DIAGNOSIS — it’s all that really matters.  We can’t a treat a patient if we don’t know what they have.  And aside from antibiotics, surgery, maybe chemotherapy, much of our therapeutics achieves comparatively little.  By all means, do what you can to relieve asthma symptoms, control hypertension, ease gastroesophageal reflux, minimize pain, etc.  But please — always think to yourself, “Hey, Dude, what’s the DIAGNOSIS?”

A 44-year-old woman, volunteer at my colleague’s free clinic, mentioned her bad headache requiring lots of ibuprofen.  My colleague urged she see her primary, who prescribed an antibiotic for sinusitis.  A week later, symptoms unimproved, they changed the antibiotic.  The headache persisted another week, they performed a head CT and lumbar puncture, both were normal.  Into the fourth week, her cerebral aneurysm ruptured completely, and the lady died.

What went wrong?

When aneurysms rupture, causing a Subarachnoid Hemorrhage (SAH), over half the time patients succumb then-and-there.  But up to 50% of patients experience a “warning leak” [a.k.a. “sentinel leak”], which seals over.  This is obviously unstable, and will fully burst a few weeks later [rarely, in just 1-2 days], but offers a window period for life-saving surgery.  It’s a diagnosis NOT TO MISS.

A non-contrast head CT is virtually 100% sensitive within the first 12 hours, over 90% within the first 24 hours [i.e. no false-negative scans].  But this sensitivity decreases to 60% by the first week, as blood is resorbed.  Similarly, a lumbar puncture (LP) is highly sensitive throughout the first week, much less so after two.  So if you entertain the diagnosis of SAH at all, don’t delay.

Who warrants this work-up (head CT and LP) for an acute headache?  Anyone with both oftwo clinical parameters:

  1. New onset of the “worst headache ever.”  “Worst” is defined by a patient’s prior headaches and pains; the intensity depends on how much blood leaks, but the headache is always the “worst ever.”
  2. Sudden Onset: Pain often reaches its maximal intensity immediately.  Makes sense — a drop of blood spurts out, the pain begins then & there.

If the sudden onset of headache occurred during a valsalva, it should raise suspicion.  This includes coughing, lifting, sex, etc.  Sadly, some women die of SAH while pushing during labor.  Patients with polycystic kidney disease are at high risk of cerebral aneurysms.  A family history of SAH [often reported as a “stroke,” but at an age too young for atherosclerosis] may suggest rare genetic syndromes.

Certainly, anybody with a new headache deserves a complete neurologic exam, especially of the cranial nerves.  But in most patients with a sentinel leak, this will be normal.  The most common clinical error is to assume that patients with SAH look ill — true for the fully ruptured aneurysm, but not for the treatable “sentinel leak.”  After all, the patient described above went to work with her “worst headache ever,” to an unpaid volunteer job at that!

MORAL: a patient with the very sudden onset of the “worst headache ever,” even if not appearing in much pain, should get a non-contrast CT & an LP, most expeditiously obtained via an emergency department.  If 2 weeks have elapsed since symptom onset, consider a neurologist opinion regarding an angiogram.

MORAL #2: When a patient’s condition isn’t improving, NEVER change an antibiotic without a very specific reason.  Instead, question your diagnosis.

This Blog encapsulates over 30 years of experience, anecdotes, pearls, pitfalls, clinical tricks, PubMed searches, & literature reviews.  And, occasionally, the very best teacher of all — really bad outcomes [hopefully those of colleagues & others, not one’s own].  It deals primarily with adults and school-age children in an out-patient setting.  We’ll explore a comprehensive range of Differential Diagnosis, one “chief-complaint” at a time.

The key to Diagnosis is History-Taking, much more so than physical exam or other testing.  A few crucial suggestions:

  • Put yourself in a patient’s shoes — truly imagine what’s been going on.  Pardon the cliché, but try to “feel their pain” (or vomit, diarrhea, whatever).
  • CHRONOLOGY of illness is the key parameter — get it right.  Are symptoms new or recurrent; intermittent or continuous?  I’ve seen countless clinicians fail here.
  • Don’t ask a single question without knowing what you’re trying to rule out or in — obtain data systematically.
  • Consider the natural history of each diagnostic entity — is the patient’s presentation a possible course of untreated illness?  [or, if they’d tried a therapy, of treated disease].
  • Always inquire if the patient’s illness is “getting better,” “getting worse,” or “staying the same”.  A progressive course is always more concerning.

