Pancreatic Enzymes: Lipase Wins, Amylase Is Overrated, and DKA Fakes Pancreatitis

 

Pancreatic Enzymes: Lipase Wins, Amylase Is Overrated, and DKA Fakes Pancreatitis

One enzyme is enough for diagnosis. The other one misleads you. And the magnitude doesn't predict severity.

Lipase vs. Amylase: Lipase Wins

Lipase is the preferred test for diagnosing acute pancreatitis. It's more sensitive (82–100%) and more specific (82–99%) than amylase. It rises within 4–8 hours, peaks at 24 hours, and remains elevated for 8–14 days (longer than amylase). Current ACG guidelines recommend lipase as the sole enzymatic criterion—amylase adds nothing and is no longer recommended as a standalone test for pancreatitis.

Acute pancreatitis is diagnosed when 2 of 3 criteria are met: (1) characteristic abdominal pain, (2) lipase ≥3× upper limit of normal, (3) characteristic findings on imaging. Most cases can be diagnosed without imaging if pain + lipase are both positive.

The Pitfalls

1. Amylase Has Too Many Non-Pancreatic Sources

Amylase is produced by the salivary glands, fallopian tubes, lungs, and small intestine—not just the pancreas. Elevated amylase can be caused by: parotitis (mumps), bowel obstruction, perforated ulcer, ectopic pregnancy, renal failure (decreased clearance), macroamylasemia (a benign condition where amylase binds to immunoglobulins and stays elevated chronically). If you only order amylase, you'll over-diagnose pancreatitis.

2. DKA Causes Elevated Lipase Without True Pancreatitis

The DKA Trap

Up to 16–25% of DKA patients have lipase >3× ULN without radiographic pancreatitis. The mechanism is likely metabolic (ketoacidosis + dehydration causing pancreatic ischemia) rather than true inflammatory pancreatitis. Don't diagnose pancreatitis in DKA based on lipase alone. Correlate with imaging and clinical presentation. If the abdominal pain resolves with DKA treatment, it wasn't pancreatitis.

3. Magnitude Doesn't Predict Severity

A lipase of 5,000 doesn't mean worse pancreatitis than a lipase of 500. The degree of elevation does NOT correlate with disease severity, complications, or prognosis. Severity is assessed by clinical criteria (BISAP score, Ranson's criteria, CT severity index), not enzyme levels.

4. Chronic Pancreatitis May Have Normal Lipase

In advanced chronic pancreatitis, the gland is so fibrosed and atrophied that it can no longer produce enough enzyme to elevate the lipase during acute exacerbations. A "burned-out pancreas" will have normal lipase even during pain flares. Diagnose chronic pancreatitis by imaging (calcifications on CT, pancreatic duct changes on MRCP).

5. Renal Failure Elevates Both Enzymes

Both lipase and amylase are renally cleared. CKD/ESRD patients commonly have chronically elevated levels (typically 2–3× ULN) without pancreatitis. Use clinical presentation and imaging, not enzyme levels alone, to diagnose pancreatitis in renal patients.

6. Macroamylasemia

A benign condition where amylase forms complexes with immunoglobulins that are too large for renal clearance, causing chronically elevated serum amylase. Lipase is normal. Diagnose by checking amylase-creatinine clearance ratio (low in macroamylasemia) or by serum amylase electrophoresis. This is one more reason lipase is superior—macroamylasemia doesn't affect lipase.

Bottom Line

Order lipase, not amylase. Don't diagnose pancreatitis in DKA based on lipase alone. The magnitude doesn't predict severity. Chronic pancreatitis can have normal enzymes. And CKD patients will have chronically elevated baseline levels. Acute pancreatitis is a clinical diagnosis supported by lipase, not a lipase diagnosis supported by pain.

Stay sharp out there.

Procalcitonin: When It Helps, When It Doesn't, and When It Fools You

 

Procalcitonin: When It Helps, When It Doesn't, and When It Fools You

The antibiotic stewardship tool that's powerful in the right context and useless in the wrong one.

