Workup for Old People

[ACal]

Over a year out of residency, working a lot of evening/night shifts, still uncomfortable with old patients that come in at night...

2 scenarios:


1. 85 yo F with dementia, a million other medical problems, unwitnessed GLF fall at SNF, A&O x1 maybe x2 at baseline, doesn't remember what happened. Maybe feels a little weak recently. Probably on aspirin/plavix/Eliquis (Full Code)
I usually scan their head/neck and film whatever hurts. Basic labs/trop/ekg/urine. If I don't find anything, usually some mild leukocytosis or baseline anemia, can't really admit them for anything, maybe high risk syncope? But usually back to the SNF...I don't feel too worried about these. (Maybe I should be? Anything important not to miss?)


2. 70 yo F with HTN, DM, HLD, comes in feeling weak and "dizzy" for a day. Maybe some vertigo but not really. (Can never tell me if they feel the room spinning) Labs/cardiac workup are normal....sometimes I'll get a CT head noncon (usually normal and useless). No focal neuro deficits. They can still walk but maybe a lot slower. Nothing to admit them for. Not really even a SNF candidate. What do you do with these?

---

Members don't see this ad.

 

[WildEMDoc]

1. Unwitnessed falls in debilitated people I generally only do whatever traumatic workup is indicated as long as there is no other concern (not been eating/fevers). Very rarely do I do labs, but if I do some screening ones its only a CBC/BMP never a trop. Trop too often indeterminant and meaningless. Always sending them back unless something egregious wrong (SNFs can take care of a lot of acute on chronic disease as long as they haven't had any big changes).

2. Labs, +/- CT head depending on history. Tele in the ED for ~2 hours. Discharge if they are at their functional baseline and labs etc are normal.

 

[lymphocyte]

Over a year out of residency, working a lot of evening/night shifts, still uncomfortable with old patients that come in at night...

2 scenarios:


1. 85 yo F with dementia, a million other medical problems, unwitnessed GLF fall at SNF, A&O x1 maybe x2 at baseline, doesn't remember what happened. Maybe feels a little weak recently. Probably on aspirin/plavix/Eliquis (Full Code)
I usually scan their head/neck and film whatever hurts. Basic labs/trop/ekg/urine. If I don't find anything, usually some mild leukocytosis or baseline anemia, can't really admit them for anything, maybe high risk syncope? But usually back to the SNF...I don't feel too worried about these. (Maybe I should be? Anything important not to miss?)


2. 70 yo F with HTN, DM, HLD, comes in feeling weak and "dizzy" for a day. Maybe some vertigo but not really. (Can never tell me if they feel the room spinning) Labs/cardiac workup are normal....sometimes I'll get a CT head noncon (usually normal and useless). No focal neuro deficits. They can still walk but maybe a lot slower. Nothing to admit them for. Not really even a SNF candidate. What do you do with these?

Less is more. I wouldn't get any bloods in the first case, especially not a troponin or even a urine with no urinary symptoms.

Second person is not "old" in the same way the first one is. They're independent from home. They may have neurological symptoms. They're walking different, which is a big functional change. You don't know what's going on. I would admit them to short stay until they're at least feeling better.

Also, using symptoms to decipher vertigo is unreliable and imprecise. One study of over a thousand people found more than 50% changed their symptom type within 10 minutes of being asked. It's confusing to us because it's confusing to the patient.

I ask what's triggered it and how long it's been going for. If it's acute and ongoing, hasn't really changed, there's cerebrovascular risk factors, and there isn't something obvious like an arrhythmia or hypokalemia, they get admitted for either what we call "dizz phys" (vertigo physiotherapy) in short stay or a neurology consult depending on my HiNTs exam.

There's now reasonable evidence that trained EPs (and not just neuro-ophthalmologists) can outperform early MRI when it comes to the HiNTs exam. Neuroimaging is not always helpful, and even early MRIs miss posterior circulation strokes about 15% of the time.

Check out this iconoclastic paper and Peter Johns on YouTube.

https://onlinelibrary.wiley.com/doi/full/10.1111/acem.12234

 

[Groove]

Less is more. I wouldn't get any bloods in the first case, especially not a troponin or even a urine with no urinary symptoms.

Second person is not "old" in the same way the first one is. They're independent from home. They may have neurological symptoms. They're walking different, which is a big functional change. You don't know what's going on. I would admit them to short stay until they're at least feeling better.

Also, using symptoms to decipher vertigo is unreliable and imprecise. One study of over a thousand people found more than 50% changed their symptom type within 10 minutes of being asked. It's confusing to us because it's confusing to them.

I ask what's triggered it and how long it's been going for. If it's acute and ongoing, hasn't really changed, there's cerebrovascular risk factors, and there isn't something obvious like an arrhythmia or hypokalemia, they get admitted for either what we call "dizz phys" (vertigo physiotherapy) in short stay or a neurology consult depending on my HiNTs exam.

There's now reasonable evidence that trained EPs (and not just neuro-ophthalmologists) can outperform early MRI when it comes to the HiNTs exam. Neuroimaging is not always helpful, and even early MRIs miss posterior circulation strokes about 15% of the time.

