OSCE stations

Atypical Presentations of Common Diseases: 12 Wrong Doors

Mostafa Ibrahim10 min read
Atypical Presentations of Common Diseases: 12 Wrong Doors

What an atypical presentation actually is

An atypical presentation is a common disease arriving under a label that fits the symptom but not the disease. The first doctor's explanation is reasonable, and wrong. The wrong label sticks, sometimes through more than one visit, until someone re-opens the differential and recognises the underlying diagnosis that was there all along.

This is a recognition problem, not a knowledge gap. You already know myeloma. You miss it when it walks in stamped as something else, like multiple myeloma labelled as arthritis and wear and tear, bladder cancer called a water infection twice, or myasthenia gravis dismissed as being run down.

Upstream is widening your list. That lives in generating the differential list. Downstream is the moment you already have a list, somebody ticked the wrong item on it, and you need to recognise the pattern drift that says the applied label does not belong.

This is not the same as the overnight danger mindset. That piece, the overnight danger list, is about rare but lethal things you must not miss on call. This piece is about ordinary diseases wearing the wrong label for months. Different job, different muscle memory. The skill is hearing the label and asking what disease it might be hiding.

The 12 diseases and the door each one arrived through

Our misses are rarely exotic, they are familiar diseases that walk in through the wrong door. Here are twelve atypical presentations of common diseases that arrived with a tidy label already attached. Use the right-hand line to prise them back open, then test the story again.

  • Multiple myeloma The label it arrived under: Arthritis and wear and tear, twice, with physiotherapy · The line that reopens it: Back pain constant, nocturnal, at rest, not eased by position
  • Bladder cancer The label it arrived under: A water infection, two courses of antibiotics, no proven infection · The line that reopens it: Visible haematuria, painless, keeps coming back
  • Acute mesenteric ischaemia The label it arrived under: A bad stomach bug, something he ate · The line that reopens it: Excruciating pain, soft abdomen, no guarding, no rebound
  • Myxoedema coma The label it arrived under: Dementia, old age, the heating was off · The line that reopens it: Hypothermia at 31.8 degrees with no shivering, bradycardia of 42
  • Wernicke's encephalopathy The label it arrived under: Drunk, from the way he walks and talks · The line that reopens it: Does not drink, bariatric surgery four months ago, vomiting daily since
  • Myasthenia gravis The label it arrived under: Tired, stressed, run down · The line that reopens it: Weak on waking, worse by evening, normal sensation and preserved reflexes
  • Haemophagocytic lymphohistiocytosis The label it arrived under: Sepsis, an infection they have not found yet · The line that reopens it: Three antibiotic courses, negative infection screen, worsening anyway
  • Cardiac amyloidosis The label it arrived under: Heart failure, high blood pressure, old age · The line that reopens it: Thick ventricular walls on echo, low voltages on ECG
  • Haemochromatosis The label it arrived under: Alcohol, or a fatty liver, or middle age · The line that reopens it: Few units a week, brother with iron overload
  • Insulinoma The label it arrived under: Epilepsy, then panic attacks, then drunk at work · The line that reopens it: Episodes when fasting, gone within minutes of eating
  • Carcinoid syndrome The label it arrived under: IBS, rosacea, asthma, anxiety, the male menopause · The line that reopens it: Diarrhoea continues while fasting, wakes him at night
  • Wilson's disease The label it arrived under: Depression, and separately a fatty liver · The line that reopens it: Aged 24, two years abnormal liver tests, new tremor

Each middle-column label was a fair first pass. None survives the third-column line once you press it.

Flow chart to reopen a settled diagnosis: symptom fit, did prior plan help, second team problem, conflicting findings.

Pattern one: the symptom contradicts its own label

Sometimes the label explains the symptom, but the symptom carries a detail the label cannot cover. The tell is usually about timing, character, or what relieves it, and if that clashes with the label, the door you're at is the wrong one.

Gordon, 68, has had several episodes of visible red or pink blood in the urine, sometimes with clots, over weeks. No fever, no burning, no significant loin pain. He was given antibiotics twice for a presumed infection, with no benefit and no proven infection. He smokes and worked for years with industrial dyes and chemicals. For people aged 45 and over, unexplained visible haematuria without urinary tract infection, or visible haematuria that persists or recurs after successful treatment, triggers suspected cancer pathway referral per the NICE referral threshold. The feature that breaks the infection story here is that the bleeding is painless.

Alan, 68, keeps being told it is arthritis and wear and tear. No change with analgesia and physiotherapy, and he has lost about a stone in four months without trying. Mechanical pain eases with rest, unlike this. His back pain is constant, present at night and at rest, and not relieved by changing position.

Douglas, 55, collected labels for two years, IBS, rosacea, anxiety, the male menopause. None of the treatments touched his symptoms, and nobody reviewed the diary he kept of attacks and triggers. The diarrhoea is watery and secretory, it continues even when he fasts and it wakes him at night. IBS does not cause nocturnal or fasting diarrhoea, weight loss, flushing or wheeze. The contradiction is that this pattern is organic and active during fasting, and it wakes him from sleep.

Spot the wrong door clue Speak or type to an AI patient, take the history yourself, and see if you ask the question that reopens it, one free case a week, Start free

Pattern two: the label was treated, and nothing happened

A diagnosis that has been treated properly and produced no response has been tested, and it failed the test. Most people re-treat it; better to reopen with the cheap test nobody sent, like Alan's back pain where physiotherapy replaced serum protein electrophoresis and a serum free light-chain assay to confirm a paraprotein, the test that reopens it.

HLH. Douglas, 38, had three weeks of high swinging fevers with drenching night sweats, three courses of broad-spectrum antibiotics, and no improvement as he deteriorated. An exhaustive infection screen found no source, his counts fell across all three lines, his spleen enlarged, and failing to send the ferritin is the commonest reason this sepsis mimic is missed.

