Welcome to Last Minute Lecture.
This free chapter overview is designed to help students review and understand key concepts.
These summaries supplement not replaced the original textbook and may not be redistributed or resold.
For complete coverage, always consult the official text.
Welcome.
We are really glad you're here studying with us today.
Yeah, it's great to have you.
Because if you're a college student stepping into the complexities of clinical biochemistry for the first time,
you know the material can get incredibly dense, incredibly fast.
Oh, absolutely.
It can feel like hitting a brick wall.
Right.
So consider this deep dive, your dedicated one -on -one tutoring session to cut through all that noise.
Today, we are focusing entirely on chapter one of clinical biochemistry and metabolic medicine, the eighth edition.
The foundational chapter.
Exactly.
We're looking at requesting laboratory tests and interpreting the results.
And I am thrilled we're doing this deep dive.
Yeah.
Because here is the fundamental reality of clinical biochemistry.
Getting a lab result back is not a simple pass or fail grade.
It really isn't.
It's not this magical piece of paper that just hands you a definitive diagnosis on a silver platter.
Our mission today is to understand the biochemical, the physiological, and especially the statistical why behind those numbers.
I think the best place to start is, well, with a basic question that actually gets overlooked in practice a lot.
Why do we even order these tests to begin with?
Right.
What's the purpose?
Mostly, we're looking to diagnose or screen for a metabolic disease or maybe monitor a patient's response to treatment, detect complications.
Sometimes it's for medical, legal, or research reasons, but there's this massive temptation, especially when you're first starting out, to just adopt a scatter gun approach.
Oh, the scatter gun approach.
You're drawing blood, so the instinct is to just check off every single box on the form.
Which is a highly dangerous instinct, frankly.
The reality is that more information isn't always better.
Right.
In fact, over -investigating a patient causes tangible harm.
You're subjecting them to unnecessary discomfort.
You can severely delay their actual treatment while waiting for completely irrelevant results, and it's a huge waste of laboratory resources.
So you really have to approach every
critical filter.
You do, and the golden rule is straightforward.
Before you request a test, you have to ask yourself, will this result actually influence the clinical management of the patient?
Will it change what we do?
Exactly.
If the answer is no, it's just noise.
Okay, let's unpack this, because this applies heavily to timing as well.
Let's say you have a patient admitted to the hospital.
There's this assumption that they need their blood drawn every single morning at 6 a .m.
A very common assumption.
But metabolically, that just doesn't make any sense for a lot of parameters, does it?
It doesn't.
I mean, you rarely need laboratory investigations more than once a day, unless the patient is undergoing highly intensive, rapidly changing therapy.
Right.
The timing should be dictated by two things.
One, how quickly are significant changes actually likely to occur in the body, and two, again, while seeing that change alter your
So think about thyroid stimulating hormone, TSH, or even plasma proteins.
The biochemistry of the body simply doesn't move fast enough for those markers to change significantly in less than a week.
So ordering a TSH test every 24 hours is literally just a complete waste of time.
Right, but on the other hand, you have markers that do change rapidly.
Take plasma transaminases in a patient with acute hepatitis.
Those enzyme activities might fluctuate significantly
They might.
But here's the catch.
Once you've already made the diagnosis of acute hepatitis, watching that transaminase number bounce around every day isn't going to change your management of the condition.
Oh, I see.
So even though the number changes fast,
repeated daily testing still isn't justified.
Contrast that with something like plasma potassium.
Yes.
If you have a patient on large doses of diuretics, their potassium levels can plummet incredibly fast.
Very fast.
And unlike the hepatitis example, a rapid dream in potassium dictates an immediate change in clinical intervention.
You have to adjust their electrolytes right away.
So frequent monitoring is absolutely necessary there.
You also have to factor in the reality of the lab environment itself.
There's a great visual in the text, figure 1 .1, showing a modern laboratory.
They are built around these massive,
highly automated analyzers, capable of processing hundreds of samples a day.