Procalcitonin (PCT) is a peptide precursor of calcitonin that rises specifically in response to bacterial infection and remains low in viral infections and most autoimmune flares. It's gained traction as an antibiotic stewardship tool, but it's only useful in specific clinical scenarios—and ordering it indiscriminately creates more confusion than clarity.

How It Works

In health, PCT is produced only by thyroid C cells and is undetectable (<0.05 ng/mL). During bacterial infection, virtually every tissue in the body begins producing PCT in response to bacterial endotoxins and pro-inflammatory cytokines. Viral infections do NOT trigger this response (interferon-gamma actually suppresses PCT production), creating the bacterial-vs-viral distinction that makes PCT clinically useful.

Where PCT Changes Management (Evidence-Based)

• Lower respiratory tract infections: The strongest evidence. PCT-guided algorithms reduce antibiotic use in community-acquired pneumonia, acute bronchitis, and COPD exacerbations without increasing adverse outcomes. PCT <0.25 ng/mL strongly argues against bacterial pneumonia.
• Sepsis: PCT >0.5 supports bacterial sepsis. Serial PCT trending downward guides antibiotic de-escalation and discontinuation. Many ICU protocols use PCT to shorten antibiotic courses by 2–3 days.
• Neonatal sepsis: PCT at birth helps guide antibiotic duration in suspected early-onset neonatal sepsis (rises within 6–12 hours of infection).

Where PCT Does NOT Help

• UTI: PCT is unreliable for diagnosing UTI; localized infections don't always raise systemic PCT. Exception: pyelonephritis with systemic involvement may elevate PCT.
• Skin and soft tissue infections: Cellulitis, abscesses—PCT doesn't add to clinical assessment.
• Immunocompromised patients: Neutropenic fever, transplant recipients—PCT may be blunted or unreliable.
• Post-surgical patients: PCT rises after major surgery (especially cardiac, abdominal) as a nonspecific inflammatory response, limiting its specificity.

The Pitfalls

False Elevations (Non-Bacterial Causes)

• Major surgery or trauma (first 24–48 hours)
• Medullary thyroid carcinoma (C cells produce PCT constitutively)
• Severe burns, heat stroke
• Cardiogenic shock (without infection)
• Some autoimmune conditions: Kawasaki disease, anti-MDA5 dermatomyositis with severe inflammation, and Adult-onset Still's disease can elevate PCT
• Neonates: physiologic rise in the first 24–48 hours of life (age-specific norms required)

The Autoimmune Confounder

Most autoimmune flares (SLE, RA, vasculitis) do NOT raise PCT significantly, which makes it useful for distinguishing flare from infection in these patients. However, there are exceptions: severe systemic inflammation from macrophage activation syndrome (MAS), Kawasaki disease, and some inflammatory myopathies can elevate PCT without bacterial infection. Use it as one piece of the puzzle, not a standalone rule-out.

Interpretation Thresholds

PCT Level	Interpretation
<0.1 ng/mL	Bacterial infection very unlikely. Consider withholding/stopping antibiotics.
0.1–0.25	Bacterial infection unlikely. Antibiotics generally not recommended.
0.25–0.5	Possible bacterial infection. Consider antibiotics based on clinical context.
>0.5	Bacterial infection likely. Initiate/continue antibiotics.
>2.0	High likelihood of severe bacterial infection/sepsis.
>10	Severe sepsis/septic shock. Very high mortality risk.

Bottom Line

PCT is a powerful antibiotic stewardship tool when used in the right context: lower respiratory tract infections and sepsis. It's not useful for UTI, cellulitis, or post-surgical fever. It can help distinguish autoimmune flare from infection in your rheumatic patients (with exceptions). And the trend matters more than the single value—a PCT that's falling supports de-escalation; one that's rising demands investigation.

