Laboratory interpretation in the emergency department — the systematic high-yield read

The emergency department laboratory panel is a defined set of high-turnaround tests that drive time-critical decisions long before the formal work-up returns: lactate, troponin, lipase, beta-hCG, D-dimer, the coagulation studies with INR, the venous blood gas, C-reactive protein, and creatine kinase, with the full blood count, urea and electrolytes, and glucose woven through. The Fellowship skill is not memorising reference ranges — it is applying a structured read to every abnormal value: the range, the critical threshold, the clinical correlation, the true-versus-spurious question, and the action the result triggers. A lactate of 5 mmol per litre means one thing in a septic, mottled patient and another in a well marathon runner; a troponin of 60 ng per litre is acute ischaemia in one patient and chronic renal leak in another. This topic is the diagnostic-interpretation framework, not a disease, and it is built to answer any SAQ panel, any data-viva, and any SCQ on an abnormal number. [1]

When to use the framework and the principles of the systematic read

The high-yield panel is sent on any patient who is, or may become, unwell: the septic or shocked patient, chest pain, abdominal pain, the pregnant or possibly pregnant woman, the bleeding or anticoagulated patient, the reduced-conscious-state patient, and any deterioration. The results are read by the treating clinician at the bedside, acted on in parallel with resuscitation, and never deferred to a formal report. Three principles govern the read. First, treat the patient, not the number — a glucose of 2.9 mmol per litre in an asymptomatic, chronically controlled diabetic is observed; the same value in a seizing patient is treated immediately. Second, trend over single value — a lactate falling from 6 to 3 over two hours tells a story a single value cannot. Third, every surprising result gets a spurious screen before a dangerous intervention — a potassium of 7.1 mmol per litre in a calm, well patient with a normal ECG is haemolysed until proven otherwise.[16]

The six-step read for any abnormal value

Every abnormal result is run through the same six-step sequence so that none is acted on reflexively under cognitive load. [1]

Critical values — the numbers that trigger an immediate call

A critical value is one that carries an imminent risk to life or limb and that triggers a telephone call from the laboratory and an immediate clinician action. The list is examinable verbatim in the SCQ paper and is reproduced here in the ANZ units.[1]

Lactate — the perfusion and sepsis marker

Lactate is the single most prognostic value on the emergency panel. It is the end-product of anaerobic glycolysis and rises whenever oxygen delivery fails to meet demand (type A — shock, hypoxia, seizures, mesenteric ischaemia, exercise) or when aerobic metabolism is deranged (type B — metformin, malignancy, toxins); the type A causes dominate the emergency department.[3] A lactate at or above 2 mmol per litre is hyperlactataemia; at or above 4 mmol per litre it is severe and, in the septic patient, it reclassifies the case to septic shock or severe sepsis and triggers the resuscitation bundle.[1] The Surviving Sepsis Campaign guides resuscitation to reduce lactate to normal or by at least 10 per cent per hour, and lactate-guided therapy improves outcome in the critically ill.[2] Two pitfalls recur: the well patient with a single mildly raised lactate (often from a tourniquet or a struggling sample) is rechecked, not over-treated; and a rising lactate despite fluids in the abdominal-pain patient is mesenteric ischaemia or bowel ischaemia until disproven. Beta-blockers blunt the lactate response and can mask severity.

Troponin — the myocardial-injury marker

Cardiac troponin (I or T) is the preferred biomarker of myocardial injury. The high-sensitivity assays report in nanograms per litre and are interpreted against the 99th percentile upper reference limit of the local assay, with a rise or fall (the delta) defining an acute process.[4] A single high-sensitivity troponin below the 99th percentile at presentation, in a low-risk chest-pain patient with onset more than three hours earlier, safely rules out myocardial infarction; the validated 0-hour and 1-hour algorithms accelerate this further. The recurring trap is the chronically elevated troponin. Troponin rises in any strain on the heart, not only ischaemia — sepsis, pulmonary embolism, heart failure, renal failure, stroke, myocarditis, and tachyarrhythmia all elevate it. In chronic kidney disease a troponin of 60 ng per litre may be the patient's baseline; the delta decides. The examiner wants the candidate to distinguish acute myocardial infarction (a rise and fall with ischaemic evidence) from chronic myocardial injury (a persistently elevated troponin) — and to treat the underlying cause, not the number.

