Guillain-Barré Syndrome, Part I: What Was Measured, and When

The textbook version of a patient with Guillain-Barré has the diagnosis settled early and the argument confined to timing. Cases that reach review often do not look like that. The name covers several distinct diseases rather than one, and the two tests that would ordinarily settle the question are frequently unhelpful when decisions are being made: the spinal fluid can be normal in the first week, and nerve conduction studies (electrical tests that measure how fast and how strongly signals travel along a nerve) are not obtainable for hospitalized patients at every facility. The 2023 European Academy of Neurology and Peripheral Nerve Society guideline is the current reference for diagnosis and treatment, and it is explicit about what a normal test does not establish.¹ So the decisions that matter are made before anything has been confirmed, on the examination and on measurements repeated over time. Whether those were done is recorded, or not recorded, in the chart.

The Disease, Briefly

Guillain-Barré syndrome is an acute immune-mediated polyradiculoneuropathy (a rapidly developing disorder in which the immune system attacks the nerve roots and the peripheral nerves running out to the limbs). The guideline describes it as the most common cause of acute flaccid paralysis (sudden weakness with loss of muscle tone), with an annual global incidence of approximately one to two per 100,000 person-years, rising with age.¹ A general emergency department may see it once every few years, which is the epidemiological fact underlying most of the failure modes below.

The syndrome is not one disease, electrically speaking. Acute inflammatory demyelinating polyradiculoneuropathy (AIDP) is the demyelinating subtype, in which the attack falls on the myelin sheath (the fatty insulation that lets a nerve conduct quickly). The axonal subtypes, acute motor axonal neuropathy (AMAN) and its sensorimotor counterpart (AMSAN), damage the conducting fiber itself rather than its insulation, which matters because the fiber recovers more slowly. The demyelinating form predominates in every region studied, though the axonal share varies several-fold between them.² A caution travels with any such figure: in the Dutch cohort that validated the Brighton criteria (a standardized case definition used to grade how confidently a diagnosis can be made), only about three-fifths of confirmed patients could be assigned to a distinct electrophysiologic subtype at all.³

The course is monophasic (a single episode that worsens, plateaus, and then recovers, rather than relapsing). Weakness reaches its worst point, the nadir, within two weeks in roughly four-fifths of patients and within four weeks in nearly all of them.³ In the Dutch validation cohort, 95 percent of patients had a monophasic course, and one in ten of those had a treatment-related fluctuation (a transient worsening after immunotherapy has been given); 5 percent worsened more than eight weeks after onset, which raises the question of a different diagnosis.³

Recovery is common and incomplete. In the international cohort, 83 percent of patients in Europe and the Americas were walking unaided at one year, and mortality at one year in that group was 5 percent.² The complement is the part that reaches counsel: roughly one in six was not walking unaided a year later.

Guillain-Barré Names a Family, Not a Disease

The label covers several diseases whose separation is itself a live question, and two of the divisions matter to a record.

The first is the boundary with the chronic form. About 5 percent of patients initially diagnosed with Guillain-Barré later turn out to have acute-onset chronic inflammatory demyelinating polyradiculoneuropathy (A-CIDP, a related immune neuropathy that keeps progressing instead of resolving after a single episode), which the guideline says should be treated as chronic disease.¹ The distinguishing feature is time, and the guideline fixes two points on it.¹ A record that does not fix those dates cannot support either answer.

The second is the variants. Miller Fisher syndrome presents with eye-movement paralysis, incoordination, and areflexia (absent reflexes) rather than ascending weakness, so the differential it raises is brainstem (stroke or other injury at the base of the brain) rather than peripheral nerve. The guideline advises considering anti-GQ1b antibody testing (a blood test for an antibody against one component of the nerve membrane) when Miller Fisher is suspected.¹ This can be a send-out study rather than an in-house one, which means the confirmation arrives days after the decisions it might have informed.

A Normal Test Is Not a Negative One

The single most consequential misreading in these cases is treating an early normal test as evidence against the diagnosis. A normal result means the expected abnormality has not appeared yet, which in the first days is what the disease itself predicts; a negative result would mean the disease had been excluded, and none of these three tests delivers that early. Each of the three is frequently normal at first, and the guideline says so in terms.

