Home / Policy / Too Many CT Scans … or Not Enough?

Too Many CT Scans … or Not Enough?

Scary findings.

Patients using Coumadin who have any head injury need repeat CT scans.

The study looked at 116 patients who were taking Coumadin and who had any head injury with a GCS of 14 or 15 – regardless of loss of consciousness (patients with lower GCS were presumably at higher risk of intracranial bleeding). CT scans were performed on all patients. Of those initial 116, nineteen patients (16%) had bleeding on their initial exam. Of the remaining 97 patients with normal initial CT scans, ten refused to be in the study. Repeat CT scans were performed on the remaining 87 patients 24 hours after the first normal CT scan and showed 5 cases of new hemorrhage. Three of those patients required hospitalization and one delayed bleeding patient required brain surgery.
Even after a normal CT scan 24 hours later … two additional patients still developed symptomatic subdural hematomas — one patient 2 days later, one patient 8 days later. Both of those patients had INRs greater than 3.0. The study recommends admitting patients overnight and repeating CT scans in 24 hours. Original study here (.pdf).

While admission and repeat CT scan for minor head trauma hasn’t become the standard of care in the United States, this study raises questions about the optimal care of minor head injuries in patients taking blood thinners.

Also at issue is the Medicare policy not to pay for “normal” CT scans of the head in atraumatic headaches. Will this policy spill over to deny elderly nursing home patients from receiving CT scans when they can’t remember whether they have hit their heads?


  1. I saw this study and then imagined the conversation after calling the hospitalist to admit someone with “minor head injury” for 24 hours. Now I can’t stop laughing…..

  2. Options would be to develop a collaborative guideline with trauma unit ( general surgery , not neurosurgery ) or keep patient in ED observation unit till second scan. Or if GCS remains 15 with normal neurology observe and scan at 24 hours only?

  3. Shaking my head…I think some things are best left unstudied…For as many head CTs as we get on older patients on coumadin this could lead to a giant new cost burden that I’m not sure is entirely warranted.

    Instead of spending the money for an inpatient admission- a happy medium could be to have the patient return to the ED in 24 hours for a followup CT. That would at least save the cost of the admission (although are two ED visits worse than an ED visit and an admission as far as costs? I honestly don’t know…)

    Of the 5 patients with the delayed bleeding (~6% of the total)- only 1 of those 5 had changes in neuro status during their observation period and only 1 (I’m assuming it was the same patient) got a craniotomy.

    So all of this would be adding a burden of 87 admissions to catch only 1 decline in neuro status and surgically intervene on 1 patient. That is a yield of 1.1%. Not even our “low risk chest pain units” have a rule-in rate that low. So if you really don’t want to miss anyone and can’t stand a 1% risk tolerance then having the patient return for a repeat head CT would be a better solution.

    I think there are better places to be spending our healthcare dollars. I’m not advocating rationing or euthanizing the elderly- just that the suggestions by the authors of the study (both implied and not implied) seem really overblown and unnecessary.

  4. What about the radiation? Although ..brain bleed trumps that. And perhaps doesn’t matter as much in elderly population ..but younger people can be on coumadin too.

    I would not want to have to make the kinds of decisions physicians routinely have to make ..between a rock and a hard place decisions, but then that is why you all wear the white coat and I do not.

  5. My Gram is on coumadin…and the one thing I have noticed over the last three years is that she has lost her ability to verbally communicate with any coherence. She has remained quite mobile and spry,and I have not noticed any falling or unexplained bruises. I wonder if this has an effect somehow in her decreased capacity and it has not been tested?

    • It could be a plethora of things. Any of multiple forms of dementia. NPH – although usually causes balance issues. Stroke/expressive aphasia. Psych issue. Medications. Chronic bleed is a possibility, but less likely. Has her doctor been notified/noted a change?

  6. Matt must be salivating. I have a headache.

  7. That image appears to be an epidural rather than a subdural hematoma.

  8. Ugh. What a disastrous study. I wonder how many of the patients in the group that have bleeds picked up 24 hours later had any new or worsening symptoms. (no I did not read the study – guess I should)
    Regardless, this will be a mess trying to get these people admitted and have this paid for. I think a much better study of several thousand people would have to be done to show enough benefit to warrant the monumental costs.

  9. Oh, and yeah Scalpel, I thought that looked like an epidural myself.

  10. So…. why not wait 24 hrs to do the CT scan in the absence of concerning S & S?

  11. Scalpel is still around – YAY! :)

Leave a Reply

Your email address will not be published. Required fields are marked *