Stay sharp out there.

Hemoglobin Electrophoresis & Sickle Cell Testing: Interpreting the Patterns

 

Hemoglobin Electrophoresis & Sickle Cell Testing: Interpreting the Patterns

Trait vs. disease, thalassemia clues on the CBC, and why the newborn screen needs follow-up.

Hemoglobin electrophoresis separates hemoglobin variants by their electrical charge, identifying the type and proportion of each hemoglobin present. It's the definitive test for diagnosing hemoglobinopathies—but it's also one of the most misunderstood results in primary care.

Normal Hemoglobin Composition

• HbA (adult hemoglobin): 95–98% (two alpha + two beta globin chains)
• HbA2: 2–3.5% (two alpha + two delta chains)
• HbF (fetal hemoglobin): <1% in adults (two alpha + two gamma chains; predominates in utero and declines after birth)

The Key Patterns

Condition	HbA	HbS	HbA2	HbF	CBC Clues
Normal	95–98%	0	2–3.5%	<1%	Normal
Sickle cell trait (AS)	55–60%	35–45%	Normal	Normal	Usually normal CBC. No anemia. Benign in most contexts.
Sickle cell disease (SS)	0	80–95%	Normal	2–20%	Chronic hemolytic anemia, reticulocytosis, sickle cells on smear
HbSC disease	0	~45% S, ~45% C	Normal	Low	Milder anemia than SS; target cells on smear; still has vaso-occlusive crises
Beta-thalassemia trait	90–95%	0	>3.5%	Normal/slightly ↑	Microcytic anemia with normal RDW and elevated RBC count. Mentzer index (MCV/RBC) <13 favors thalassemia.
Beta-thalassemia major	0–10%	0	Variable	60–90%	Severe transfusion-dependent anemia from infancy
Alpha-thalassemia trait (2-gene deletion)	Normal pattern	0	Normal	Normal	Microcytic anemia with normal electrophoresis. Diagnosis by exclusion or genetic testing.
HbH disease (3-gene alpha deletion)	Reduced	0	Reduced	Normal	HbH (beta-4 tetramers) visible on special stain or HPLC. Moderate hemolytic anemia.

The Alpha-Thalassemia Trap

Alpha-thalassemia trait (1–2 gene deletions) has a completely normal hemoglobin electrophoresis. The only clues are microcytosis with normal iron studies and normal HbA2. If you've ruled out iron deficiency and the electrophoresis is normal, alpha-thalassemia trait is the diagnosis by exclusion (confirm with genetic testing if needed). This is extremely common in African American, Southeast Asian, and Mediterranean populations.

Thalassemia Trait vs. Iron Deficiency: The CBC Distinction

Both cause microcytic anemia. The CBC pattern distinguishes them:

• Iron deficiency: Low MCV, high RDW (unequal cells), low/normal RBC count, low ferritin, high TIBC
• Thalassemia trait: Low MCV, normal RDW (uniformly small cells), elevated RBC count, normal iron studies, elevated HbA2 (beta-thal) or normal electrophoresis (alpha-thal)

Newborn Screening Follow-Up

Pediatric Critical Point

All US states screen for hemoglobinopathies on the newborn metabolic panel. Results are reported as the hemoglobins detected in order of quantity: FA = normal (Fetal predominant, Adult present). FAS = sickle cell trait. FS = possible sickle cell disease (no HbA detected—needs urgent confirmatory testing). FSC = HbSC disease. Any abnormal newborn screen requires confirmatory hemoglobin electrophoresis by 2–3 months of age and genetic counseling. Early identification of sickle cell disease enables penicillin prophylaxis starting at 2 months, which dramatically reduces mortality from pneumococcal sepsis.

Connections to This Series

• HbA1c post: Hemoglobinopathies make A1c unreliable. HbS, HbC, HbE interfere with assays.
• CBC post: MCV + RDW + RBC count pattern distinguishes thalassemia from iron deficiency.
• Iron studies post: Always check iron studies before diagnosing thalassemia. Coexisting iron deficiency can mask thalassemia, and iron overload from transfusions is a major complication.