Cerebrospinal fluid (CSF, the clear fluid surrounding the brain and spinal cord) is the classic example. The expected abnormality is albuminocytologic dissociation, a raised protein level with a normal white blood cell count, and it is genuinely characteristic once it appears. It frequently has not appeared yet. In the Dutch cohort, protein was elevated in 49 percent of patients tapped on the first day of weakness and in 88 percent after two weeks.³ Roughly half of patients who undergo a lumbar puncture (a spinal tap, drawing fluid from the lower back with a needle) on day one will have a normal protein. The guideline states the consequence directly: normal cerebrospinal fluid protein is common during the first week of the disease and does not exclude Guillain-Barré syndrome.¹ A record that documents a normal tap in the first days and reasons from it to discharge has used a test outside the window in which it discriminates.

Nerve conduction studies carry the same limitation, and a second one on top of it. The guideline states outright that a normal electrodiagnostic examination in the first week does not exclude the diagnosis.¹ The electrical hallmarks of demyelination (slowed conduction, blocked signals, and delayed responses from the nerve roots) take time to develop, and an early study that is normal or equivocal is consistent with the disease rather than against it. The second limitation is practical: inpatient electrodiagnostic testing is not available on demand at every facility. A record containing no nerve conduction study therefore does not establish that one was considered and declined, and the question the record has to answer is which of those it was.

Reflexes are the third. Absent or reduced reflexes (the tendon jerks elicited with a reflex hammer at the knee, ankle, and elbow) are the expected finding and the most useful single item on the bedside examination. That makes their absence from a record its own kind of evidence, because a reflex examination costs nothing and takes under a minute. The complication is that reflexes are not invariably lost early. The guideline places hyperreflexia (exaggerated rather than diminished reflexes) among the findings that make Guillain-Barré less likely, and then states in the same line that initial hyperreflexia does not exclude it.¹ The finding cuts in one direction only: preserved reflexes lower the probability without closing the question.

Pain compounds all three. In a long-term follow-up of Dutch patients, 36 percent reported pain in the two weeks before any weakness began.⁴ Severe back or limb pain preceding weakness is the presentation most easily attributed to a musculoskeletal cause (a strain, a disc, a mechanical back problem), and it arrives before the findings that would correct the attribution.

One thing the literature cannot supply is a rate. There is no reliable published figure for how often Guillain-Barré is initially missed, because the cohorts that would measure it enroll patients whose diagnosis was eventually confirmed. Patients who were sent home and never returned to a study center are structurally invisible to this literature. What exists is narrower and still useful: in a United Kingdom tertiary center series, 22.7 percent of patients had a diagnostic delay of five days or more, a figure describing referrals to that center rather than a population.⁵ An opinion that asserts a general missed-diagnosis rate is asserting something the evidence does not contain.

The Decision Space After the Diagnosis

Once Guillain-Barré is suspected, the chart has six decisions to document. Their absence is itself an analytical fact.

• Severity, classified and recorded. The Guillain-Barré disability scale runs across seven grades, 0 through 6, and a chart records the number rather than the phrase. Grade 3 or more is the entry that matters, because it is the point at which a patient can no longer walk unaided, and it is not an academic instrument in this setting. Both immunotherapy recommendations in the guideline are gated on it: they apply to patients unable to walk unaided.¹ A record that never grades function has omitted the variable that determines whether treatment was indicated at all.
• Respiratory function, measured rather than observed. The guideline advises assessing decline in respiratory function by measuring forced vital capacity (FVC, the volume of air a patient can exhale in one full breath) and single breath count (the numbers a patient can count aloud on one breath).¹ The measurement is the record. A nursing note reporting comfortable breathing is an observation of a patient whose oxygen saturation stays normal until it does not.
• Ventilation risk, stratified at admission. The Erasmus Guillain-Barré Respiratory Insufficiency Score (EGRIS) combines days from onset, facial or bulbar weakness, and a summed muscle strength score (the Medical Research Council sum score, which totals graded strength across six muscle groups on both sides) into an admission estimate of the risk of needing a ventilator.⁶ These models stratify groups rather than predicting an individual, and which model fits a given record is work for an expert; what belongs in the chart is evidence that the risk was estimated at all.
• Autonomic involvement, actively sought. The disease can affect the autonomic nerves (the involuntary system governing heart rate, blood pressure, and gut motility) as well as the ones moving muscle. The guideline lists autonomic instability (fluctuations in blood pressure or heart rate) among the risk factors for requiring mechanical ventilation that it advises assessing regularly throughout the admission.¹
• The immunotherapy decision, with its date. Which treatment, intravenous immunoglobulin (pooled antibodies given into a vein) or plasma exchange (a procedure filtering antibodies out of the blood); when it started; and how that timing sat against the onset of weakness.
• Escalation criteria, written before they are needed. The threshold at which this patient goes to intensive care, recorded in advance, is what converts a series of measurements into a plan.

