Triphasic waves on an EEG are one of the most recognizable yet frequently misinterpreted patterns in clinical electroencephalography. When ordered as part of an eeg test for a patient with altered mental status, these distinctive three-phase complexes can point clinicians toward hepatic encephalopathy, uremia, sepsis-associated encephalopathy, or even nonconvulsive status epilepticus. Understanding triphasic waves matters because the correct interpretation directly changes treatment decisions, from initiating lactulose for ammonia clearance to starting urgent antiseizure therapy in ambiguous cases.
The classic triphasic wave consists of three sequential phases: a small initial negative deflection, a dominant positive peak, and a slower terminal negative wave. They typically occur at 1.5 to 2.5 Hz, are bilateral, and show a frontocentral maximum with a measurable anterior-posterior lag. Most appear on a background of generalized slowing, and they often wax and wane with the patient's level of consciousness, fading during deeper coma and reappearing with mild stimulation or arousal.
Historically, triphasic waves were considered nearly pathognomonic for hepatic encephalopathy after Bickford and Butt described them in 1955. Modern data has shifted that view considerably. Today, neurologists recognize that any toxic-metabolic insult to the cortex can produce indistinguishable patterns, and a growing literature shows overlap with generalized periodic discharges seen in nonconvulsive seizures. That overlap is exactly why getting the read right requires careful clinical correlation, ammonia levels, renal panels, and sometimes a benzodiazepine trial.
For trainees and technologists, learning to identify triphasic waves quickly is a high-yield skill. They appear on roughly 5 to 10 percent of inpatient EEGs ordered for encephalopathy workup, and missing them or mislabeling them as epileptiform spikes leads to unnecessary antiseizure drug exposure. Knowing how long is an eeg test typically runs in this setting also helps families understand why a routine 20 to 30 minute study may need to extend into longer monitoring.
This guide walks through the morphology, the differential diagnosis, the bedside maneuvers that distinguish metabolic triphasics from epileptiform discharges, and the practical workflow for reporting. We will cover what causes triphasic waves, how to grade their reactivity, why the frontal predominance matters, and how the Salzburg criteria for nonconvulsive status epilepticus apply when the pattern is ambiguous. Real case examples and bedside tips are woven throughout.
By the end you will be able to recognize the wave morphology, list at least six causes beyond liver disease, run the standard benzodiazepine challenge, and write a clinically useful EEG report that helps the primary team treat the underlying condition rather than chase the wave itself. Whether you are a neurology resident, an EEG technologist, a hospitalist, or a curious patient researching your own results, this article gives you a complete working framework.
We will also address common patient questions about cost, preparation, side effects, and what an abnormal report actually means for prognosis. Triphasic waves alone are not a death sentence, but they do flag a sick brain that needs urgent attention to the underlying metabolic derangement, infection, or medication toxicity driving the pattern.
A small initial negative deflection is followed by a tall, dominant positive peak and then a longer, slower terminal negative wave. The positive phase typically has the highest amplitude, giving the wave its recognizable silhouette on standard bipolar montages.
Amplitude is maximal over Fz, F3, F4, and Cz electrodes, with progressive attenuation toward the occipital region. This anterior dominance is a useful localizing clue that separates triphasic waves from posterior dominant slow-wave activity.
A measurable 100–150 millisecond delay exists between frontal and posterior channels, visible on referential montages. This lag, sometimes called the AP gradient, supports a metabolic origin and is rarely seen with true epileptiform discharges.
Triphasic waves often increase in frequency or amplitude with arousal or noxious stimulation, then fade as the patient returns to baseline. Reactivity is generally preserved, which contrasts with the fixed, drug-resistant patterns of nonconvulsive status epilepticus.
The pattern is bilateral and largely symmetric across hemispheres, although mild asymmetry can occur with focal structural lesions overlaid on the metabolic background. Persistent strong asymmetry should prompt imaging to rule out a stroke or mass.
The differential diagnosis for triphasic waves on an eeg test is broader than most clinicians initially appreciate. Hepatic encephalopathy remains the prototypical cause, particularly in patients with cirrhosis, acute liver failure, or portosystemic shunts where ammonia and other gut-derived neurotoxins cross into the brain. In these patients, the appearance of triphasic waves often correlates with West Haven grade II or III encephalopathy and tends to resolve as ammonia clears with lactulose and rifaximin therapy.
