Tag Archives: laboratory tests

A Blood Test For TBI? Part 3

The FDA announced approval of a blood test that incorporates both GFAP and UCH-L1. Approval was based on two as yet to be published studeis titled Evaluation of Biomarkers of Traumatic Brain Injury (ALERT-TBI) and Evaluation of Biomarkers of Traumatic Brain Injury Extension Study (ALERT-TBIx), and passed after less than 6 months of evaluation. Yes, more silly acronyms, I know.

The studies were designed to “evaluate the utility of the Banyan UCH-L1/GFAP Detection Assay as an aid in the evaluation of suspected traumatic brain injury (Glasgow Coma Scale score 9-15) in conjunction with other clinical information within 12 hours of injury to assist in determining the need for a CT scan of the head.”

The former study started in 2012 and involved 2011 patients! The latter had only 119 patients, starting in 2015. Now, I have no access to their data, so I can’t tell what the FDA saw.

From Banyan Biomarkers’ website:

“The CT scan is widely available to assist clinicians in the evaluation of TBI, however, CT scans do not provide a clear and objective answer and scanning may increase the risk for radiation-induced cancer. Furthermore, over 90% of patients presenting to the emergency department with mild TBI, sometimes described as “concussion”, have a negative CT scan. Despite these limitations, nearly all patients are sent for a CT, which results in increased costs to the healthcare system and unnecessary patient exposure to radiation.”

Here are the (very) few factoids that I can find:

  • CT scan results were compared to the Brain Trauma Indicator (BTI) blood test (GFAP + UCH-L1)
  • BTI predicted a positive CT scan 98% of the time
  • It predicted a negative CT scan 99.6% of the time
  • Time to process the test is currently 4 hours

Bottom line: Sounds promising, right? Based on the data summarized over the last two days, I wouldn’t be too excited about this test, but the FDA was able to look at a study that I can’t. It appears that the negative predictive value is excellent, so I can see the application.

That being said, 4 hours is way to long. We can’t have a patient sitting in the ED waiting for the results to come back to decide whether they need a head CT. And how long will it take the assay to be widely available?

The devil will be in the details. What types of intracranial lesions were detected. Are the negative predictive values the same for subarachnoid, subdural, epidural, or intraparenchymal bleeds? And finally, how expensive will it be? How does the cost for the test compare to the cost of a CT scan done in 5 minutes?

I’ll let you know more as the details emerge. But don’t look for, or plan to use, this test at your hospital any time soon. There’s more work to do!

Reference: Banyan Biomarkers (banyanbio.com)

A Blood Test For TBI? Part 2

Yesterday, I wrote about one blood biomarker, GFAP, and its possible application in detecting traumatic brain injury (TBI). Today, I’ll discuss a complementary marker called UCH-L1.

Fewer studies have been done looking at the utility of UCH-L1 in detecting TBI than of GFAP.  A review article published last month pooled existing literature to get a sense of how good this biomarker really is. It also examined the risk of bias due to the small numbers of studies involved.

Here are the factoids:

  • Only 38 abstracts were eligible, but full text was available for analysis in only 13 (meaning it was only an abstract and never passed muster for publication). The authors of the published studies were contacted for additional information, which is an interesting (and helpful) practice.
  • Of all of those, only 4 were selected for meta-analysis! This significantly limited the value of the analysis.
  • Serum UCH-L1 has a high accuracy in predicting CT findings in mild to moderate TBI, but there is a high risk of bias affecting this result
  • Plasma UCH-L1 has a moderate accuracy predicting CT findings across all GCS levels, with a low risk of bias
  • Pooling all studies, this is high accuracy in predicting CT findings in patients with TBI across all GCS levels, but there is a high risk of bias affecting the results

Bottom line: UCH-L1 show promise as a predictor of CT findings in patients with TBI. However, the research papers were few and far between, and the possibility of bias was high. What does this mean? That using this test alone is better than a coin toss, but not good enough to change our practice in ordering CT scans in head injured patients. More well-designed studies are needed tell us whether this new (and undoubtedly expensive) test is worth the trouble.

Tomorrow, I’ll discuss a blood test incorporating both UCH-L1 and GFAP that was recently approved by the FDA.

Reference: The diagnostic values of UCH-L1 in traumatic brain injury: A meta-analysis. Brain Inj 32(1):1-17, 2018.

A Blood Test For TBI? Part 1

Traumatic brain injury (TBI) is an extremely common problem encountered by trauma professionals. Diagnostic and management pathways are fairly well-defined, and rely mainly on physical examination, as well as CT imaging in select cases.

In recent years, work has been done to identify markers of brain injury in the blood. The theory is that the injured brain may release substances that can be assayed with a simple blood test. The presence of these blood markers could then influence our use of CT for diagnosis, decision to admit or send home, and possibly help identify patients likely to develop post-concussive symptoms.

Two particular biomarkers are being evaluated: UCH-L1 and GFAP. A recently published review examined the current status of GFAP in diagnosis of head injury.

Here are the factoids:

  • A total of 27 pertinent research papers were identified for review, and 24 of 27 demonstrated a positive association between GFAP levels and TBI
  • GFAP prediction of intracranial pathology by CT scan was good to excellent
  • GFAP appeared to be able to discriminate between mass lesions and diffuse injury
  • There was considerable variability in the average GFAP values. This means that the cutoff value that predicts significant injury is not yet clear.
  • The number of pediatric studies reviewed was low, so the results may not be generalizable to children
  • GFAP may be elevated in patients with orthopedic injuries, and this was not well controlled for in the studies reviewed. It is unclear whether GFAP can be used in patients with fractures.

Bottom line: GFAP looks promising as a marker for detecting significant TBI in some trauma patients. 