Though I know point -of -care testing is getting more common to shorten turnaround times.
It is, but for the main lab, they have strict workflows.
If you need an immediate answer that will instantly alter management, you have to clearly flag that sample as urgent.
And that communication definitely goes both ways.
Labs have what are called panic limits.
Yes, panic limits.
These are highly abnormal, critically life -threatening results.
And when a machine hits a panic limit, the lab staff are required to immediately contact the medical team.
Immediately.
Which highlights a logistical point that's just as vital as the biochemistry -accurate patient location details.
Oh yeah.
If a lab tech sees a life -threatening potassium level, they cannot spend 20 minutes trying to figure out which ward the patient is actually on.
That perfectly underscores the most important rule of clinical interpretation.
Treat the patient, not the number.
Always.
Once that lab report is in your hands, the interpretation begins.
And the first rules are deceptively simple.
First, is it the right patient?
Mix -ups happen.
Second, does the result fit the clinical findings?
Because context is everything.
If the patient sitting in front of you looks perfectly healthy, but their lab results indicate they should be in a coma.
You don't assume they have superhuman tolerance.
Exactly.
Your immediate assumption should be an analytical error or a mixed -up sample.
Which brings us to defining normal.
We constantly compare results against reference ranges.
But how does a lab actually define normal?
Well, a reference range is typically established by taking test results from a designated healthy population and calculating the spread.
Good.
For most biochemical tests, these results fall into a Gaussian distribution, you know, the classic bell curve.
Sometimes it's skewed, but usually the reference range is set to include 95 % of that healthy population.
Wait, if the reference range intentionally only covers 95 % of healthy people, doesn't that inherently mean 5 % of completely healthy people will be flagged as abnormal?
That is precisely what it means.
It's a huge statistical trap.
2 .5 % of perfectly healthy individuals will fall above the upper limit, and 2 .5 % will fall below the lower limit.
Just naturally.
Just naturally.
So a result outside the reference range does not guarantee disease.
The further from the mean, the higher the probability of true abnormality, sure.
But conversely, a normal result doesn't guarantee health.
There's an incredible compounding effect here too, right?
Figure 1 .2 in the text illustrates this overlap between normal and ill populations perfectly.
And it introduces the multi -test trap.
The math here is mind -blowing.
It really is.
Run one test, there's a 5 % chance of an abnormal result just by luck.
What happens when you run a massive panel of 20 tests?
There is a 64 % chance that at least one result will be statistically abnormal.
Over 60%.
Just by chance.
This is exactly why the scatter gun approach is so dangerous.
Let's look at case one from the text to make this real.
Let's do it.
A four -year -old boy comes in abdominal pain.
The team draws blood, and his alkaline phosphatase or ALP comes back at 326 units per liter.
The lab report states the adult reference range is less than 250.
So on paper, it's flagged as highly elevated.
But the reveal here is all about physiology.
ALP is an enzyme heavily involved in bone turnover.
And what is the four -year -old boy doing?
Drawing.
Exactly.
His bones are actively developing.
The issue wasn't the patient, it was the computer.
It used adult reference ranges.
Right.
For a four -year -old child, a normal ALP is actually anywhere from 60 to 425.
So that child is completely, perfectly normal.
And that leads to a much broader issue regarding demographic variables with the book calls between individual differences.
Age is one, sex is another.
Males naturally have higher year rate.
Premenopausal women have higher HDL.
But ethnicity is a factor that frequently gets overlooked.
It's a critical factor.
Look at case 2.
We have a 54 -year -old Nigerian man who presents with chest pain.
Normal ECG, but his creatine kinase, his CK, is 498 against a reference range of less than 250.
Which usually suggests a heart attack.
Right.
But his troponin T, a highly specific marker for cardiac injury, was perfectly normal, less than 20.
So his heart is fine.
Why is his CK double the limit?
Because of his racial origin.
Normal CK can be two to three times higher in black populations compared to the white UK population that the reference range was originally built on.
So his high CK was completely physiological.
Entirely physiological.