Bottom Line

Hemoglobin electrophoresis diagnoses sickle cell disease, beta-thalassemia, and hemoglobin variants definitively. Alpha-thalassemia trait is the exception—it has a normal electrophoresis and is diagnosed by exclusion. Use the CBC pattern (MCV, RDW, RBC count) to guide when to order electrophoresis, and always follow up abnormal newborn screens promptly.

Stay sharp out there.

Fecal Calprotectin & Hyperinflammation Markers: IBD vs. IBS, and When Ferritin >10,000 Is a Diagnosis

 

Fecal Calprotectin & Hyperinflammation Markers: IBD vs. IBS, and When Ferritin >10,000 Is a Diagnosis

The stool test that saves your patient a colonoscopy and the serum ferritin level that means macrophage activation, not iron overload.

Part 1: Fecal Calprotectin—IBD vs. IBS

Fecal calprotectin is a neutrophil-derived protein released into the gut lumen during intestinal inflammation. It's the single best non-invasive test for distinguishing inflammatory bowel disease (IBD) from irritable bowel syndrome (IBS)—a distinction that saves patients from unnecessary colonoscopies and saves the healthcare system significant cost.

How to Use It

• Calprotectin <50 µg/g: IBD very unlikely. Supports IBS or functional GI disorder. Negative predictive value >95%.
• 50–150: Borderline. Repeat in 4–6 weeks. If persistently elevated, refer for GI evaluation.
• >150: Intestinal inflammation likely. Correlate with clinical picture. Refer for colonoscopy/endoscopy to differentiate Crohn's, UC, or other causes.
• >250: Strongly suggestive of active IBD.

Also useful for monitoring IBD activity: rising calprotectin in a known IBD patient may predict relapse before symptoms appear, enabling preemptive treatment adjustment.

Pitfalls

• NSAIDs cause intestinal inflammation and elevate calprotectin—stop NSAIDs for 2 weeks before testing.
• PPIs may mildly elevate calprotectin in some patients.
• GI infections (bacterial gastroenteritis) elevate calprotectin—test after the infection resolves.
• Colorectal cancer can elevate calprotectin—an elevated result in an older patient still warrants colonoscopy.
• Children <4 years have physiologically higher calprotectin levels; use age-adjusted thresholds.
• Celiac disease can mildly elevate calprotectin.

Part 2: Hyperinflammation Markers—When Ferritin Tells a Different Story

Ferritin as a Hyperinflammation Marker

In the iron studies post, we discussed ferritin as an acute-phase reactant that masks iron deficiency. But extremely elevated ferritin (>1,000–10,000+) is a different diagnostic entity—it signals macrophage activation and hyperinflammation, not iron overload.

Ferritin >10,000 = Think MAS or HLH

Hemophagocytic lymphohistiocytosis (HLH) and macrophage activation syndrome (MAS) are life-threatening hyperinflammatory conditions where macrophages go into overdrive, phagocytosing blood cells and releasing massive cytokines. Ferritin >10,000 ng/mL is a classic diagnostic clue (sensitivity ~90% for HLH in children). Other features: high fevers, cytopenias, hepatosplenomegaly, elevated triglycerides, low fibrinogen, elevated soluble IL-2 receptor. MAS is the rheumatologic variant, most commonly complicating systemic JIA (children) or Adult-onset Still's disease.

Adult-Onset Still's Disease

Ferritin >1,000 (often >5,000–10,000+) in a young adult with quotidian (daily spiking) fevers, evanescent salmon-colored rash, arthritis, sore throat, and leukocytosis is Adult-onset Still's disease until proven otherwise. A unique feature: the glycosylated ferritin fraction drops to <20% (normal >50%), which helps confirm the diagnosis. This specialized test isn't available everywhere but is highly specific when obtainable.