The Monitoring the Diagnosis Requires

Monitoring in Guillain-Barré is not generic ward observation, and the guideline does not leave its frequency to judgment. Forced vital capacity should be checked between three and six times a day depending on severity, and while the patient is still declining, four-hourly monitoring is described as likely appropriate.¹ That is a published cadence, and a chart can be read against it.

The action thresholds are equally specific. A fall in forced vital capacity of more than 30 percent below the predicted baseline (the volume expected for a person of that age, sex, and height) should alert concern; a fall of more than 30 percent within 24 hours likely indicates immediate transfer to intensive care; and a decline of 50 percent in under 24 hours likely indicates the need for ventilation.¹ The first compares a patient against a population norm; the two that follow track change across a day, and so require an earlier reading to exist.

Two absolute volumes sit underneath those trends. Elective ventilation (intubation performed as a planned step rather than as a rescue) should be considered at 20 mL/kg of body weight, about 1.4 liters in a 70-kilogram adult, and at 10 mL/kg ventilation is described as almost inevitable.¹ A single breath count below 20 is offered as a bedside tool for the transfer decision.¹ In the international cohort behind the guideline's prognostic model, 10 percent of patients required mechanical ventilation within the first week from study entry, in a group that included mild cases and variants but excluded anyone already ventilated on entry, which makes it a floor rather than a whole-illness rate.¹

Two cautions belong with these numbers.

The first concerns what oxygen saturation can show. Neuromuscular respiratory failure is a failure of the bellows rather than of gas exchange: the lungs work and the muscles driving them do not, so carbon dioxide accumulates while oxygen saturation stays normal. In an audit of blood gas results from patients with stable chronic hypercapnic respiratory failure, among samples already showing raised carbon dioxide, 54 percent of the group with neuromuscular disease other than motor neuron disease, and 36 percent of the motor neuron disease group, had a normal oxygen reading at the same time, against none of the patients without neuromuscular disease.⁷ Two limits travel with that finding: the samples were selected for already being abnormal, so they cannot say how often the combination occurs, and the population was chronic rather than acute. What it establishes is a mechanism rather than a rate. It does not follow that arterial blood gas measurement (a blood sample drawn from an artery to measure oxygen and carbon dioxide directly) is also late; in that study the blood gas is what detected what the oximeter (the fingertip probe reading oxygen saturation) missed. The defensible statement is narrow: normal saturation does not exclude ventilatory failure in a neuromuscular patient.

The second cuts against the intuitive argument. The guideline ranks how confident it is in each predictor of respiratory failure, and it treats three clinical findings as better established than the vital capacity itself: a shorter time from onset to admission, bulbar involvement (weakness of the muscles of speech and swallowing, supplied by nerves arising from the brainstem), and a summed muscle strength score below 20 of a possible 60.¹ Single breath count is the least established of the three measurements, and the guideline notes it may be more an indicator than a predictor. A case theory resting on the spirometer as the missing instrument overstates the guideline's own ranking, which is on the page for either side to read. The stronger argument is that the examination findings predicting respiratory failure most reliably are the ones a record most often lacks.

One technical discrepancy is worth knowing. The guideline gives its optional pressure thresholds as a maximal expiratory pressure below 30 cmH2O or a maximal inspiratory pressure below 40 cmH2O.¹ These two measurements are the hardest a patient can blow out, and the hardest they can suck in, against a closed gauge. The 2001 paper from which the familiar bedside rule derives gives the thresholds the other way around, inspiratory 30 and expiratory 40.⁸ The two documents transpose the pressures. A report quoting one set against a record generated under the other invites a question it should answer first.

What the Literature Will and Will Not Carry

Three calibrations matter.