Check out this iconoclastic paper and Peter Johns on YouTube.

https://onlinelibrary.wiley.com/doi/full/10.1111/acem.12234

That's from Dr. Jonathan Edlow. He's got some of the most exquisite articles on dizziness I've ever read. Check out a couple of his articles from 2018 in Annals (not sure where to link a free version...)

Diagnosing Patients With Acute-Onset Persistent Dizziness

Managing Patients With Acute Episodic Dizziness

I remember when it first came out in Annals I had to read it twice to fully digest. I was so impressed that I sent him an email praising his work and thanking him for writing it which seemed to really make his day. He's got such a great system.

 

[ACal]

I agree the lab workup is usually low yield for the first case.
But if it was an unwitnessed fall it could be syncopal and doesn't that at least warrant a basic workup? I'm a bit paranoid since I find significant pathology every once in a while in these patients (had a bilateral PE, significant AKI, heart block, NSTEMI). Honestly I feel like they shouldn't be treated aggressively but I do the labs mostly as defensive measure to justify "I met a standard of care". Thoughts?

 

[WildEMDoc]

I agree the lab workup is usually low yield for the first case.
But if it was an unwitnessed fall it could be syncopal and doesn't that at least warrant a basic workup? I'm a bit paranoid since I find significant pathology every once in a while in these patients (had a bilateral PE, significant AKI, heart block, NSTEMI). Honestly I feel like they shouldn't be treated aggressively but I do the labs mostly as defensive measure to justify "I met a standard of care". Thoughts?

Could always have been syncope. But thinking everyone is a syncope is what will slow their dispo and lead to unnecessary testing.

Not a one size fits all for this one. Based on the history/exam is where I decide if I need to do anything besides traumatic injuries. Things that shift me one way or another are things like, do they fall a lot, do they normally use a walker and nobody saw one nearby, where were they found, are they totally at baseline now, did they pee/stool themselves, where are their goals of care and functional status at baseline (mildly demented vs aaox0/completely out of it at baseline

Nobody will fault you for working these people up though, besides maybe a disgruntled hospitalist that is going to obs them for their abnormal labs or whatever. A lot of my more conservative colleagues work them all up and hand them over to me not infrequently and I’m more than happy to still get their patient on obs for whatever comes out of the shotgun labs. That’s just the way some folks practice and at the end of the day you have to decide where you fall.

 

[lymphocyte]

I agree the lab workup is usually low yield for the first case.
But if it was an unwitnessed fall it could be syncopal and doesn't that at least warrant a basic workup? I'm a bit paranoid since I find significant pathology every once in a while in these patients (had a bilateral PE, significant AKI, heart block, NSTEMI). Honestly I feel like they shouldn't be treated aggressively but I do the labs mostly as defensive measure to justify "I met a standard of care". Thoughts?

Trust your gestalt when it comes to syncope. Your clinical judgement is as good or better than the best validated decision tools at predicting 30 day outcomes. If you think they need the work-up, do the work up.

For me, normal signs and back to normal in an otherwise demented 85 year old is good enough. I always try to call NOK, explain I don't exactly know what happened, grandma looks okay with normal scans, is back to normal, but I want them to see their usual doctor in 48 hours to make sure things are going okay. Most NOK are actually pretty reasonable. Safety net and discharge instructions are key.

It could always be X, Y, and Z. All of medicine is probability and managing risk in a rational way.

International Validation of the Canadian Syncope Risk Score - American College of Cardiology

Mollie McDermott, MD, MS

www.acc.org

 

[JacobMcCandles]

There should be a reason you are getting a particular lab test. Getting random troponins isn't good medicine and isn't good for the patient. You'll clog your department with people who are either getting admitted for a minimally elevated troponin or waiting on their repeat troponin.

You should really think about what a hospital stay is going to accomplish for the patient. People with dementia and a million other medical problems typically aren't going to benefit from a hospital stay for non-specific complaints vs. going back to their facility.

 

[Zebra Hunter]

I almost never get labs on the first patient given the downstream effects of incidental findings in the demented nursing home patient. Maybe a head CT/cervical spine CT at most is what I’ll order. I don’t typically bother with an ECG either. If they are at their neuro baseline, have normal vitals, and otherwise appear pleasantly demented what is the patient centered benefit of finding the sodium of 124 or the potassium of 2.9, or the questionable UTI you found? Hell even if they are having a STEMI, none of my cardiologist are taking patients so demented that they can’t remember falling 30 minutes ago to the cath lab.

 

[Rekt]

1: Agree with most people above. Labs and even ECG are a giant waste. Radiographs/CT - discharge back to lala land.
2: CT scans are a waste is dizzy patients. Either its vertigo or not. Decide to MRI them or do nothing. I agree trops are worthless, but most people order them. Some "expert" would love to throw you under the bus for not ordering a trop on dizziness.

 

[RustedFox]

1: Agree with most people above. Labs and even ECG are a giant waste. Radiographs/CT - discharge back to lala land.
2: CT scans are a waste is dizzy patients. Either its vertigo or not. Decide to MRI them or do nothing. I agree trops are worthless, but most people order them. Some "expert" would love to throw you under the bus for not ordering a trop on dizziness.