Tomorrow, I’ll take a look at the other biomarker, UCH-L1, and the following day I’ll discuss the recent FDA approval of an assay for both of these by a US company, Banyan Biomarkers.

Reference: A systematic review of the usefulness of glial fibrillary acidic protein for predicting intracranial lesions following head trauma. Frontiers in Neurology 8(652):1-16.

I have no financial interest in Banyan Biomarkers.

Serial Lab Testing: Worthwhile or Worthless?: Final Answer

In my last two posts, I detailed the serum sodium measurements in a hypothetical patient two ways. The first was a listing of daily values, and the second provided values obtained every six hours or so. It also showed the sodium supplementation that was ordered based on those values. (I’ve included the table at the bottom of this post)

What did you think? Did the extra determinations help you decide what, if any, treatment was needed? Did the therapies ordered help?

Here are my thoughts:

  • Overall, there was not a huge or rapid decline in sodium values. Given the initial values, I would not have started a saline infusion on day 1, just watched a few daily values and the patients physical exam. The infusion only provided 3gm of salt per day, and the serum Na remained fairly stable for the first 3 days.
  • There was a significant amount of intra-day variation seen on the six hour table. You need to know the normal “within-person ” variation for any lab test you order. If two assays on specimens drawn at the same time can vary by 5%, you must factor this in to your decision making. If the value is 3% lower than the previous draw, the difference could represent normal variation. Obtaining more frequent assays exacerbates the amount of variation you see and my be confusing.
  • From day 5 to 6, the sodium appeared to be rising without any salt supplementation! But then a higher dose was given, and one of the intra-day values dropped to 124. What’s up with that? More variation?!
  • Is the morbidity of frequent blood draws worth it if there is no clinical change in the patient’s exam? What morbidity, you ask? Sleep disturbances, with all the cascading problems like delirium, sundowning, administration of additional meds to compensate, and on and on. Unnecessary medication or interventions. Plus it does not promote patient or family satisfaction at all.

Bottom line: Unless your patient has a clinical problem that may deteriorate rapidly, serial lab determinations are probably not of much value. The example patient was many days out from a TBI with some extra-axial blood. So yes, he could develop hyponatremia, but it would have probably surfaced earlier. Know your within-person  variability, which for sodium is roughly +2 meq. Is your new value within that limit? Then it is statistically the same as the first value unless you see a trend over several measurements. And as always, if you note a marked change in just one value, repeat it immediately before beginning any more drastic interventions.

Reference: Biological variation of laboratory analytes based on the 1999-2002 national health and nutrition examination survey. Natl Health Statistic Reports 21:March 1, 2010.

Day/Time Na Treatment NaCl per day
Day 1 18:30 131
Day 1 22:54 132 0.9% NS @ 125/hr 3G
Day 2 05:59 133 continues 3G
Day 2 12:19 129 continues
Day 2 17:50 129 continues
Day 3 07:18 127 continues
Day 3 12:09 127 continues
Day 3 17:58 126 continues
Day 3 23:53 126 continues
Day 4 07:45 125 continues
Day 4 11:38 122 2% NS @ 25/hr 6G
Day 4 15:25 125 continues
Day 4 19:31 125 continues
Day 5 00:06 122 continues 6G
Day 5 04:04 126 continues
Day 5 08:01 122 continues
Day 5 11:50 132 stop
Day 5 16:14 126
Day 5 19:26 127
Day 6 00:20 129 9.2G
Day 6 04:42 127 2% NS @ 40/hr
Day 6 08:30 124 continues
Day 6 12:29 127 stop
Day 6 16:16 127 Salt tabs 2G tid
Day 6 20:28 132 continues
Day 7 05:22 134 Salt tabs 2G qid 8G
Day 7 12:33 135 continues
Day 8 07:02 131 stop None
Day 8 13:33 136

Serial Lab Testing: Worthwhile or Worthless? Part 2

Yesterday, I posted a series of sodium levels that were drawn daily. There was no change in clinical status as the levels varied from 131 to 125 and back up.

Now let me give you a bit more information. The patient was actually getting serial checks every 6 hours (or more)! Here’s the updated chart:

Day/Time Na Treatment NaCl per day
Day 1 18:30 131
Day 1 22:54 132 0.9% NS @ 125/hr 3G
Day 2 05:59 133 continues 3G
Day 2 12:19 129 continues
Day 2 17:50 129 continues
Day 3 07:18 127 continues
Day 3 12:09 127 continues
Day 3 17:58 126 continues
Day 3 23:53 126 continues
Day 4 07:45 125 continues
Day 4 11:38 122 2% NS @ 25/hr 6G
Day 4 15:25 125 continues
Day 4 19:31 125 continues
Day 5 00:06 122 continues 6G
Day 5 04:04 126 continues
Day 5 08:01 122 continues
Day 5 11:50 132 stop
Day 5 16:14 126
Day 5 19:26 127
Day 6 00:20 129 9.2G
Day 6 04:42 127 2% NS @ 40/hr
Day 6 08:30 124 continues
Day 6 12:29 127 stop
Day 6 16:16 127 Salt tabs 2G tid
Day 6 20:28 132 continues
Day 7 05:22 134 Salt tabs 2G qid 8G
Day 7 12:33 135 continues
Day 8 07:02 131 stop None
Day 8 13:33 136

Confused? Me, too! This poor person had 30 blood draws in 8 days, with 6 per day for two of those days. Carefully look at the amount of salt given in each 24 hour period, and look at the sodium levels for that day.

See the variability, even when getting high doses of sodium chloride? What does this tell you? Was the salt administration helpful? Was seeing the lab value every 4-6 hours valuable?

Tell me what you think. Leave comments or tweet your opinions. Next, I’ll discuss the known variability of the serum sodium assay, and give you my opinion on the value of serial testing.