Other Hyperinflammation Markers

• Soluble IL-2 receptor (sIL-2R/sCD25): Markedly elevated in HLH/MAS. Included in the HLH-2004 diagnostic criteria.
• Lactate dehydrogenase (LDH): Elevated in hemolysis, tissue destruction, and hyperinflammatory states. Non-specific but part of the pattern.
• Triglycerides: Paradoxically elevated in HLH/MAS (cytokine-mediated lipolysis suppression).
• Fibrinogen: Low in HLH/MAS (consumed by DIC-like process), which is the opposite of what you'd expect in inflammation (fibrinogen is normally an acute-phase reactant that rises).

Bottom Line

Fecal calprotectin is your primary care tool for distinguishing IBD from IBS non-invasively—a negative result (<50) has >95% NPV for ruling out IBD. Stop NSAIDs before testing. On the hyperinflammation side, learn to recognize the pattern: ferritin >10,000 + cytopenias + fevers + hepatosplenomegaly = HLH/MAS until proven otherwise. These are rare but lethal conditions where early recognition saves lives.

Stay sharp out there.

Uric Acid: The Gout Pitfall, the Asymptomatic Hyperuricemia Debate, and Tumor Lysis

 

Uric Acid: The Gout Pitfall, the Asymptomatic Hyperuricemia Debate, and Tumor Lysis

Uric acid is often NORMAL during a gout flare. Asymptomatic hyperuricemia almost never needs treatment. And yet both mistakes happen daily.

Uric acid is a deceptively simple lab that generates outsized confusion. The two most common errors: using it to rule out gout during an acute flare (when it's often normal) and treating asymptomatic hyperuricemia (which guidelines recommend against).

The #1 Pitfall: Normal Uric Acid During Acute Gout

During an acute gout flare, serum uric acid drops in up to 40% of patients due to the inflammatory response (IL-6 increases renal uric acid excretion) and redistribution of urate into the inflammatory joint. A normal serum uric acid does NOT rule out gout. The gold standard for acute gout diagnosis is synovial fluid analysis showing negatively birefringent monosodium urate crystals under polarized light. If you can't aspirate the joint, diagnose gout clinically (monoarticular, first MTP involvement, rapid onset, prior attacks, response to colchicine) and check uric acid 2–4 weeks after the flare resolves for a more accurate baseline.

Asymptomatic Hyperuricemia: Don't Treat It

Elevated uric acid (>6.8 mg/dL is the saturation point) without gout, tophi, or uric acid nephrolithiasis is asymptomatic hyperuricemia. Current ACR/EULAR guidelines recommend against urate-lowering therapy for asymptomatic hyperuricemia. There is no proven benefit in preventing gout, CKD progression, or cardiovascular events in this population. Most hyperuricemic patients never develop gout. Address modifiable factors: reduce alcohol (especially beer), limit purine-rich foods, stop or reduce thiazides/loop diuretics if possible, manage metabolic syndrome.

When Uric Acid Is Useful

• Establishing baseline for gout management: Check 2–4 weeks after a flare. Target <6 mg/dL (some guidelines say <5 for tophaceous gout) on urate-lowering therapy (allopurinol, febuxostat).
• Monitoring urate-lowering therapy: Titrate allopurinol to target uric acid, not to a fixed dose.
• Tumor lysis syndrome (TLS): Uric acid rises rapidly as tumor cells lyse (after chemotherapy for high-tumor-burden malignancies: ALL, Burkitt's, high-grade lymphomas). Elevated uric acid + hyperkalemia + hyperphosphatemia + hypocalcemia = TLS. Prevent with aggressive IV hydration, rasburicase, or allopurinol.
• Preeclampsia: Elevated uric acid is associated with preeclampsia severity, though it's a marker, not a diagnostic criterion.
• CKD monitoring: Uric acid rises as GFR declines. Whether treating hyperuricemia slows CKD progression remains debated—current evidence does not support routine treatment.