Lipase — the pancreatitis marker

Serum lipase is more sensitive and specific than amylase for acute pancreatitis and is the preferred test. The diagnostic threshold is lipase at least three times the upper limit of normal, in the right clinical setting (epigastric pain radiating through to the back, nausea, vomiting).[5] Lipase rises within hours, peaks in 24, and falls over days; a normal lipase at presentation in severe pain does not exclude pancreatitis and is repeated. Amylase is lower-yield — it is falsely normal in hypertriglyceridaemia and in alcohol-related pancreatitis, and it rises in non-pancreatic abdominal pathology (perforated viscus, ischaemic bowel, salivary disease). The degree of lipase elevation does not correlate with severity — the Atlanta classification grades severity on organ failure and local complications, not on the lipase number.[6] Lipase is also mildly raised in renal failure and in diabetic ketoacidosis; correlate clinically before committing to a pancreatitis admission.

Beta-hCG — the pregnancy test

A quantitative or urine beta-human chorionic gonadotropin is sent on every woman of reproductive age with abdominal pain, collapse, vaginal bleeding, or a missed period, and before any imaging or procedure that could harm a pregnancy. A urine beta-hCG becomes positive at around the time of the missed period (roughly 10 days after conception); a serum quantitative beta-hCG is more sensitive and is measured in international units per litre. The discriminatory zone — the beta-hCG level at which an intrauterine pregnancy should be visible on transvaginal ultrasound — sits at 1500 to 3500 IU per litre: a beta-hCG above this zone with an empty uterus on scan is an ectopic pregnancy until proven otherwise.[7] Two traps recur. The hook effect produces a falsely low beta-hCG in molar or multiple pregnancy at very high titres; dilution resolves it. And a negative urine beta-hCG does not fully exclude an early ectopic in the woman with a strong clinical picture — a quantitative serum level and early ultrasound are sent when the suspicion is high.

D-dimer — the fibrin marker and its pretest-probability gate

D-dimer is a breakdown product of cross-linked fibrin and rises with any thrombosis and any inflammation, haematoma, pregnancy, malignancy, or sepsis. It is therefore a rule-out, not a rule-in, test, and it is interpreted only through the lens of pretest probability. The standard cutoff is 500 micrograms per litre fibrin-equivalent units; in patients over 50 the age-adjusted cutoff (age multiplied by 10) increases specificity without losing sensitivity — a 72-year-old is ruled out below 720.[9] In the low-pretest-probability chest-pain or dyspnoea patient the PERC rule allows PE to be ruled out without any D-dimer at all.[10] In non-high-risk patients a D-dimer below cutoff, combined with a low Wells score, safely excludes PE and DVT; the Wells rule combined with D-dimer has a failure rate under 2 per cent.[8] The error is reflexive over-imaging: chasing a positive D-dimer in a high-probability patient who needs imaging regardless, or in a pregnant or post-surgical patient in whom it is almost always positive. A positive D-dimer never confirms VTE — it prompts imaging in the patient whose pretest probability warrants it.

Coagulation studies — INR and bleeding reversal

The INR standardises prothrombin-time reporting for warfarin monitoring, with a therapeutic range of 2.0 to 3.0 for most indications (2.5 to 3.5 for mechanical mitral valves). An INR above 5 with major bleeding is a critical value requiring urgent reversal. The first-line reversal agent for life-threatening warfarin-related bleeding is four-factor prothrombin complex concentrate (4F-PCC) 25 to 50 IU per kg intravenously plus vitamin K 5 to 10 mg intravenously, which corrects the INR within minutes, far faster and more completely than fresh frozen plasma.[11] Fresh frozen plasma is not recommended for routine warfarin reversal — it requires large volume, corrects slowly and incompletely, and is named on the ASH Choosing Wisely list of practices to question.[16] For the raised INR with no bleeding, oral vitamin K 2.5 to 5 mg reverses over 24 hours; 4F-PCC is reserved for major or life-threatening bleeding. The DOACs do not show reliably on the INR — a normal INR does not exclude apixaban or rivaroxaban effect, and a specific anti-Xa level or a calibrated assay is needed when reversal (idarucizumab for dabigatran, andexanet alfa for anti-Xa agents) is being considered. A spuriously high INR from an underfilled citrate tube is a common pre-analytical artefact — recheck on a correctly filled sample before reversing.