Insulinoma. Simon, 34, had a year and a half of episodes investigated as epilepsy, a normal EEG and MRI, and an anticonvulsant that made no difference at all. Between episodes he was completely well, which is why 'seizures' that ignore anticonvulsants and clean imaging redirect you to a metabolic cause.

Wernicke's. Craig, 42, kept being called drunk because of the way he walked and talked, despite not drinking. Four months after bariatric surgery, vomiting most days and off his prescribed supplements, the classic triad is present in only a minority, which is exactly how this gets missed.

Myasthenia. Owen, 40, is nearly normal on waking then progressively worse, slurred and nasal the longer he talks, has choked on fluids, and is breathless lying flat. That combination sits under NICE on progressive weakness, which says to refer adults with slowly progressive limb or neck weakness over weeks to months, and to refer urgently if there is any evidence of swallowing impairment.

Worked case: 68-year-old man with four months of back pain. What he was told, what he got, and six facts that never fit.

Pattern three: two problems, two clinics, one disease

Two parallel diagnoses under two different teams can hide one unifying disease you don't see. Letters go outward, they don't go sideways, so nobody owns the join.

Wilson's disease. Elliot, 24, carries depression and a fatty liver as separate labels, he's had two years of abnormal liver tests despite almost no alcohol, a tremor, slurred speech, and personality change.

Cardiac amyloidosis. Raymond, 70, has high blood pressure and heart failure, and he's assumed it was age. Both wrists had carpal tunnel surgery, his lumbar spinal stenosis is longstanding, and breathlessness and swelling worsen despite tablets he's tolerating poorly. Heart failure with preserved ejection fraction is a syndrome, not a cause.

Haemochromatosis. Martin, 52, has nine months of exhaustion and 2nd and 3rd MCP pain with about 30 minutes of morning stiffness, plus a grey bronze skin first noticed by his wife. His GP says his blood sugar is creeping up, he's lost his libido, he drinks a few units a week, and his older brother has iron overload. The trap is anchoring on alcohol or fatty liver.

Pattern four: two findings that do not agree

When two pieces of data point in opposite directions, the instinct is to trust the bit that flatters your working diagnosis. From what I can tell, the disagreement is usually the real finding, the thread you should pull.

  • Peritonism you plan to see What you actually have: Excruciating pain, abdomen soft with mild tenderness, no guarding or rebound, bowel sounds present, atrial fibrillation and anticoagulant recently stopped, mesenteric ischaemia
  • Hypothermia with shivering from cold What you actually have: 31.8 degrees without shivering, bradycardic at 42, slow shallow breaths, heating off is a precipitant not the whole explanation, myxoedema coma
  • Thick walls should give high QRS What you actually have: Echo shows thick ventricular walls with low QRS voltages on the ECG, the key discordance in cardiac amyloidosis
  • In sepsis, fibrinogen rises What you actually have: Raised triglycerides with low fibrinogen, HLH

Peritonism appears late and signals infarcted bowel, waiting for it is why the diagnosis is missed. The discordant result is the one worth chasing, a confirmatory one usually is not, and if you are choosing investigations see which tests earn their place.

Age is the commonest wrong door

Three patients had their symptoms filed under ageing. Alan, 68, was told twice it was arthritis and that he was just getting old. Bernard, 74, had months of being slower, forgetful, badly constipated, hoarse and puffy faced, and his daughter called it dementia and just getting old.

Raymond, 70, assumed it was his heart and old age. Nobody challenged that early story.

Two over 65 were given completely different labels. Gordon, 68, was told it was a water infection. Frank, 72, thought he had a bad stomach bug or something he ate.

Seven of the twelve are 24 to 55. Martin, 52, did his own filing, putting nine months of exhaustion and aching hands down to age, work stress and too little exercise, and wondering if the sugar and the aches were just middle age.

The label changes, the mechanism does not. A plausible early explanation that never got retested. That is sequence failure, not knowledge, see how doctors actually reason.

Simon, 34, has lived with this for a year and a half, unconvinced by the epilepsy label, because the tablets changed nothing and eating always helps.

What this page cannot teach you

If you're here for atypical presentations of common diseases, the best known is myocardial infarction in women, and this page cannot teach it from our cases. Every one of the twelve patients above is a man, 73 of 74 library cases are male, and the single female case is obstetric, so get that teaching elsewhere.

Take the myxoedema coma case. It is decompensated hypothyroidism, and hypothyroidism is up to ten times commoner in women than men, yet our library patient is Bernard, a 74-year-old man. The point to carry is the pattern, not the demographics that happened to come attached here.

No physical examination. No case has an examination block, so where a case mentions a soft abdomen or absent shivering you're told it and reason from it, you never elicit it. There is also no psychiatry case and no family-medicine case.

Diagnosica is an AI clinical simulator. You take a history from an AI patient by voice or text, order investigations, commit to a diagnosis and a management plan, defend your reasoning to an AI senior, and get a scorecard calibrated to the published exam mark sheets. Phone or laptop, no hardware.

All twelve here are AI patients in the library carrying the same wrong labels as in this post. For example, Alan, Gordon, Bernard and Elliot. The library is 74 cases across 14 specialties in three difficulty bands, and every case is written and signed off by a doctor.

Live and early, so expect rough edges. One free case a week. Standard is $29 a month or $249 a year, and UK cards see £22 a month or £189 a year. "It doesn't give advice about real patients, and it's not a diagnostic system. Treat every output as educational." If you want to try it now, take one of these histories.

Educational use only. Not medical advice. AI-generated; verify clinically against primary sources.

Clinical review pending.