Here’s our last posting on laboratory tests; completing those tests in the typical “Comprehensive Metabolic Panel,” plus a few other miscellaneous blood chemistries.
Since around 40% of total serum Calcium is bound to serum Albumin, changes in the latter affect results of the former. Therefore, never order a Calcium level without also ordering an Albumin. The correction factor is as follows:
- For every 1.0 g/dL Albumin above or below 4.0, correct the Calcium by 0.8 mg/dL.
- In other words: (Albumin minus 4.0) x 0.8 = correction to add/subtract to Calcium
Some examples of spurious hypercalcemia (assuming lab normals are 8.5-10.3):
- Calcium “high” at 11.0, but Albumin is 5.0. So Calcium is really 10.2.
- Calcium “normal” at 10.3, but Albumin low at 3.0. So Calcium is really 11.1.
Calculating low serum calciums works the same, going downward.
Another option is to order an “Ionized Calcium” giving you the free calcium, which is always the most accurate. The problems with that are 1) it’s more expensive; 2) it’s an extra step after having obtained the initial CMP; and 3) the serum has to be frozen, and lab often forgets to do so, so you have to call the patient back to redraw blood.
Unfortunately, in renal failure the above calculations don’t work well, so the lower the GFR (i.e. the higher the creatinine), the more you should order an ionized calcium. This is also true in acid-base disorders (maybe suspected by a very low serum bicarb). Multiple Myeloma can also skew serum calcium upward; in patients with known disease, ionized calcium is necessary, but for the undiagnosed patient, it just means that Multiple Myeloma is in the differential of hypercalcemia.
Serum glucose measurements are accurate; the main question is how they are used. Diabetes Mellitus can be diagnosed by two abnormal measurements on different days: Fasting Blood Sugar (FBS) ≥126 mg/dL [after at least 8 hours without caloric intake) and/or Glycated Hemoglobin (A1C) ≥6.5%. A Glucose >200 mg/dL at 2 hrs. by formal glucose tolerance test (75 gm. glucose) can also be used, but this is rarely performed outside of pregnancy.
“Two measurements” means 2 of the same, or 1 of each FBS & A1C. Some clinicians also tend to include any random glucose >200, but this is potentially inaccurate if, for example, the patient had consumed a larger glucose load (>75 gm) only 1 hr. ago. A random glucose of >200 should only be used as a criteria if accompanied by symptoms like polydipsia / polyuria. Of course, those don’t tend to occur until the glucose approaches 350 (at which point I’m happy to say “You’ve Got Diabetes!” even without them).
I’m not sure what “2 different days” is exactly meant to accomplish. If the purpose is to rule out lab error, any 2 days would suffice. If it’s to determine persistent pathophysiology, then an FBS today & tomorrow would suffice. But a slightly abnormal A1C (say, 6.6%) today & tomorrow would be measuring the same 2-3 month period that A1C’s cover. So I’d repeat the A1C at least 3 months later.
An aside, regarding Errors in A1C Measurement — Today’s assays tend to be accurate. Racial differences are probably too minimal to matter. The main source of error has to do with RBC production & survival. Iron, B12, and folate deficiencies result in false-high A1C’s, while hemolysis or correction of these deficiencies results in false-low values. The A1C may underestimate mean glucose in patients with high MCVs (especially if due to HIV antivirals). Abnormal hemoglobins (Hb S, etc.) may interfere depending on the A1C method used; see the website http://www.ngsp.org/interf.asp if you have such patients.
Don’t make a diagnosis of “Diabetes” from fingerstick measurements of either glucose or A1C. Always repeat it through a bona fide lab using venous blood. In terms of known diabetics who use glucometers, measurements can easily become unreliable due to dirty fingers (maybe with spots of sugar on them!) or lack of calibration.
I tend not to like to label patients. For example, when I was practicing in the early 1980’s, we defined diabetes by an FBS >140. Then I left the country for 4 years, only to discover upon my return that the number had dropped to 126. Why? And, why the “126” instead of a more rounded 120, 125, 130, etc?
Well, it turns out that Europe’s units are millimoles (mmol) instead of mg/dL. And the nice, round “7.0 mmol” used there corresponded to our 126 mg/dL. So we conceded, and lowered our definition. That didn’t sound exactly “scientific” to me.
Obviously, diabetes is a bad disease. But I’m not sure what the difference is between, “A1C 6.7% = It’s Diabetes — Bad!” and “A1C 6.7% = Diabetes Under Control — Good!” (“control” = <7.0%). I suppose if they didn’t know they had diabetes, and were that high on their baseline diet, it’d be “bad.” Whereas if they were following medical advice and modified their diet to reach 6.7% it’d be “good.”
Then there’s “Pre-Diabetes.” We used to say an FBS >100 was “impaired fasting glucose,” but now it’s an actual disease / condition. What it really means is that such persons have an increased probability of being diagnosed with frank diabetes in the future. To me, explaining this in terms of risk is much more honest than slapping on a label.
I work primarily with a Latino population; lots of my patients have FBS >100. And genetically there’s lots of diabetes among Latinos. As such, the label “Pre-Diabetes” seems frankly meaningless to me. I simply advise people that it’ll be healthiest for them to avoid lots of sugars & carbohydrates, & maintain as normal a weight as possible.
Click for a story about a Horrible Practice one endocrinologist used to diagnose Diabetes.
The CMP also measures “Proteins” — maybe just the total Protein & Albumin, or maybe the Globulins as well. If the latter isn’t included, you have to calculate it by subtracting:
Globulin = Total protein MINUS albumin. Normal Globulin is ≤4.0 g/dL.