You must work somewhere where you can order MRIs with relative ease.

 

[Rekt]

You must work somewhere where you can order MRIs with relative ease.

When I trained it was still a big deal and I rarely got them. I still practice that way. I can't really tell where I'm at now yet as far as culture but it seems less taboo. I'll still admit for MRI too.

 

[RoyBasch]

Over a year out of residency, working a lot of evening/night shifts, still uncomfortable with old patients that come in at night...

2 scenarios:


1. 85 yo F with dementia, a million other medical problems, unwitnessed GLF fall at SNF, A&O x1 maybe x2 at baseline, doesn't remember what happened. Maybe feels a little weak recently. Probably on aspirin/plavix/Eliquis (Full Code)
I usually scan their head/neck and film whatever hurts. Basic labs/trop/ekg/urine. If I don't find anything, usually some mild leukocytosis or baseline anemia, can't really admit them for anything, maybe high risk syncope? But usually back to the SNF...I don't feel too worried about these. (Maybe I should be? Anything important not to miss?)


2. 70 yo F with HTN, DM, HLD, comes in feeling weak and "dizzy" for a day. Maybe some vertigo but not really. (Can never tell me if they feel the room spinning) Labs/cardiac workup are normal....sometimes I'll get a CT head noncon (usually normal and useless). No focal neuro deficits. They can still walk but maybe a lot slower. Nothing to admit them for. Not really even a SNF candidate. What do you do with these?

Patients being in a SNF helps with their disposition, because they can go back. The ones that are hard are the ones coming from home. Often times they have poor social support (the real reason they are in the ER at all). They are chronically progressing in the dementia/debility/dwindling and cannot easily attend to their ADLs at home. In this case you can consider admit for placement vs. discharge. This often depends on the standard of care/practice pattern of your given hospital/region.

I think one thing that can help with these permanently "grey area" decision making is shared decision making with the patient (or their family often over the phone, if patient is demented). I do take into account (especially when talking to family/DPOA) code status, baseline functionality, etc. I have a spiel I give these patients where I explain dizziness/syncope can potentially have a cardiac etiology that is difficult to exclude. If they want to be admitted for tele-obs and further workup like an echo, so be it. Alternatively if they want to go home and follow up with a cardiologist as an outpatient for a TTE and event monitor they can. Frequently these patients WANT to go home.

Particularly as the patient's baseline status deteriorates, the patient's families often want less rather than more done. That being said, there are families that wanted everything done on the 95 year old A&Ox0 end stage dementia ("He's a fighter, I swear!") +/- PEG tube patient. If so, don't take it personally. Is that ultimately futile? Sure. But it's no sweat off my sack, just get 'em admitted and move on.

That being said, these cases are/CAN be HARD. Elderly comorbid patients are at risk for devastating illnesses. They often have vague and non specific presentation without focal signs more common to healthier middle age patients with the same pathology. The elderly weak and dizzy can be nothing (most likely) but they also can be occultly septic, have a vascular catastrophe in their chest or abdomen, occult trauma from unmentioned falls, etc. So you have to tread lightly.

 

[Groove]

A lot of good advice in here. I'm inclined to agree with @lymphocyte in that I primarily operate with a clinical gestalt approach. Either I do the minimal obligatory imaging while skipping labs if I can justify it or I do a full work up based on pretest probability for an occult process based entirely on my initial clinical impression. Another way to save some time is to reach out to POA/family right off the bat and see if they can shed any additional history that would influence your work up. I also use this to gauge any resources that might be lacking at the SNF/home because if there is any reason to put them in for PT/OT/CM/SW assessments then dispo becomes easy. I also feel out the family to see if there is any resistance to discharge. If there is...then I cast a bigger net to help justify the admission.

Big exceptions would be a messy shift where you get dumped these 2 patients from your colleague during sign out (who doesn't know what to do with them) and immediately are slammed with a few strokes and/or trauma alerts. In that case, just shotgun and pan scan while putting out the other fires and circle back to see if you caught anything and disposition based on those results. It's inefficient and messy and weak but the quickest route to disposition during extenuating circumstances and you won't lose sleep or be tempted to kick them out the door too early while you put out your other fires.

If you are ever really having decision paralysis, especially early in your career don't be afraid to "phone a friend" and call up your friendly neighborhood consultant whether that be neuro or cards. Many times it lets you bounce the case off someone with a different perspective + will provide a way to guarantee they can be seen in 2-3 days. Or you can get buy in from them for further inpatient work up to buff an admission which will supersede any push back from medicine. You can often sell these either way to the consultant depending on how you present the case.

The last two approaches are kind of weak but won't kill your flow on a busy night and I have been guilty of doing them especially earlier in my career. Nowadays not so much. The more patients you see throughout your career, the easier these get and you often times can really just kind of .... trust your gut.

Agree with @RoyBasch these kinds of pts can be nothing or everything and you have to be careful. Don't rush them but on the other hand being thorough does not always = massive work up either.