Okay.
So we have differences between people, but then we have within individual variations, our own bodies trick the tests constantly.
They do.
You have regular circadian rhythms.
Plasma iron drops by 50 % from morning to evening.
50%.
Yes.
Cortisol spikes with you have monthly changes like Ostradiol during the menstrual cycle, even seasonal changes like vitamin D peaking in the summer.
And then you have random variations.
Meals obviously affect glucose, which is why fasting is required.
Yeah.
But here's where it gets really interesting.
I want to talk about the physical blood draw.
Ah, plasma versus serum.
Yes.
Could you explain the actual difference because people use the words interchangeably?
They do, but biochemically they're entirely different.
Plasma is the aqueous phase of anticoagulated blood.
So it hasn't clotted.
Right.
Serum is the aqueous phase of clotted blood.
That clotting process changes the chemistry.
Dramatically.
Potassium is higher in serum because during clotting cells rupture and leak their intracellular potassium into the fluid.
And total protein is lower in serum because fibrinogen is actually removed during the clotting process.
That's fascinating.
And speaking of potassium, there's cellular versus extracellular issue.
Yes.
Normal plasma potassium might completely hide dangerous intracellular depletion like you see in diabetic ketoacidosis.
Wow.
And then there's posture.
This blew my mind.
Simply lying down recumbency for 30 minutes drops your protein and albumin levels by up to 15%.
Just because fluid shifts in the body, it drops the concentration.
Incredible.
Okay.
We need to introduce the final hurdle.
All right.
Diagnostic performance.
The math of medicine.
The lab machinery itself.
Right.
Estimations should be reproducible to well within 5%.
Sodium and calcium are very precise.
Hormones, less so.
This is measured by the coefficient of variation or CV.
How does that math work?
It's simply the standard deviation divided by the mean times 100.
It measures in precision.
You want that CV percentage to be as small as possible.
Makes sense.
Let's introduce case three to talk about sensitivity and specificity.
Say we have 100 patients with chest pain and a test finds 80 of them are positive.
Then we have 100 patients without chest pain and the test finds 95 or negative.
Okay.
So let's walk through it.
Sensitivity is your true positive rate.
Out of the 100 sick people, we caught 80.
So 80 over 100 is an 80 % sensitivity.
And specificity.
That's your true negative rate.
Out of 100 healthy people, we correctly identified 95.
95 % specificity.
And in real life, this translates into predictive values.
Exactly.
Negative predictive value, which is true negatives divided by all negatives, is vital for screening.
We need it to be high so we don't miss sick people.
And positive predictive value is vital so we don't treat healthy people unnecessarily.
Precisely.
But what happens when we move the goalposts?
Like changing the cutoff limits for a test.
It's a seesaw effect.
If you lower the cutoff, you increase sensitivity, you catch more disease, but you ruin specificity because you get more false positives.
And raising it does the exact opposite.
Right.
We visualize this using ROC curves.
Receiver operating characteristic curves.
Basically, you plot this trade off on a graph.
Like figure 1 .3 in the book.
Yes.
A bigger area under the curve means a better test.
Test A beats test B.
And test C, which is just a straight diagonal line, is literally just a coin flip.
And clinicians use likelihood ratios to figure out the statistical odds of a test being useful, right?
Yeah.
A positive likelihood ratio is just sensitivity divided by 1 minus specificity.
It tells you the odds that a positive result actually means the patient has the disease.
So what does this all mean?
We've covered so much ground today.
To synthesize all this, biochemical tests are tools.
They are not verdicts.
You always have to look at the patient, communicate with the lab, and remember the massive variables at play.
It's so true.
I'll leave you with this brain teaser to think about.
If a simple 30 minute nap, the time of day, and the fact that blood clotted in a tube can entirely change what is quote unquote normal,
at what point does a reference range stop being a scientific law and start being just a subjective snapshot of a single microscopic moment in time?
That is a great question.
Thank you for joining us on this deep dive.
We'll catch you next time.