This is important, because Globulins are elevated in many systemic diseases, including Multiple Myeloma, paraproteinemias (amyloidosis, etc.), and all sorts of chronic infectious / inflammatory conditions. We’ve diagnosed 2 cases of asymptomatic HIV infection by simply noting high globulins on a lab slip. Aside from screening for HIV & Hep C, however, I wouldn’t embark on wide-scale testing (a “Grope-O-Gram”). Anyone over 50 with high globulins should get a Serum Protein Electrophoresis (SPEP) to rule out Multiple Myeloma. See below; see also our blurb Calculating Serum Globulins.
OTHER BLOOD TESTS (NOT ON THE “CMP”)
Serum Protein Electrophoresis (SPEP) — This blood test identifies a variety of unusual plasma cell disorders including multiple myeloma, amyloidosis, and Waldenström macroglobulinemia. The diseases are rare but the test is easy and cheap. An abnormal result consists of a “spike” on the electrophoretic gel due to a monoclonal protein (called an “M-Spike” in jargon).
If found, the protein must then be identified. That’s a lot more complex; refer monoclonal spikes to Hematology. However, I’ve never found one. Usually the lab result returns “Polyclonal elevations, no monoclonal spike.”
There’s NO NEED to order the SPEP whenever you find elevated serum globulins >4.0 g/dL that are otherwise explained. These include known connective tissue disorders (rheumatoid arthritis, SLE, etc. et.) and chronic infections like HIV and Hep C. However, the test is may be useful in the following contexts:
- Unexplained anemia, weakness, or fatigue
- Osteopenia, lytic lesions on bone x-ray, non-traumatic fractures
- Renal insufficiency with a normal urine sediment
- Heavy proteinuria in patients over 40
- Peripheral neuropathy of unknown etiology
- Recurrent infections
- Elevated ESR (sedimentation rate)
Creatine Kinase (CK) — Previously (& occasionally still) known as the “Creatine Phosphokinase” (CPK). It’s the standard test to evaluate muscle weakness, for conditions like polymyositis and dermatomyositis. One common use is to rule out myopathy due to statins, which can cause rhabdomyolysis, with renal failure and even death.
With rhabdo, CK elevations are enormous, up to 100-times upper limit of normal. Since exercise, black race, muscle strain, hypothyroidism, IM shots, recent surgery, and idiopathicity (if that’s a word) can cause minor rises, I ignore CK results unless they’re in the 1,000’s. One study found that large numbers of people may have “high” CKs according to printed laboratory reference “normals,”, compared to the conventional 2.5% as defined by Bell curve.
As such, it’s unnecessary to draw a “baseline CK” before prescribing a statin. It’s even more unnecessary to monitor CKs during therapy.
However, statins can cause a very low-grade myopathy even without a CK elevation. Patients complain of easy fatigability or bilateral muscle soreness, usually within 6 months of treatment, but even up to 2 years. It’s fine to draw a CK, but the only decent diagnosis is by discontinuation & rechallenge. Be sure that symptoms are bilateral & somewhat generalized, since one small study found patients might complain of aches and pains after reading package insert warnings.
As long as we now seem to be on statins (instead of the CK), I’ll mention our patient who developed frank rhabdo when a hospitalist discharged him with H. pylori therapy (including clarithromycin) without thinking over the med list. Always look up new meds when someone’s on a statin (actually, when someone’s on anything!).
Lactate Dehydrogenase (LDH) — The LDH is a useful clue in the following circumstances:
- Normocytic Anemia — if LDH elevated, suspect hemolysis
- Jaundice — if increased Indirect Bilirubin (i.e. high Total Bili with normal Direct Bili) and LDH elevated, suspect hemolysis
- LDH elevated in Lymphoma — might order it in patient with fevers & sweats.
- Among HIV patients with fever, a high LDH suggests a pulmonary focus (or lymphoma), whereas a very high LDH (e.g. >600) suggests fungal infection.
Uric Acid — There are only 2 contexts to order a serum Uric Acid: 1) to diagnose Gout; 2) if you’re dealing with kidney stones.
Gout can occur with a normal uric acid. In fact, if you want to see if it’s high, test at least 2 weeks after the acute attack. The gold standard for diagnosing gout is finding uric acid crystals from joint fluid (by a polarizing microscope). Arthrocentesis for this is commonly done from the knee, rarely from the ankle or big toe.
I had a patient who had an equivocal diagnosis; rheumatology said he should drop in at urgent care for a tap if it recurred. So he did, nobody on that day knew how to tap a big toe, so they sent him home. Next time he saw another Rheum fellow, they said the same thing!
Even worse, lots of people have high serum uric acids & will never get gout. So as you’ve undoubtedly guessed, the diagnosis is clinical. Sounds like gout, uric acid high, give allopurinol & keep increasing it until the level is under 6.0 mg/dL.
For kidney stones, you might make a diagnosis by finding uric acid crystals in a urinalysis. You can send a recovered stone for analysis, but unless the lab identifies the nidus (revealing the metabolic condition that started it all), it may not be helpful. Here, a serum uric acid is of minimal importance compared to a 24-hour urine value. If the latter is high, you’ve got your likely diagnosis. Give allopurinol… [see above paragraph].
Allopurinol is an easy drug, if you’re not one of the unfortunate few who get a rash with lethal hypersensitivity reaction. Asians are genetically at higher risk. Always give appropriate warnings whenever prescribing it. I’ve seen patients from underdeveloped countries who’d seen docs for any non-specific “joint pain” or “arthritis,” & get placed on the drug if a serum uric acid was high. NOT! Never treat hyperuricemia in & of itself, in a clinical situation not compatible with gout.
And that’s it for our discussion of laboratory tests. On to some other as-yet undetermined topic next time.