Uremic encephalopathy is the second classic cause. Patients with acute kidney injury, end-stage renal disease before dialysis, or dialysis disequilibrium syndrome can develop identical triphasic patterns. The mechanism involves accumulation of guanidino compounds, parathyroid hormone, and other middle molecules that disrupt cortical neuronal function. Unlike hepatic cases, uremic triphasics often resolve dramatically within hours of effective dialysis, providing a useful diagnostic confirmation.
Sepsis-associated encephalopathy and other infection-related delirium states produce triphasic waves in roughly 15 to 20 percent of cases with severe systemic illness. Septic patients in the ICU with persistent altered mental status despite hemodynamic stabilization frequently show this pattern, and its presence portends worse outcomes including longer ICU stays and higher mortality. Cytokine-mediated cortical dysfunction and blood-brain barrier disruption appear to be the primary mechanisms.
Medication toxicity deserves special attention because it is the most reversible cause. Lithium, valproate, baclofen, cefepime, ifosfamide, levetiracetam, and tiagabine have all been reported to induce triphasic waves, sometimes at therapeutic doses in patients with renal impairment. Cefepime neurotoxicity has become particularly common in hospitalized patients and can be confused with nonconvulsive status epilepticus. A careful medication review with attention to renal dosing is essential for every patient with new triphasics.
Anoxic brain injury after cardiac arrest, hyponatremia, hypoglycemia, severe hyperglycemia with diabetic ketoacidosis, hypercalcemia, and hypothyroidism with myxedema coma round out the metabolic causes. Hashimoto encephalopathy and other autoimmune encephalitides can occasionally produce triphasic-appearing waves, although their patterns more often show focal features. A glance at what is a eeg test staffing model in your hospital helps explain why turnaround for these urgent reads varies.
Structural causes are rarer but important. Bilateral thalamic infarcts, Creutzfeldt-Jakob disease in its later stages, and large frontal lobe hemorrhages can mimic triphasic morphology. CJD typically produces periodic sharp wave complexes that share the three-phase appearance but occur at a slightly faster 1 Hz rhythm with a different clinical context of rapidly progressive dementia, myoclonus, and characteristic MRI findings in the cortical ribbon and basal ganglia.
Finally, nonconvulsive status epilepticus remains the diagnostic mimic that drives most consultations. The Salzburg consensus criteria help separate ictal patterns from triphasic waves, but in practice the overlap is real and a benzodiazepine trial is often the only way to be sure. We will return to this distinction in detail later in the article.
Hepatic triphasic waves classically appear at 1.5 to 2 Hz with a strong frontocentral predominance and a clear anterior-posterior lag of about 100 milliseconds. The background shows diffuse theta and delta slowing, often with loss of the normal posterior dominant rhythm. Reactivity to noxious stimulation is preserved in early grades but progressively diminishes as encephalopathy deepens toward coma.
Ammonia levels above 100 micromoles per liter strongly correlate with the pattern, though the relationship is not linear. Treatment with lactulose, rifaximin, and management of precipitants like gastrointestinal bleeding or infection typically clears the triphasics within 24 to 72 hours. Persistence beyond this window suggests an additional insult or progression to acute-on-chronic liver failure.
Uremic triphasic waves often run slightly faster, around 2 to 2.5 Hz, and may show somewhat sharper transients than hepatic patterns. Patients with blood urea nitrogen above 100 milligrams per deciliter are at highest risk, though the absolute number matters less than the rate of rise. Pre-dialysis EEGs in chronic kidney disease patients commonly show these features alongside diffuse slowing.
One of the most satisfying clinical observations is the rapid resolution of triphasic waves within hours of starting effective renal replacement therapy. Continuous renal replacement may produce more gradual improvement than intermittent hemodialysis. Dialysis disequilibrium syndrome itself can paradoxically worsen the pattern in the first session, so neuromonitoring during initial dialysis in severely uremic patients can be informative.
Drug-induced triphasic waves from cefepime, lithium, ifosfamide, and valproate often present with the most concerning clinical picture because the underlying cause is not immediately obvious. Pattern morphology is essentially identical to metabolic causes, so diagnosis requires a careful medication reconciliation and review of recent dose changes, renal function, and drug levels where available.