The venous blood gas — the parallel reads

The venous blood gas is faster, safer, and less painful than the arterial, and for most emergency questions it is sufficient. The venous pH is within 0.03 of the arterial; the venous bicarbonate and base excess are within 2; and the venous lactate and electrolytes are clinically equivalent to the arterial.[12] The VBG is therefore the pragmatic screening tool for the lactate, the potassium, the sodium, the glucose, and the acid-base status, drawn from any cannula at the moment of cannulation. It is not reliable for the PaO2 (meaningless in a venous sample) or for the precise PaCO2 in respiratory failure, where an arterial sample is taken to confirm hypercapnia before non-invasive ventilation. The VBG also carries the haemoglobin and haematocrit (roughly equivalent to the laboratory value in the non-hypoxaemic sample), which makes it the first read on the bleeding or the anaemic patient. A VBG with a glucose of 2.9 mmol per litre and a reduced conscious state is treated before the formal laboratory glucose returns.

C-reactive protein — the inflammation marker

C-reactive protein is a hepatic acute-phase reactant driven by interleukin-6; it rises within 6 to 12 hours of infection or tissue injury and peaks at 48 hours. It is a supporting marker, not a diagnostic one — a normal CRP does not exclude serious bacterial infection (it lags the clinical picture in the first hours, and is often normal in early meningococcal disease and in immunosuppression), and a raised CRP is non-specific (trauma, surgery, malignancy, autoimmune flare, and tissue infarction all raise it). The CRP trend over 24 to 48 hours is more useful than the single value, and it helps separate bacterial from viral infection and track the response to antibiotics.[13] Procalcitonin, also a supporting marker, rises faster and is more specific for bacterial infection and sepsis, and procalcitonin-guided algorithms safely shorten antibiotic courses in respiratory infection.[13] Neither CRP nor procalcitonin replaces clinical judgement — the decision to give antibiotics in the septic or the febrile neutropenic patient is made on clinical grounds before any marker returns.

Creatine kinase — the muscle-breakdown marker

Creatine kinase (CK) is released from damaged skeletal and cardiac muscle. A CK above 1000 units per litre, or at least five times the upper limit of normal, with the right clinical picture defines rhabdomyolysis; the common causes are exertional, trauma, prolonged immobilisation, statins and other drugs, seizures, the illicit-drug and alcohol-related collapse, and the infectious and inflammatory myositis.[14] The danger of rhabdomyolysis is acute kidney injury from pigment nephropathy, and the laboratory clues are a raised CK with an anion-gap metabolic acidosis, hyperkalaemia, hyperphosphataemia, hypocalcaemia, and a rising creatinine, with urine positive for blood on dipstick but with few red cells (myoglobin cross-reacts with the dipstick). A CK of 14 000 units per litre in the collapsed patient is rhabdomyolysis until proven otherwise, and early aggressive intravenous fluid is the cornerstone — the target is a urine output of 200 to 300 mL per hour until the CK falls. CK-MB and the troponin distinguish cardiac from skeletal muscle; the troponin is the cardiac marker of choice.

Differential diagnosis — true versus spurious, and the common abnormal patterns

A surprising abnormal value is sorted into true (a real finding that fits or redirects the clinical picture) or spurious (a pre-analytical artefact). The spurious screen is run before any dangerous intervention, because acting on a false number — reversing a normal patient's "INR of 9", or giving calcium for "hyperkalaemia of 7.5" in a calm, well patient — causes harm. The common spurious patterns and their true differentials are listed below. [1]

The common true abnormal patterns carry short lists of can't-miss causes that the candidate names unprompted. [1]

Management driven by the critical result

Several critical results immediately trigger a drug, and the doses are committed to memory because the unstable patient cannot wait for a lookup. The overriding principle is to treat the patient, not the number — and to recheck any surprising value before a dangerous intervention. [1]

Special populations

Pregnancy alters the normal ranges: D-dimer rises progressively through the trimesters and is rarely normal at term, so it has limited rule-out value; CRP is mildly raised; the beta-hCG is the must-do test in any woman of reproductive age with abdominal pain. Renal failure chronically elevates troponin (the delta decides) and biases the lactate in the metformin-treated patient; the urea and creatinine trend drives dialysis decisions. The elderly have an age-adjusted D-dimer (age multiplied by 10, over 50) and a troponin in which the delta matters more than the absolute; frailty lowers the transfusion threshold's tolerance for anaemia. The anticoagulated patient is interpreted through the INR (warfarin) or specific assays (DOACs), and a normal INR does not exclude a DOAC effect — the history of the last dose and the assay decide. [1]