 

[thegenius]

At best....if there is some reason to be concerned for syncope in these unwitnessed falls...I'll do the following:

1) put them on the monitor. That isn't recorded in the chart. I'll look at the rhythm. If it's goofy...then I'll get an EKG.

2) look at their past EKGs. if they have a propensity to go in / out of dangerous arrhythmias, then I'll get an EKG.

Sometimes I'll look through the chart and purposely NOT GET AN EKG. Because all of their past ones are goofy and I don't want to potentially deal with an abnormal one.

Likewise...I won't order labs unless I really have to.

FIGHT THE URGE to look under the hood on these cats. It will regularly cause you a headache and I guarantee you won't do anything to prolong their life.

These days I don't even scan the c-spine if they can move it without pain, squeeze my hands equally, and move their legs normally.

 

[ACal]

FIGHT THE URGE to look under the hood on these cats. It will regularly cause you a headache and I guarantee you won't do anything to prolong their life.

Agree 100% in that the workup will probably not prolong their life, or even increase their quality of life. I guess I just feel like there is a lot of liability in not doing a basic workup in someone whose complaint is they been feeling unwell and had a fall.

 

[gro2001]

Over a year out of residency, working a lot of evening/night shifts, still uncomfortable with old patients that come in at night...

2 scenarios:


1. 85 yo F with dementia, a million other medical problems, unwitnessed GLF fall at SNF, A&O x1 maybe x2 at baseline, doesn't remember what happened. Maybe feels a little weak recently. Probably on aspirin/plavix/Eliquis (Full Code)
I usually scan their head/neck and film whatever hurts. Basic labs/trop/ekg/urine. If I don't find anything, usually some mild leukocytosis or baseline anemia, can't really admit them for anything, maybe high risk syncope? But usually back to the SNF...I don't feel too worried about these. (Maybe I should be? Anything important not to miss?)


2. 70 yo F with HTN, DM, HLD, comes in feeling weak and "dizzy" for a day. Maybe some vertigo but not really. (Can never tell me if they feel the room spinning) Labs/cardiac workup are normal....sometimes I'll get a CT head noncon (usually normal and useless). No focal neuro deficits. They can still walk but maybe a lot slower. Nothing to admit them for. Not really even a SNF candidate. What do you do with these?

One thing I would add on 2 is that if I was concerned enough to get a CTH, I would add a CTA of the head/neck. Between the two, CTA is more likely to be useful in a case like this. Maybe you'll find a very tight vertebral or basilar artery which could both explain their symptoms and end up preventing a stroke down the line (definitely don't get a CTA without a non-con though).

 

[Janders]

I love that the advice ranges the spectrum from (A) avoid getting an EKG if you can to (B) strongly consider a CTA Head.

We’re scientists!

[and I do agree, if I’m really working up DIZZINESS that doesn’t seem like generalized weakness or clear orthostatic syncope, a CTA may very well be reasonable]

These cases got easier for me when I made them patient and family focused. Sounds corny, but ask the patient, the daughter, the HCP… whoever is there or who you have a phone number for, what they want. Do they just want a once over? Do they just want the ouchy elbow they hit checked out, and don’t care about the fall/weakness? Do they want transferred to Man’s Greatest Hospital for a 2 week specialist work up? Did they chose to come in? Did someone in the family force them in? How are they doing at home?

Turns out people like it when the doctor calls them, it rarely takes >5min, and the majority of the time you will rapidly discover that (1) This is an acute, significant change in functional baseline and they want it aggressively worked up OR (2) Nana falls all the time, they have 16hr of private care arranged every day, they really just want to keep her out of the hospital if at all possible. I’m honest with people about the utility of our tests, and even that getting a stack of “normal” things in the ED doesn’t mean Nana is fine… she is 90yo and weak. This year I’ve had a solid handful of VERY pleasant conversations with families that led to palliative care consultations, DNR/DNI/DNH MOLST forms, and likely avoided me or my partner running a futile code on Nana in the next 1-2 years. The kind of conversations that make you remember why you did this while you drudge through burnout land. And if the conversation isn’t productive, and second-cousin HCP Bob who hasn’t seen Nana in 5 years demands admission… at least you know how the board game is set up. No biggie.

Of course some patient’s / families aren’t sure what they want; in that case I do what I believe is a reasonable workup, then loop back around, tell them the results, and offer them something off the menu of Home w/ urgent PCP follow up, admission for tele-obs understanding we might not necessarily find anything, or OBSV for PT/CM in the AM to work on functional mobility, home resources, and possible Rehab placement.

In that we do a pretty solid job sending people home AND doing ED OBSV for a fair volume of elderly PT/CM, our hospitalists tend not to push back when we do ask to bring this type of patient in for an overnight obsv on their service (I can’t say they NEVER push back, but its infrequent).

 

[RoyBasch]

These cases got easier for me when I made them patient and family focused. Sounds corny, but ask the patient, the daughter, the HCP… whoever is there or who you have a phone number for, what they want. Do they just want a once over? Do they just want the ouchy elbow they hit checked out, and don’t care about the fall/weakness? Do they want transferred to Man’s Greatest Hospital for a 2 week specialist work up? Did they chose to come in? Did someone in the family force them in? How are they doing at home?