Septic encephalopathy triphasics tend to be more variable in frequency and may show greater background discontinuity. Patients in the ICU with persistent triphasics despite hemodynamic improvement have markedly worse outcomes. Empirical discontinuation of suspect medications combined with supportive care and treatment of the underlying infection usually leads to gradual EEG improvement over several days.
Administer 1 to 2 mg of midazolam or 1 mg of lorazepam IV while recording EEG and watching the patient. Clinical improvement with EEG resolution suggests nonconvulsive status epilepticus. EEG suppression without clinical improvement is the typical metabolic response. No change at all argues against both diagnoses and pushes you back to imaging and broader metabolic workup.
Distinguishing triphasic waves from epileptiform discharges is the single most consequential decision an electroencephalographer makes in this context. The 2021 American Clinical Neurophysiology Society Standardized Critical Care EEG Terminology folded triphasic waves into the broader category of generalized periodic discharges with triphasic morphology, acknowledging that the same wave can sit anywhere on a continuum from purely metabolic to overtly ictal. This terminology change reflects clinical reality but adds interpretive complexity.
Several features favor a metabolic origin. A clear anterior-posterior lag of 100 milliseconds or more, frequency below 2.5 Hz, preserved reactivity to stimulation, a frontocentral maximum, and a clinical picture dominated by encephalopathy rather than focal deficits all point toward metabolic disease. The waves tend to be smooth and rounded rather than sharply peaked, and they often wax and wane with arousal level over minutes to hours.
Features favoring an ictal interpretation include frequency above 2.5 Hz, evolving morphology with progressive changes in shape and field, sharper transients, lack of reactivity, associated subtle clinical signs like eye deviation, twitching, or autonomic changes, and a clinical history of prior seizures or acute brain injury. The Salzburg criteria formalize these features and provide structured guidance for ambiguous cases.
The benzodiazepine trial remains the workhorse diagnostic maneuver. Administering 1 to 2 mg of midazolam intravenously while continuously recording EEG and observing the patient can produce one of three outcomes. Clinical improvement with EEG resolution strongly suggests nonconvulsive status epilepticus and justifies escalation of antiseizure therapy. EEG suppression without clinical change is the typical metabolic response. No change at all suggests the pattern is neither truly ictal nor benzodiazepine-responsive.
Reviewing actual eeg test side effects alongside the tracing helps trainees develop pattern recognition. The classic teaching montage shows generalized 2 Hz triphasic complexes with a clear frontal maximum on Fp1-F3 and Fp2-F4 channels, fading as you move down to T3-T5 and O1-O2. The morphology is reproducible study to study within the same patient, which itself is a useful clue when comparing serial EEGs.
One pitfall worth flagging is the patient with both metabolic dysfunction and structural brain disease. An elderly cirrhotic patient with prior stroke can show triphasic waves with superimposed focal epileptiform discharges, and treatment needs to address both. Similarly, post-cardiac-arrest patients often have a hypoxic-ischemic background with periodic patterns that defy clean categorization and require individualized assessment using burst suppression ratios, somatosensory evoked potentials, and clinical exam.
Inter-rater agreement for triphasic waves versus GPDs without triphasic morphology is moderate at best, even among experienced electroencephalographers. This reality argues for over-reading rather than under-reading uncertainty in EEG reports and for direct conversation with the primary team about what the pattern does and does not mean. A confidently wrong report is worse than a carefully hedged one.
Writing a useful EEG report for a patient with triphasic waves requires more than morphological description. The clinical team wants to know whether the pattern represents seizures, what underlying cause it suggests, what they should do next, and how worried they should be about prognosis. A report that lists features without integrating them into a clinical recommendation falls short of what modern critical care neurology demands.
The ideal report opens with a brief technical summary, then describes the predominant pattern including frequency, distribution, reactivity, and morphology. Next comes the impression section, which should explicitly state whether the findings are most consistent with toxic-metabolic encephalopathy, possible nonconvulsive seizures, or a true ictal pattern. Recommendations follow, covering metabolic workup, medication review, and whether continuous EEG monitoring or repeat studies are warranted.