Common errors and pitfalls

The recurring errors are well described. Acting on a single surprising value without a spurious screen — the haemolysed potassium of 7.5 in a calm, well patient is rechecked, not treated with calcium. Treating the number rather than the patient — transfusing a haemoglobin of 68 g per litre in the chronically anaemic, euvolaemic, asymptomatic patient. Chasing the D-dimer in a high-probability patient who needs imaging regardless, or in a pregnant patient in whom it is almost always positive. Declaring "myocardial infarction" on a single troponin in chronic kidney disease, where the delta and the ischaemic features decide. Over-reversing the INR of 9 with no bleeding — oral vitamin K, not 4F-PCC, is the answer. Misreading a normal lipase in hypertriglyceridaemic or alcoholic pancreatitis (lipase can be falsely low). Using fresh frozen plasma for warfarin reversal when 4F-PCC is available.[16] Forgetting that the venous gas gives the lactate, potassium, glucose, and haemoglobin within two minutes, and waiting instead for the formal panel.

Evidence and regional guidelines

The lactate-clearance recommendation traces to the Surviving Sepsis Campaign 2021 guidelines (Evans and colleagues) and to the Jansen multicentre randomised trial of lactate-guided therapy.[1][2] The pathophysiology of lactic acidosis is reviewed by Kraut and Madias.[3] Single high-sensitivity troponin rule-out is validated by Sandoval and colleagues.[4] The initial management of acute pancreatitis follows the American Gastroenterological Association guideline (Crockett and colleagues) and the revised Atlanta classification (Banks and colleagues).[5][6] Ectopic-pregnancy evaluation, including the discriminatory zone, is reviewed by Robertson and colleagues.[7] The Wells rule combined with D-dimer, the age-adjusted cutoff (ADJUST-PE), and the PERC rule are the standard PE rule-out evidence.[8][9][10] Four-factor prothrombin complex concentrate for warfarin reversal is reviewed by Quinlan and colleagues.[11] The venous-versus-arterial agreement for pH, bicarbonate, base excess, and lactate is established by Middleton and colleagues.[12] The role of CRP and procalcitonin in bacterial infection and antibiotic stewardship is summarised by Schuetz and colleagues.[13] The diagnosis and management of rhabdomyolysis is reviewed by Cervellin and colleagues.[14] The Wells clinical model for deep vein thrombosis is summarised in the rational clinical examination series.[15] The ASH Choosing Wisely campaign identifies fresh frozen plasma for warfarin reversal as a practice to question.[16]

Exam pearls

• Run the six-step read on every abnormal value: range, critical value, clinical correlation, true or spurious, action, trend. The method earns the marks, not the one-liner.
• Quote the critical values: lactate at or above 4, K above 6.5, glucose at or below 2.5, INR above 5, platelets below 20, Hb below 70 g/L, pH at or below 7.2. These are examinable verbatim.
• Troponin is interpreted against the 99th percentile with a delta. Chronic kidney disease and sepsis chronically elevate it; a flat troponin with no delta is chronic leak, not acute infarction.
• D-dimer is a rule-out test interpreted through pretest probability. Age-adjusted cutoff (age multiplied by 10) over 50; PERC rules out without any D-dimer in the very-low-risk patient.
• Lipase at least three times the upper limit confirms pancreatitis in the right clinical setting; the level does not correlate with severity.
• A positive beta-hCG with abdominal pain is ectopic until scanned. The discriminatory zone is 1500 to 3500 IU per litre; suspect the hook effect when the picture and the titre disagree.
• Reverse warfarin bleeding with 4F-PCC 25 to 50 IU per kg plus vitamin K 5 to 10 mg, not fresh frozen plasma. FFP is an ASH Choosing Wisely "don't."
• A surprising result is spurious until screened: haemolysis (high K), underfill (high INR), delay (high K, low glucose), hook effect (low hCG). Recheck before any dangerous intervention. [1]

Exam practice

Red flags