Medication Pitfalls

Thiazides and loop diuretics raise uric acid (decreased renal excretion). Losartan and fenofibrate lower it (uricosuric effects). Low-dose aspirin raises uric acid; high-dose aspirin lowers it. Consider these medication effects before attributing hyperuricemia to diet or genetics.

Bottom Line

Don't use uric acid to diagnose or exclude acute gout—it's unreliable during flares. Don't treat asymptomatic hyperuricemia. Do use it to monitor urate-lowering therapy (target <6) and watch for it in tumor lysis syndrome. And always check the medication list before blaming the patient's diet.

Stay sharp out there.

Lipid Panel: Beyond Total Cholesterol—Non-HDL, ApoB, Lp(a), and Secondary Causes

LDL-C is calculated, not measured. Non-HDL may be better. And nobody's checking Lp(a) yet even though they should be.

The lipid panel is ordered on nearly every adult in primary care, yet most providers look at the LDL-C, compare it to a target, and stop. The panel actually contains much more information—and the advanced markers that are increasingly recommended (non-HDL, apoB, Lp(a)) are the ones that predict cardiovascular events better than LDL alone.

The Standard Panel

• Total cholesterol: LDL + HDL + VLDL. Largely supplanted by its components for clinical decisions.
• LDL-C: Usually calculated by the Friedewald equation (TC − HDL − TG/5). Inaccurate when triglycerides >400 (equation breaks down) or when LDL is very low. Direct LDL measurement is available but not standard.
• HDL-C: Higher is generally protective. Very low HDL (<40 men, <50 women) increases cardiovascular risk. Pharmacologically raising HDL has NOT been shown to reduce events.
• Triglycerides: Fasting preferred for accuracy but non-fasting is acceptable for screening. >500 mg/dL = pancreatitis risk requiring treatment regardless of cardiovascular risk.
• Non-HDL cholesterol = Total cholesterol − HDL. Captures ALL atherogenic lipoproteins (LDL + VLDL + IDL + Lp(a)). Better predictor of cardiovascular risk than LDL-C alone, especially in patients with elevated triglycerides where calculated LDL is unreliable. Target: typically LDL goal + 30.

Advanced Markers: When to Go Beyond the Standard Panel

ApoB (Apolipoprotein B)

Each atherogenic lipoprotein particle (LDL, VLDL, IDL, Lp(a)) contains exactly one apoB molecule. ApoB therefore counts the total number of atherogenic particles, making it the single best measure of atherogenic burden. It's superior to LDL-C in patients with metabolic syndrome, diabetes, obesity, or high triglycerides where LDL-C may underestimate risk (discordance between particle number and cholesterol content). Consider ordering when LDL-C is at goal but the patient has metabolic syndrome or premature ASCVD.

Lp(a) (Lipoprotein(a))

The most underordered cardiovascular risk marker in medicine. Lp(a) is a genetically determined, highly atherogenic lipoprotein that is NOT lowered by statins, diet, or exercise. Elevated Lp(a) (>50 mg/dL or >125 nmol/L) independently increases ASCVD risk 2–3 fold and aortic stenosis risk. Current guidelines (2024 ESC, AHA/ACC consideration) recommend checking Lp(a) at least once in every adult's lifetime since it's genetically fixed. If elevated, it changes risk stratification (may warrant more aggressive LDL lowering) and will have targeted therapies (antisense oligonucleotides) entering the market.

When to Order Lp(a)

Check at least once in all adults. Especially important in: premature ASCVD (men <55, women <65), family history of premature ASCVD, recurrent events despite optimal statin therapy, borderline risk where Lp(a) could change the treatment decision, and calcific aortic stenosis.