Turns out people like it when the doctor calls them, it rarely takes >5min, and the majority of the time you will rapidly discover that (1) This is an acute, significant change in functional baseline and they want it aggressively worked up OR (2) Nana falls all the time, they have 16hr of private care arranged every day, they really just want to keep her out of the hospital if at all possible. I’m honest with people about the utility of our tests, and even that getting a stack of “normal” things in the ED doesn’t mean Nana is fine… she is 90yo and weak. This year I’ve had a solid handful of VERY pleasant conversations with families that led to palliative care consultations, DNR/DNI/DNH MOLST forms, and likely avoided me or my partner running a futile code on Nana in the next 1-2 years. The kind of conversations that make you remember why you did this while you drudge through burnout land. And if the conversation isn’t productive, and second-cousin HCP Bob who hasn’t seen Nana in 5 years demands admission… at least you know how the board game is set up. No biggie.

Of course some patient’s / families aren’t sure what they want; in that case I do what I believe is a reasonable workup, then loop back around, tell them the results, and offer them something off the menu of Home w/ urgent PCP follow up, admission for tele-obs understanding we might not necessarily find anything, or OBSV for PT/CM in the AM to work on functional mobility, home resources, and possible Rehab placement.

This is really the master level move in these cases. I find when you talk to these people they usually want LESS done as far as testing, dispo, etc.

 

[gro2001]

I love that the advice ranges the spectrum from (A) avoid getting an EKG if you can to (B) strongly consider a CTA Head.

To clarify: I am not saying every dizzy person needs a CTA of the head. I am saying that for those patients who you feel need a CTH, you should also get a a CTA. Most people don't need either. Some do. For those, an incomplete/wrong imaging strategy gives false reassurance but doesn't look for the thing you want to look for.

 

[Janders]

To clarify: I am not saying every dizzy person needs a CTA of the head. I am saying that for those patients who you feel need a CTH, you should also get a a CTA. Most people don't need either. Some do. For those, an incomplete/wrong imaging strategy gives false reassurance but doesn't look for the thing you want to look for.

Someone actually just sent these to me like two days ago—

Diagnostic accuracy of neuroimaging in emergency department patients with acute vertigo or dizziness: A systematic review and meta-analysis for the guidelines for reasonable and appropriate care in the emergency department​

Diagnostic accuracy of neuroimaging in emergency department patients with acute vertigo or dizziness: A systematic review and meta-analysis for the guidelines for reasonable and appropriate care in the emergency department - PubMed

Noncontrast CT has very low sensitivity and MRI will miss approximately one in five patients with stroke if imaging is obtained early after symptom onset. The evidence does not support neuroimaging as the only tool for ruling out stroke and other central causes in patients with acute dizziness...

pubmed.ncbi.nlm.nih.gov

Results: We included studies that reported diagnostic test accuracy. From 6309 titles, 460 articles were retrieved, and 12 were included: noncontrast CT scan-six studies, 771 patients, pooled sensitivity 28.5% (95% confidence interval [CI] 14.4%-48.5%, moderate certainty) and specificity 98.9% (95% CI 93.4%-99.8%, moderate certainty); MRI-five studies, 943 patients, sensitivity 79.8% (95% CI 71.4%-86.2%, high certainty) and specificity 98.8% (95% CI 96.2%-100%, high certainty); CTA-one study, 153 patients, sensitivity 14.3% (95% CI 1.8%-42.8%) and specificity 97.7% (95% CI 93.8%-99.6%), CT had higher sensitivity than CTA (21.4% and 14.3%) for central etiology; MRA-one study, 24 patients, sensitivity 60.0% (95% CI 26.2%-87.8%) and specificity 92.9% (95% CI 66.1%-99.8%); US-three studies, 258 patients, sensitivity ranged from 30% to 53.6%, specificity from 94.9% to 100%.

Conclusions: Noncontrast CT has very low sensitivity and MRI will miss approximately one in five patients with stroke if imaging is obtained early after symptom onset. The evidence does not support neuroimaging as the only tool for ruling out stroke and other central causes in patients with acute dizziness or vertigo presenting to the ED.



AND

https://journals.sagepub.com/doi/full/10.1177/1971400920988665#:~:text=The%20American%20College%20of%20Radiology,dizziness%20as%20the%20initial%20symptom.

Conclusion​

Both non-contrast computed tomography and computed tomography angiogram of the head and neck have low diagnostic yield for the detection of central causes of dizziness, However, non-contrast computed tomography has higher sensitivity and positive predictive value than computed tomography angiogram, implying a lack of diagnostic advantage from the routine use of computed tomography angiogram in the emergency department for the investigation of isolated dizziness


Clearly there is a reason dizziness is tricky.

