, a meaningful minority actually have a different underlying disease.
Treatment failure may have a genetic explanation. A therapeutic agent targeting the underlying cause of a common disease may not work in a patient whose symptoms come from a distinct rare disease mechanism. This could explain why some patients in clinical trials do not respond to otherwise effective drugs.
Clinical trial results may be skewed. If patient recruitment for trials depends on clinical signs and symptoms, patients with hidden rare diseases may dilute or distort the measured treatment effect, potentially causing effective drugs to appear less effective than they truly are.
Disease severity matters. The finding that MS patients with rare variants had more severe disease (higher ARMSS scores) suggests that hidden genetic conditions can worsen outcomes in patients who carry them.
Rare diseases can masquerade as common ones. CADASIL misdiagnosed as multiple sclerosis, common variable immunodeficiency misdiagnosed as IBD, and ichthyosis vulgaris or Noonan's syndrome misdiagnosed as atopic dermatitis are all documented in this single study.
Underdiagnosis is widespread. The Noonan's syndrome analysis — where only 5 of 73 probable cases had a formal diagnosis — shows that rare diseases frequently go completely unrecognized in the health system.

The researchers concluded that their study demonstrates "the value of systematic genome sequencing in understanding the phenotypic heterogeneity of common diseases and identifying failure to diagnose rare diseases." They also highlighted "the benefits of deep molecular phenotyping in clinical trials and patient care."

Study Limitations: What This Research Could Not Prove

Every study has limitations, and this one is no exception. Understanding these caveats is important for interpreting the results appropriately.

Electronic health record limitations. Because disease diagnoses in the U.K. Biobank were extracted from electronic health records, there is a risk of misdiagnosis due to administrative errors, disease misclassification, or a lack of specific ICD codes for some rare diseases. This could inflate or deflate the true rates of hidden rare diseases.
Incomplete penetrance and variable expressivity. Rare disease–associated variants discovered in clinical studies (where participants are specifically recruited for gene discovery) often show incomplete penetrance (meaning not everyone with the variant develops the disease) or variable expressivity (meaning the same variant can cause different symptoms in different people) in the general population. Carrying a pathogenic variant does not guarantee that the rare disease is actually causing the patient's symptoms.
Incomplete health records. General practitioner records were available for only 45% of the U.K. Biobank cohort. For the 91 participants without a Noonan's syndrome diagnosis in hospital records, it is unknown whether a diagnosis exists in their primary care records.
Technological limitations. Directly determining the chromosomal phase of FLG variants was not possible in 4 participants due to the physical distance between variants and the limitations of short-read sequencing technology, though maximum-likelihood estimates supported trans configuration (both copies affected).
Selection bias in follow-up cohorts. The Genuity Science cohort excluded participants with suspected rare diseases, which likely explains the lower rate of hidden rare diseases (1.18% vs. 2.86% in the U.K. Biobank). The clinical trial cohort, by design, recruited only patients with moderate-to-severe disease, which likely explains the higher rate (4.73%).
Statistical context. The ARMSS score analysis included only 31 carriers and 3,158 noncarriers — a relatively small carrier group. The odds ratio findings for dementia (14.79) and convulsion (3.54), while highly significant (P = 0.002), were based on small numbers of events.

Recommendations for Patients

This research does not mean that everyone with multiple sclerosis, IBD, or atopic dermatitis needs immediate genetic testing. But it does suggest some practical steps that may help patients receive the most accurate diagnosis and best possible care.

Ask about genetic testing if your disease is atypical. If your symptoms do not quite fit the typical pattern of your diagnosed condition, if your disease is unusually severe, or if you have unusual coexisting conditions, ask your doctor whether genetic testing might be appropriate.
Discuss treatment failure openly. If a standard treatment that works for most patients has not helped you, mention this to your doctor. The study shows that some non-responders may actually have a different underlying disease mechanism.
Consider whether you have a family history of your condition. Monogenic (single-gene) diseases often run in families. A strong family history of a "common" disease — or of related symptoms in multiple relatives — may be a clue worth investigating.
Be aware of overlapping symptoms. Conditions like CADASIL (which can mimic multiple sclerosis), common variable immunodeficiency (which can mimic IBD), and ichthyosis vulgaris (which can mimic atopic dermatitis) have specific treatments that differ from those for the common disease they resemble.
Participate in research with sequencing when possible. The researchers note that deep molecular phenotyping — combining genetic and RNA (transcriptome) information — can reveal why some patients do not respond to therapy. Clinical trials that include genetic testing may improve both research and individual patient care.
Know that a misdiagnosis is not the end of the road. A correct molecular diagnosis can open the door to appropriate treatments, clinical trial eligibility, and a clearer understanding of your condition.