Treatment is directed at the underlying cause rather than the EEG pattern itself. Hepatic encephalopathy requires lactulose titrated to three to four soft stools daily, rifaximin 550 mg twice daily, and aggressive identification of precipitants including infection, GI bleeding, electrolyte disturbance, and constipation. Uremic patients need optimization of dialysis prescription. Septic patients need source control, appropriate antibiotics, and hemodynamic support. Drug-induced cases need discontinuation or dose reduction.
Prognosis varies substantially by underlying cause. Reversible drug toxicity and dialyzable uremia often clear within days with complete neurological recovery. Hepatic encephalopathy outcomes depend on liver disease severity and reversibility of precipitants. Sepsis-associated encephalopathy with persistent triphasic waves portends in-hospital mortality of 30 to 40 percent and increased risk of long-term cognitive impairment in survivors. Anoxic injury cases carry the worst prognosis when the pattern persists beyond 72 hours.
Patients and families often ask about the eeg test cost and what to expect during the study. A routine 20 to 30 minute EEG in the hospital typically costs $200 to $1,000 with insurance, while continuous EEG monitoring over 24 to 72 hours can reach $3,000 to $10,000 depending on facility and duration. Most insurance plans cover medically necessary EEGs ordered for encephalopathy workup without significant out-of-pocket costs beyond standard deductibles.
Side effects of the EEG itself are minimal. The test is noninvasive, painless, and carries no radiation exposure. Skin irritation from the electrode paste, mild headache from prolonged immobility, and rare scalp infections at electrode sites are the main concerns. Hyperventilation activation can briefly worsen lightheadedness, and photic stimulation occasionally provokes brief absence or myoclonic activity in patients with primary generalized epilepsy.
Follow-up EEGs are particularly valuable for tracking response to treatment. A second study 24 to 48 hours after initiating therapy can confirm pattern resolution and support de-escalation of monitoring. Persistent or worsening triphasic waves despite appropriate treatment should prompt reconsideration of the diagnosis, search for additional contributing causes, and discussion of goals of care if multiple organ systems are failing.
Practical preparation for an EEG in a patient with suspected metabolic encephalopathy starts before the technologist arrives at the bedside. Ensure the patient has clean, dry hair if possible, though in the ICU this is often not feasible. Document baseline mental status, recent medications, and time of last dose of any sedating agent so the reading neurologist can correlate background changes with pharmacology. Coordinate with nursing on sedation holidays when safe.
For the technologist, attention to electrode impedance, careful eye and EMG monitoring channels, and accurate annotation of patient state changes are essential for high-quality interpretation. Mark every nursing intervention, vital sign change, family interaction, and medication administration. These annotations transform an ambiguous record into a clinically actionable study by tying EEG changes to identifiable bedside events.
Continuous EEG monitoring has become the standard of care for any ICU patient with unexplained altered mental status, and triphasic wave patients often benefit from at least 24 hours of recording. Quantitative EEG trends including color density spectral array, amplitude-integrated EEG, and rhythmicity spectrograms help bedside clinicians screen for evolution into clear ictal patterns between formal interpretations. These tools do not replace expert review but extend the reach of limited neurology resources.
For trainees building skill in this area, deliberate practice with annotated case archives is the fastest path to competence. Aim to review at least 50 to 100 examples of triphasic waves across different etiologies to build robust pattern recognition. Pair each EEG with the clinical context, lab values, and outcome to anchor the visual pattern in real disease. Several professional society educational portals and textbooks provide curated case libraries for this purpose.
Communication with the primary team often matters more than the report itself. A five-minute phone call to discuss the findings, suggest specific workup, and clarify what the pattern does and does not mean prevents misunderstanding and accelerates appropriate treatment. Reading neurologists who build these communication habits become invaluable members of the multidisciplinary care team for critically ill patients with neurological dysfunction.
Documentation discipline pays dividends during readmissions and serial monitoring. Note the exact electrode locations of maximum amplitude, the precise frequency, reactivity testing results, and benzodiazepine trial outcomes with specific drug, dose, time, and observed effects. This level of detail allows future interpreters to make meaningful comparisons rather than starting from scratch each study.
Finally, do not forget that the EEG is a tool, not a diagnosis. Triphasic waves are a sign of brain dysfunction, and the patient in front of you needs a clinician who can integrate the EEG with everything else known about their case to make the best possible decision. The technology is sophisticated, but the judgment is human, and that judgment improves with every patient you care for thoughtfully.