 

[gro2001]

Someone actually just sent these to me like two days ago—

Diagnostic accuracy of neuroimaging in emergency department patients with acute vertigo or dizziness: A systematic review and meta-analysis for the guidelines for reasonable and appropriate care in the emergency department​

Diagnostic accuracy of neuroimaging in emergency department patients with acute vertigo or dizziness: A systematic review and meta-analysis for the guidelines for reasonable and appropriate care in the emergency department - PubMed

Noncontrast CT has very low sensitivity and MRI will miss approximately one in five patients with stroke if imaging is obtained early after symptom onset. The evidence does not support neuroimaging as the only tool for ruling out stroke and other central causes in patients with acute dizziness...

pubmed.ncbi.nlm.nih.gov

Results: We included studies that reported diagnostic test accuracy. From 6309 titles, 460 articles were retrieved, and 12 were included: noncontrast CT scan-six studies, 771 patients, pooled sensitivity 28.5% (95% confidence interval [CI] 14.4%-48.5%, moderate certainty) and specificity 98.9% (95% CI 93.4%-99.8%, moderate certainty); MRI-five studies, 943 patients, sensitivity 79.8% (95% CI 71.4%-86.2%, high certainty) and specificity 98.8% (95% CI 96.2%-100%, high certainty); CTA-one study, 153 patients, sensitivity 14.3% (95% CI 1.8%-42.8%) and specificity 97.7% (95% CI 93.8%-99.6%), CT had higher sensitivity than CTA (21.4% and 14.3%) for central etiology; MRA-one study, 24 patients, sensitivity 60.0% (95% CI 26.2%-87.8%) and specificity 92.9% (95% CI 66.1%-99.8%); US-three studies, 258 patients, sensitivity ranged from 30% to 53.6%, specificity from 94.9% to 100%.

Conclusions: Noncontrast CT has very low sensitivity and MRI will miss approximately one in five patients with stroke if imaging is obtained early after symptom onset. The evidence does not support neuroimaging as the only tool for ruling out stroke and other central causes in patients with acute dizziness or vertigo presenting to the ED.



AND

https://journals.sagepub.com/doi/full/10.1177/1971400920988665#:~:text=The%20American%20College%20of%20Radiology,dizziness%20as%20the%20initial%20symptom.

Conclusion​

Both non-contrast computed tomography and computed tomography angiogram of the head and neck have low diagnostic yield for the detection of central causes of dizziness, However, non-contrast computed tomography has higher sensitivity and positive predictive value than computed tomography angiogram, implying a lack of diagnostic advantage from the routine use of computed tomography angiogram in the emergency department for the investigation of isolated dizziness


Clearly there is a reason dizziness is tricky.

Thanks for sharing these studies. Here is my critique:

The Shah study seems to lack some nuance. Stroke care is not necessarily binary: has stroke/doesn't have stroke. I am also interested in symptomatic intracranial atherosclerosis (maybe this person doesn't have a stroke, but this tight vessel is what's making them dizzy when they stand up and can't perfuse past the obstruction) and high grade carotid stenosis (maybe I won't find a stroke this time (TIA, very small stroke, imperfect imaging), but it certainly puts them at high risk for the next few days/weeks which I want to know about, and so on. So just dumping all the studies in the mix to get Sn/Sp numbers is not the full picture. I think it's telling that there were no neurologists on that paper.

The Guarnizo paper suffers from two problems. The first is selection bias. They looked at people who had a CTH and CTA done, and found reasonable congruence between the two. However, they don't have the population that had a CTH done (which didn't find anything much), were discharged, and then came back with a stroke a week later because really that was a TIA and they should have been put on DAPT, which maybe you/neurologist would have done if you saw how bad their ICAD was or saw a culprit lesion. Also, they say they didn't really have access to enough clinical data to be sure they weren't missing a fair amount of nuance.

Stroke is a vascular problem and needs a vascular study. If I am suspicious enough to get a CTH, I am usually suspicious enough to get a CTA.

 

[Janders]

Well following that logic, if your concern is stroke why aren’t you doing MR?

 

[southerndoc]

I thought the miss rate was 6% on stroke within 24 hours of onset. 20% seems a little high.

 

[gro2001]

Well following that logic, if your concern is stroke why aren’t you doing MR?

Because if the stroke is from say Afib, as long as their CHADS-VASC isn't crazy high, their annual risk of stroke is in the low single digits, meaning their chance of recurrence in the next few days before they follow up in neuro clinic is very low indeed. For example: 70 yo M, PMH HTN and DM -> CHADS VASC 3 = annual risk of stroke ~3% = risk of stroke in say 2 weeks before they get around to seeing a neurologist <<<3%. If the stroke is from symptomatic ICAD it's probably closer to like 3% per week. If it's from a hot carotid it could be like 3% per DAY. The entire annual risk of our 70 year old Afibber in one day, every day. So yeah, vessel imaging changes things quite dramatically.

 

[Clean7795]

If you want to be diagnostic about everything in the ER you're going to need a lot of imaging, a lot of labs, and a lot of consults.

Even shotgun neuroimaging (CT+CTA+CT perfusion+MRI) is insufficient to completely rule out a posterior circulation stroke. You will need to repeat a MRI in a few days to see if anything declares itself. Even if you see one on the initial imaging, there is no emergent intervention that's going to change anything before they see a neurologist. Usually a noncontrast CT head is reasonable to rule out large surprises.