Fasting vs. Non-Fasting

Current guidelines (2018 AHA/ACC, 2024 ESC) accept non-fasting lipid panels for initial screening. Triglycerides are the main value affected by fasting (can rise 20–30% postprandially). A fasting panel is needed when: triglycerides are >400 on non-fasting, you need accurate calculated LDL-C, or you're monitoring response to triglyceride-lowering therapy.

Secondary Causes of Dyslipidemia

Before attributing dyslipidemia to "genetics" or "diet," rule out secondary causes:

Lipid Abnormality	Secondary Causes to Consider
Elevated LDL	Hypothyroidism (check TSH!), nephrotic syndrome, obstructive liver disease, anorexia, medications (thiazides, cyclosporine, retinoids)
Elevated triglycerides	Uncontrolled diabetes/insulin resistance, alcohol, obesity, CKD, medications (steroids, beta-blockers, estrogen, retinoids, atypical antipsychotics, protease inhibitors)
Low HDL	Metabolic syndrome, smoking, sedentary lifestyle, anabolic steroids, beta-blockers, progestins
Severely elevated TG (>500)	Familial hypertriglyceridemia, uncontrolled diabetes, alcohol, medications, pregnancy

Don't Miss Hypothyroidism

Hypothyroidism is one of the most common reversible causes of elevated LDL. Always check TSH before starting a statin for new dyslipidemia. Correcting the thyroid may normalize the lipids without medication.

Pediatric Lipid Screening

• Universal screening: Non-fasting lipid panel once between ages 9–11 and again at 17–21 (NHLBI guidelines).
• Targeted screening: Children ≥2 years with family history of premature ASCVD, dyslipidemia, or risk factors (obesity, diabetes, hypertension).
• Familial hypercholesterolemia should be suspected when LDL >160 in a child or >190 in an adult. Cascade screening of family members is indicated.

Bottom Line

The lipid panel is more than LDL-C. Non-HDL is free (calculate from the standard panel) and better than LDL-C in metabolic syndrome. ApoB counts atherogenic particles directly. Lp(a) should be checked once in every adult. Always rule out secondary causes before starting a statin. And check that TSH.

Stay sharp out there.

Thyroid Function Tests: TSH, Free T4, Free T3—and When You Actually Need Each One

 

Thyroid Function Tests: TSH, Free T4, Free T3—and When You Actually Need Each One

TSH is almost always enough. Free T3 is almost never needed. And sick euthyroid syndrome fools everyone.

This post complements the thyroid antibody workup from earlier in the series. That post covered which antibody to order when you suspect autoimmune thyroid disease. This one covers the function tests—TSH, free T4, and free T3—and the interpretation patterns that trip up clinicians daily.

The Hierarchy: TSH First, Always

TSH is the single best screening test for thyroid dysfunction in outpatient primary care. The pituitary amplifies small changes in thyroid hormone levels, making TSH far more sensitive than direct hormone measurement. A normal TSH essentially rules out primary thyroid disease in an outpatient setting.

TSH	Free T4	Free T3	Interpretation
↑ High	↓ Low	—	Overt hypothyroidism. Treat with levothyroxine.
↑ High	Normal	—	Subclinical hypothyroidism. Repeat in 6–12 weeks to confirm. Treat if TSH >10, or if symptomatic + anti-TPO positive, or if pregnant/trying to conceive.
↓ Low	↑ High	—	Overt hyperthyroidism. Order TSI to confirm Graves'. Consider radioactive iodine uptake if diagnosis unclear.
↓ Low	Normal	↑ High	T3 thyrotoxicosis. One of the few times free T3 is needed. Seen in early Graves' or toxic nodule.
↓ Low	Normal	Normal	Subclinical hyperthyroidism. Repeat to confirm. Consider treatment if TSH <0.1, age >65, atrial fibrillation, or osteoporosis risk.
Normal or Low	↓ Low	—	Central hypothyroidism (pituitary or hypothalamic). TSH is inappropriately normal/low despite low T4. Rare but missed if you only check TSH.
Variable	Variable	—	Sick euthyroid syndrome (nonthyroidal illness). See below.

