13 minute read
Read the panel like a chart, not a list
The morning panel rewards a fixed order and punishes browsing. Three habits carry most of the value. First, trend before threshold: pull yesterday's and admission's values, because a creatinine of 130 falling from 180 and a creatinine of 130 rising from 90 are opposite diseases. Second, pretest probability: a mildly abnormal value in a well patient on no relevant drugs is usually noise, and chasing it is how cascades start. Third, know your units and your lab's reference intervals; a 'flag' is the lab's population statement, not a diagnosis.
Sodium is a water problem
Serum sodium is a concentration, so abnormal sodium is almost always abnormal water handling, which means ADH. The hyponatremia sequence is fixed: serum osmolality (is it truly hypotonic?), urine osmolality (is ADH acting?), then volume status with urine sodium (does ADH have a hemodynamic excuse, or is it inappropriate?). Speed of correction is a treatment decision in itself: a chronically low sodium corrected fast is how osmotic demyelination happens, so slow disease gets slow correction.
Potassium: treat the ECG, not the number
Potassium's danger is electrical, so the ECG outranks the assay. Hyperkalemia with peaked T waves or a widening QRS gets calcium first, shifting therapy second, and elimination third, before the repeat sample is back. On the low side, remember the pack it travels in: hypokalemia will not correct while magnesium is low, and both fall together with diuretics, diarrhea and refeeding. A surprising potassium in an asymptomatic patient with a normal ECG is a hemolyzed sample often enough to justify a repeat before treatment.
Creatinine and urea tell two stories
Creatinine is the kidney's speedometer, read as a slope: AKI is defined by the delta (a rise of 26 in 48 hours, or 1.5 times baseline), not by any absolute number. Interpret it against muscle mass; a frail 45 kg patient's 'normal' creatinine can hide a halved GFR. Urea rises with the same renal failure but also with its own causes: a BUN-to-creatinine rise out of proportion suggests prerenal states or an upper GI bleed (digested blood is a protein meal). The first fork in every AKI is prerenal, renal, or postrenal, and a bladder scan answers the third in two minutes.
The CBC: sort anemia by size, then by effort
Anemia's differential collapses quickly under two numbers. MCV sorts by size: microcytic sends you to iron studies (ferritin low in deficiency, RBC count paradoxically high in thalassemia trait); macrocytic to B12, folate, alcohol, liver and thyroid; normocytic to the second number. Reticulocytes sort by effort: high means the marrow is responding and cells are being lost (bleeding, hemolysis: add LDH, haptoglobin, bilirubin, and a smear); low means the marrow itself is failing (CKD, chronic disease, infiltration). White cells and platelets get the same trend treatment: a falling platelet count on heparin, or a lone falling line among three, matters more than any single value.
Liver tests come in two patterns
ALT and AST report hepatocyte injury; ALP and GGT report cholestasis; bilirubin, albumin and INR report function, which is a different question from injury. Most abnormal panels declare themselves as one of two silhouettes: hepatocellular (aminotransferases many times normal, ALP barely up) or cholestatic (ALP and GGT leading, aminotransferases modest), and the silhouette chooses the next test: hepatitis serologies and toxins for the first, biliary imaging for the second. Two refinements earn their keep: an AST roughly double the ALT hints at alcohol, and aminotransferases in the thousands mean ischemia, toxin, or acute viral hepatitis, a short and urgent list.
When the number lies
Every panel has famous liars. Potassium rises in a hemolyzed or long-tourniquet sample. Total calcium reads low whenever albumin is low; correct it (add 0.02 mmol/L per g/L of albumin deficit) or measure ionized calcium before treating. Sodium reads falsely low through severe hyperglycemia (translocational) or extreme lipids and paraproteins (pseudohyponatremia). A drip-arm sample can make glucose and electrolytes absurd. The tell is always the same: a value that changed faster than physiology allows, in a patient who does not look like the number. Repeat before you react.
Pearls
- The delta is the diagnosis: no lab value means anything without its previous value.
- Hyperkalemia with ECG changes is treated before the confirmatory repeat comes back.
- MCV then reticulocytes sorts most anemia before any smear is seen.
- Bilirubin, albumin and INR measure function; enzymes only measure injury.
- A number that changed faster than physiology allows is a specimen problem until repeated.
Connected