Because if the stroke is from say Afib, as long as their CHADS-VASC isn't crazy high, their annual risk of stroke is in the low single digits, meaning their chance of recurrence in the next few days before they follow up in neuro clinic is very low indeed. For example: 70 yo M, PMH HTN and DM -> CHADS VASC 3 = annual risk of stroke ~3% = risk of stroke in say 2 weeks before they get around to seeing a neurologist <<<3%. If the stroke is from symptomatic ICAD it's probably closer to like 3% per week. If it's from a hot carotid it could be like 3% per DAY. The entire annual risk of our 70 year old Afibber in one day, every day. So yeah, vessel imaging changes things quite dramatically.

Dizziness from a stroke is due to the posterior circulation, i.e. vertebrobasilar arteries & PCAs. The carotid arteries are very different, and the same data doesn't apply. We have that data in the carotids because we can actually do something about it, i.e. stenting & carotid endarterectomies. For the cerebellum the patient is just going to receive standard stroke prevention.

 

[ACal]

I love that the advice ranges the spectrum from (A) avoid getting an EKG if you can to (B) strongly consider a CTA Head.

We’re scientists!

[and I do agree, if I’m really working up DIZZINESS that doesn’t seem like generalized weakness or clear orthostatic syncope, a CTA may very well be reasonable]

These cases got easier for me when I made them patient and family focused. Sounds corny, but ask the patient, the daughter, the HCP… whoever is there or who you have a phone number for, what they want. Do they just want a once over? Do they just want the ouchy elbow they hit checked out, and don’t care about the fall/weakness? Do they want transferred to Man’s Greatest Hospital for a 2 week specialist work up? Did they chose to come in? Did someone in the family force them in? How are they doing at home?

Turns out people like it when the doctor calls them, it rarely takes >5min, and the majority of the time you will rapidly discover that (1) This is an acute, significant change in functional baseline and they want it aggressively worked up OR (2) Nana falls all the time, they have 16hr of private care arranged every day, they really just want to keep her out of the hospital if at all possible. I’m honest with people about the utility of our tests, and even that getting a stack of “normal” things in the ED doesn’t mean Nana is fine… she is 90yo and weak. This year I’ve had a solid handful of VERY pleasant conversations with families that led to palliative care consultations, DNR/DNI/DNH MOLST forms, and likely avoided me or my partner running a futile code on Nana in the next 1-2 years. The kind of conversations that make you remember why you did this while you drudge through burnout land. And if the conversation isn’t productive, and second-cousin HCP Bob who hasn’t seen Nana in 5 years demands admission… at least you know how the board game is set up. No biggie.

Of course some patient’s / families aren’t sure what they want; in that case I do what I believe is a reasonable workup, then loop back around, tell them the results, and offer them something off the menu of Home w/ urgent PCP follow up, admission for tele-obs understanding we might not necessarily find anything, or OBSV for PT/CM in the AM to work on functional mobility, home resources, and possible Rehab placement.

In that we do a pretty solid job sending people home AND doing ED OBSV for a fair volume of elderly PT/CM, our hospitalists tend not to push back when we do ask to bring this type of patient in for an overnight obsv on their service (I can’t say they NEVER push back, but its infrequent).

Thank you for the advice! I agree talking to the family has been very helpful, it's rough that most family can't be there in person due to COVID restrictions, but when they do show up in person it's usually even more helpful.

2 Questions:

1. If the family member there and the patient opt to do minimal workup, assuming the patient is still full code, and has a bad outcome due to missed pathology, aren't you just as liable? Sure they might like you better but at the end of the day a bad outcome right after ED discharge seem risky for lawsuits? Especially if it's filed by a different relative?

2. How do you deal with families who have no medical literacy? Or have unrealistic expectations? "Grandma has been getting weaker and weaker, she doesn't seem right, she should get admitted" Then has a stone cold normal workup? I guess I can try to admit to hospitalist for obs once in a while but doesn't seem like a consistent solution.

 

[CoolDoc1729]

Thank you for the advice! I agree talking to the family has been very helpful, it's rough that most family can't be there in person due to COVID restrictions, but when they do show up in person it's usually even more helpful.

2 Questions:

1. If the family member there and the patient opt to do minimal workup, assuming the patient is still full code, and has a bad outcome due to missed pathology, aren't you just as liable? Sure they might like you better but at the end of the day a bad outcome right after ED discharge seem risky for lawsuits? Especially if it's filed by a different relative?

2. How do you deal with families who have no medical literacy? Or have unrealistic expectations? "Grandma has been getting weaker and weaker, she doesn't seem right, she should get admitted" Then has a stone cold normal workup? I guess I can try to admit to hospitalist for obs once in a while but doesn't seem like a consistent solution.

Everyone can correct me if I’m wrong , but I feel like the “damages” are fairly limited in a bedbound 90year old that dies “due to a missed diagnosis” (in quotes because at some point there’s mortality measurable per day anyway).. a lot of the bigger lawsuits are due to either lost wages or need for expensive ongoing care, and the old ones have no wages and have Medicare … right?
Admittedly I get basically no pushback from my admitting teams, but I don’t shove people out the door if they are concerned about safety at home. I’m not sure what the solution is longer term because there will keep being more of them and less of us.