When to Order Free T3 (Almost Never)

Free T3 is overordered in primary care. The only validated indications:

• Suppressed TSH + normal free T4: Check free T3 to detect T3 thyrotoxicosis
• Monitoring amiodarone-induced thyroid disease (complex T3/T4 dynamics)
• Suspected T3 thyrotoxicosis (early Graves', toxic adenoma)

Free T3 should NOT be ordered for: hypothyroidism monitoring (it's the last to fall and the most variable), fatigue workup, or routine screening. Treat hypothyroidism by normalizing TSH, not by chasing T3 levels.

Sick Euthyroid Syndrome (Nonthyroidal Illness)

Acutely ill hospitalized patients frequently have abnormal thyroid function tests despite having no thyroid disease. The classic pattern: low T3, low or normal T4, low/normal/slightly elevated TSH. This is an adaptive response to illness, not hypothyroidism.

The Rule

Do not check thyroid function tests in acutely ill hospitalized patients unless you have a specific clinical reason to suspect thyroid disease (myxedema coma, thyroid storm, new atrial fibrillation). Abnormal results in acute illness are almost always sick euthyroid and do not warrant treatment. Recheck 6–8 weeks after recovery if needed.

Pregnancy: Trimester-Specific TSH Ranges

TSH normally drops in the first trimester due to HCG-mediated thyroid stimulation. Using non-pregnant reference ranges will over-diagnose hypothyroidism in early pregnancy:

• First trimester: Upper limit of TSH ~4.0 mIU/L (or ideally population-based trimester-specific ranges from your lab, typically 0.1–2.5)
• Second/third trimester: TSH gradually returns toward non-pregnant range
• Gestational thyrotoxicosis: Suppressed TSH with elevated free T4 in the first trimester, driven by high HCG (hyperemesis gravidarum). TSI-negative. Self-limited. Do NOT treat with antithyroid drugs—it's not Graves'.
• Overt hypothyroidism in pregnancy should be treated aggressively (target TSH <2.5 in first trimester). Levothyroxine dose typically increases 25–50% in pregnancy.

The Pitfalls

• Checking only TSH and missing central hypothyroidism: If a patient has symptoms of hypothyroidism + a pituitary history (surgery, radiation, tumor, postpartum hemorrhage/Sheehan's), check free T4 even if TSH is "normal."
• Biotin interference (from the antibody post): Biotin causes falsely low TSH and falsely high free T4, mimicking hyperthyroidism. Stop biotin supplements 48–72 hours before testing.
• Checking TFTs too soon after dose changes: TSH takes 6–8 weeks to re-equilibrate after a levothyroxine dose change. Don't recheck sooner.
• Ordering total T4 instead of free T4: Total T4 is affected by binding protein changes (pregnancy, estrogen, liver disease). Free T4 is the clinically relevant measurement.
• Over-checking in stable patients: Once hypothyroid patients are stable on levothyroxine, annual TSH is sufficient. More frequent testing creates dose-chasing.
• Treating subclinical hypothyroidism reflexively: TSH 5–10 with normal T4 doesn't always need treatment. Consider symptoms, anti-TPO status, age, cardiovascular risk, and fertility plans.

Pediatric Note

Congenital hypothyroidism is screened on the newborn metabolic panel. A high TSH on newborn screen requires urgent confirmatory testing and treatment—delayed treatment causes irreversible intellectual disability. In children, acquired hypothyroidism (usually Hashimoto's) should be suspected with growth deceleration, delayed puberty, fatigue, and constipation. TSH is the first-line test at any age.

Bottom Line

TSH first. Free T4 if TSH is abnormal. Free T3 only for suppressed TSH with normal T4. Don't test acutely ill patients unless you suspect a thyroid emergency. Use trimester-specific ranges in pregnancy. And stop biotin before testing.

Stay sharp out there.