Acute symptoms (i.e. recent onset) are the most dangerous.  This is counter-intuitive to many of my patients, who think that ten years of pain is worse than a few days’ worth.  Keep in mind conditions which can do a person in rapidly [the “DON’T MISS” category].

We’ll begin with “ACUTE HEADACHE” — a very common chief complaint, rarely serious, though occasionally associated with preventable deaths.  For this symptom, let’s define “acute” as less than 3-4 weeks duration.

The following table outlines the differential diagnosis of the “acute headache”.  Always consider conditions “Not To Miss,” which can invariably be rapidly ruled out.  Most of them present within a few days of onset.  Then, target your questions and exam to ruling-in a less-serious illness.  Even if you’re wrong, those conditions are mostly self-limiting.




Next posting we’ll cover increased Space Occupying Lesions, i.e. increased intracranial pressure, and the essence of a succinct neuro exam for the headache patient (Acute H/A – 2), then go on to rare causes NOT TO MISS (Acute H/A – 3) and more common causes (Acute H/A – 4), followed by Recurrent Headaches, and finally Chronic Headaches.


  1. Almost 3 years into my practice I still have your headache handout posted about my desk reminding me of what not to miss! and I owe my neuro exam to you. Thanks for being a great teacher! I would love to keep up with your blog! There are still many moments when I’ve thought to myself in clinic with a difficult case: “What would Steve do…” and I’ve even (already) caught myself saying to new grad clinicians struggling what med they want to give before getting to their diagnosis: “We’re here to rule OUT, not rule in” 🙂 Keep it up, Steve!

  2. As an urgent care provider, headaches are my bread and butter. Rarely do I complete a shift without seeing one or more patients with a headache. Many of those patient’s will tell me that they are experiencing their worst headache ever. If you are in the midst of a severe migraine, even if you had dozens of them before, it is likely to seem like your worst ever. From the patient’s perspective, if you gone into urgent care with a headache, waited two + hours to be seen, you are going to play up its severity in order to justify your being their in the first place. However, most, easily over 99%, of the headaches that i see are benign. The radiation exposure from a head CT is not insignificant. They are also costly. LPs are invasive and not without their own risks. Added to this is the fact that a high percentage of the patient’s I see who tell me they are experiencing their worst headache ever are drug seeking. So in this environment, the decision to order the head CT becomes increasingly difficult. I find myself ordering fewer of them than even a few years ago. I ask myself is this a mistake? Am i flirting with disaster? Or, is my hesitation to order them frequently good medicine? Not an easy question to answer. For me the biggest red flag is the patient who tells me they never have headaches and yet is experience a headache severe enough to bring them into urgent care. Nevertheless, despite ordering hundreds of CTs over the years, I don’t believe I have ever come up with a positive SAH. As for antibiotics, I couldn’t agree more. Acute bacterial sinusitis presenting solely as a headache is probably even more rare than the patient with SAH.

    • Thanx lots for the thoughts. As you note, headaches are common, but in my experience it’s rare for a “worst-headache-ever” to reach maximum intensity immediately at onset. Migraines build up over 30-60 minutes or more; I worry when it’s 30-60 seconds. When a patient says the headache “began in the morning,” I inquire exactly what time? where were they? what were they doing? If such precision makes no sense to them, I leave “sentinel leak” & go on to other diagnoses. My last suspicious history was 10 years ago — a young woman whose H/A began when second in line from the McDonald’s cashier, not 3rd or 1st (work-up was fortunately negative). You’re perfectly right about CTs.

  3. Wow. I get lots of clinical problem-solving articles via email & snail mail, but this was the best I’ve ever read. It was clear, it was engrossing (really? a 44-yr-old vol died of a headache??), and best of all, it taught me something i should know, but don’t. Keep blogging, DiagnosisDude!

  4. i have been in practice for 20 years and this is still interesting and helpful. please keep it up!!

  5. I really like your writing style, wonderful info , appreciate it for posting : D.

  6. Its such as you learn my thoughts! You seem to grasp a lot about this, like you wrote the e-book in it or something. I feel that you simply could do with some percent to power the message home a little bit, but instead of that, that is excellent blog. An excellent read. I will certainly be back.

  7. To be sure with you 100%. I would hesitation, for one thing, on the other hand think that not really worthy of bringing up.

  8. You lectured in my nursing class today. It was great! Your blog looks like it will be the same way. Thanks for contributing to the field!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s