It is worth emphasizing that this study does not suggest that standard diagnoses are wrong for the vast majority of patients. Rather, it shows that for a small but clinically meaningful minority, a deeper look at the genetic roots of their symptoms can change the course of care.

The authors' closing message is that systematic genome sequencing has real value in understanding why common diseases look so different from patient to patient (phenotypic heterogeneity) and in catching rare diseases that were never properly diagnosed. For patients enrolled in clinical trials — and for patients in everyday care — this kind of molecular detective work can make the difference between failed treatment and effective treatment.

Frequently Asked Questions

Why might a standard treatment for my common disease not work?

If your symptoms are actually caused by a rare genetic disease that mimics a common condition, the standard treatment for the common disease may not work. The study found that some patients with hidden rare diseases did not respond to standard therapies, and they may need different treatments targeting the rare disease mechanism.

Should I get genetic testing if I have multiple sclerosis, IBD, or atopic dermatitis?

The study does not recommend genetic testing for everyone with these conditions. However, if your symptoms are atypical, unusually severe, or you have unusual coexisting conditions, you may want to discuss with your doctor whether genetic testing could be helpful. A correct molecular diagnosis can guide more effective treatment.

What does it mean if I have a pathogenic variant but no symptoms?

Carrying a pathogenic variant does not guarantee that the rare disease is actually causing your symptoms. Rare disease variants can show incomplete penetrance, meaning not everyone with the variant develops the disease, or variable expressivity, meaning symptoms can vary. Your doctor can help interpret your specific situation.

How can hidden rare diseases affect clinical trial results?

If clinical trials recruit patients based on symptoms, some may actually have a different rare disease. This can dilute or distort the measured treatment effect, potentially making effective drugs appear less effective. The study found higher rates of hidden rare diseases in IBD clinical trial participants (4.73%) compared to the general IBD population.

If I have a rare disease that mimics my common diagnosis, what should I do?

If you suspect a hidden rare disease, discuss with your doctor. A correct molecular diagnosis can open the door to appropriate treatments, clinical trial eligibility, and a clearer understanding of your condition. The study emphasizes that systematic genome sequencing can help identify these hidden conditions and improve patient care.

When should a patient with multiple sclerosis, inflammatory bowel disease, or atopic dermatitis seek a second opinion about a possible hidden rare genetic disease?

Consider a second opinion if your symptoms are atypical, unusually severe, or if standard treatments fail. In a large study, 1.61% of patients with these common diagnoses actually had a hidden rare single-gene disease, and rates were higher in severe IBD trial patients (4.73%). Hidden conditions like CADASIL, common variable immunodeficiency, or ichthyosis vulgaris can mimic common diseases and require different care. A second opinion can help determine if genetic testing is appropriate. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Common Diseases in Clinical Cohorts — Not Always What They Seem

Authors: Fedik Rahimov, Ph.D., Benjamin M. Jacobs, B.M., B.Ch., Ph.D., John S. Lee, M.S., Naim A. Mahi, Ph.D., Andrew Blumenfeld, M.S., Ammar J. Alsheikh, M.D., Ph.D., Ali Abbasi, M.D., Ph.D., Mark Reppell, Ph.D., Valerie L. Pivorunas, Ph.D., Haukur J. Sigurðsson, B.Sc., Stephen Sawcer, Ph.D., Heath Guay, Ph.D., Jeffrey F. Waring, Ph.D., Howard J. Jacob, Ph.D., and Nizar Smaoui, M.D.

Funding: The study was funded by AbbVie and the NIHR Cambridge Biomedical Research Centre. AbbVie participated in the design and conduct of the study, the interpretation of data, and the review and approval of the manuscript.

Note: This patient-friendly article is based on peer-reviewed research published in the New England Journal of Medicine. It is intended for educational purposes and does not constitute medical advice. Patients with questions about their specific diagnosis or treatment should consult their health care provider.

Hidden Rare Diseases Behind Common Diagnoses: What a Landmark Study Reveals for Patients

Table of Contents