 

[GassYous]

Everyone can correct me if I’m wrong , but I feel like the “damages” are fairly limited in a bedbound 90year old that dies “due to a missed diagnosis” (in quotes because at some point there’s mortality measurable per day anyway).. a lot of the bigger lawsuits are due to either lost wages or need for expensive ongoing care, and the old ones have no wages and have Medicare … right?
Admittedly I get basically no pushback from my admitting teams, but I don’t shove people out the door if they are concerned about safety at home. I’m not sure what the solution is longer term because there will keep being more of them and less of us.

I feel like eventually things will come to a head and the people in charge will realize that doing everything for everyone is unsustainable from a financial, economic, manpower, environmental etc standpoint but I thought we were there during covid when we were running out of everything. It didn't happen though. Hail marys and million dollar workups for everyone.

 

[gro2001]

If you want to be diagnostic about everything in the ER you're going to need a lot of imaging, a lot of labs, and a lot of consults.

Even shotgun neuroimaging (CT+CTA+CT perfusion+MRI) is insufficient to completely rule out a posterior circulation stroke. You will need to repeat a MRI in a few days to see if anything declares itself. Even if you see one on the initial imaging, there is no emergent intervention that's going to change anything before they see a neurologist. Usually a noncontrast CT head is reasonable to rule out large surprises.



Dizziness from a stroke is due to the posterior circulation, i.e. vertebrobasilar arteries & PCAs. The carotid arteries are very different, and the same data doesn't apply. We have that data in the carotids because we can actually do something about it, i.e. stenting & carotid endarterectomies. For the cerebellum the patient is just going to receive standard stroke prevention.

1) I am not suggesting anyone shotgun anything. I certainly don't think I shotgun very much at all. I agree that you can't be completely sure even with an MRI. And I also agree that a CTH is reasonable and does rule out a lot of big things. A CTA does take a few more big surprises off the table though.

2) Yes, the data is for carotids. I threw it in there to demonstrate the importance of vessel imaging in stroke in principle, which is what I understood was being questioned (ie my statement that stroke is a vascular problem and needs a vascular study). Although we don't have the same data for posterior circulation strokes, and INR interventions are more limited for posterior strokes, it's inaccurate to say that nothing except standard stroke care would be done. I can think of a couple of cases where a CTA for a posterior circulation stroke would be management changing:

-some centers do absolutely thrombectomize basilars and P1 segments of the PCAs. To be fair, this is probably not the case described as usually that's reserved for disabling strokes (the original case we started discussing can still walk, so I doubt most interventionalists would take it, also the timeline of a day is probably a little late). But as a more general point of highlighting the importance of vessel imaging in posterior circulation strokes I think it still stands.

-when there is a lesion that's not intervenable but seems pressure dependent some centers have protocols for blood pressure augmentation to keep the area at risk perfused. In our system we would transfer to the mothership to do this in the stroke unit or ICU.

-not to knock my fellow ER docs' neuro exams (I've seen this with neuro residents as well) but sometimes isolated dizziness isn't just isolated dizziness and is also a little hoarseness and a bit of a facial droop, and then it's an AICA/PICA stroke, and maybe depending on how much space they have they should be admitted for a SOC watch. Again, maybe the 70 year old would end up not getting offered a SOC by the surgeons, or already has a lot of extra space and wouldn't need one even if she swells, but that's the kind of nuance I'd rather discuss after I see/don't see a sizable occlusion.

Edit: SOC = suboccipital craniectomy

 

[thegenius]

SOC?

 

[turkeyjerky]

SOC?

Sub occipital craniectomy?

 

[gro2001]

Yes, sorry, SOC = suboccipital craniotomy.

Large posterior fosa strokes don't have a lot of place to swell. There's skull on one side, tentorium on the other, so the path of least resistance is straight through the brainstem.

 

[thegenius]

what does your handle mean? It has SOCMOB
Sub Occipital Craniotomy Mother Of Boob?

 

[RustedFox]

what does your handle mean? It has SOCMOB
Sub Occipital Craniotomy Mother Of Boob?

C'mon man.

 

[gro2001]

what does your handle mean? It has SOCMOB
Sub Occipital Craniotomy Mother Of Boob?

lol no.

It's the age old EM adage: standing on corner, minding own business.

As in "I swear doc, I was just standing on the corner, minding my own business, when these 3 dudes jumped me"

 

[thegenius]

lol no.

It's the age old EM adage: standing on corner, minding own business.

As in "I swear doc, I was just standing on the corner, minding my own business, when these 3 dudes jumped me"

Oh it should be
SOC711MOB
I was standing on the corner...at the 7-11...minding my own business

then I GOT SHOT!

 

[CoolDoc1729]

lol no.

It's the age old EM adage: standing on corner, minding own business.

As in "I swear doc, I was just standing on the corner, minding my own business, when these 3 dudes jumped me"

It’s actually why I stopped going to church … didn’t want to get shot on the way ..

 

[turkeyjerky]

It’s actually why I stopped going to church … didn’t want to get shot on the way ..

also high risk of aborted sudden death while there