The treatment windows are asymmetric, and both carry a fence. The guideline strongly recommends starting intravenous immunoglobulin as soon as possible in patients unable to walk unaided, if still within the first two weeks from onset of weakness, and it carries a good practice point (a consensus recommendation issued where trial evidence is too thin to grade) extending that to the second-to-fourth week; it strongly recommends starting plasma exchange as soon as possible in the same patients within four weeks from onset.¹ A sentence asserting a single treatment window flattens two different recommendations, drops the good-practice extension belonging to only one of them, and loses the eligibility condition governing both. A patient who can still walk unaided falls outside the recommendation rather than inside a prohibition. That is a different fact from a contraindication, and different again from evidence that treatment would not have helped. What the record has to show is whether the question was reached at all.

Late treatment is unstudied rather than disproven, and the distinction is the one most likely to be lost in these cases. The trials that established benefit enrolled patients early, and the Cochrane review of immunoglobulin (a standardized synthesis of all the trials on one question) states in its own conclusion that more research is needed in patients whose treatment starts more than two weeks after onset.⁹ No published gradient relates the length of a delay to the size of the harm. An expert who testifies that a given delay caused a given outcome is reasoning past the literature, and an expert who says the treatment window is where the evidence lives, and that what happens outside it was never measured, is not.

Corticosteroids (steroid drugs given to suppress inflammation) are recommended against, which is not the same as shown to harm. The guideline strongly recommends against oral corticosteroids and weakly recommends against intravenous methylprednisolone alone or combined with immunoglobulin, the weaker recommendation resting on probable lack of efficacy rather than demonstrated harm.¹ The Cochrane review of corticosteroids attaches its delayed-recovery finding specifically to the oral route, on evidence it grades as very low quality.¹⁰ A report asserting that steroids worsen Guillain-Barré has overstated all three sources at once.

Where the Management Goes Wrong

The cases that reach review tend to fail in a small number of recurring ways.

• Weakness attributed to a mechanical or functional cause. Back and limb pain arrives before the neurological findings, the examination is recorded as non-focal, and the working diagnosis is a strain or anxiety. The reflex examination that would have redirected the workup does not appear in the note.
• A normal spinal tap read as a rule-out. The tap is done on day one, the protein is normal, and the normal result is treated as excluding a disease whose own guideline says it does not.
• An early nerve conduction study treated the same way. The study is normal or equivocal in the first days, which is expected, and is recorded as reassuring.
• Admission without a respiratory measurement. The patient is admitted for observation, and the chart contains oxygen saturations and respiratory rates but no vital capacity and no breath count. Nothing was measured that would have shown the trend.
• Oximetry used as the respiratory monitor. Saturations are recorded, they remain normal, and the decline they cannot detect proceeds until it presents as an emergency.
• Measurement without a threshold. Vital capacity is recorded once or twice, no baseline is established, no escalation criterion is decided upon, and a falling number sits in the chart without anyone owning the decision it should have triggered.
• Immunotherapy started without reference to its window. Treatment is given, the record notes the date, and nothing documents where that date sat relative to the onset of weakness or to the patient's inability to walk.
• A chronic course read as a fluctuation. The patient worsens again after treatment, the worsening is recorded as a treatment-related fluctuation, and nobody revisits the label when progression continues past eight weeks. The chart shows a relapse without a date attached to it, and the patient who needed treatment for a chronic disease is treated as a resolved acute one.
• Steroids given as the immune treatment. A course of corticosteroids appears in a chart where immunoglobulin or plasma exchange belonged.

These are not separate failure modes. They compound. A diagnosis attributed to a mechanical cause produces an admission without a neurological baseline; an admission without a baseline produces observation instead of measurement; observation instead of measurement produces a first recorded vital capacity taken after the decline has already happened, at which point there is no trend to read and no threshold to have crossed.

Reading a Guillain-Barré Case in Context

A defensible review of a case in which Guillain-Barré is central to the evidence, plaintiff or defense, works through a small set of specific questions.

• What did the first examination record about reflexes, and if reflexes are absent from the note, was any neurological examination documented at all?
• Where in the course was each test performed, counted in days from the onset of weakness rather than from presentation?
• Was a normal spinal tap or an unremarkable early nerve conduction study treated as excluding the diagnosis, and does the reasoning appear in the note or only in the disposition?
• Was function graded on admission in a way that establishes whether the patient met the condition the treatment recommendations attach to?
• Does the record contain serial vital capacity or single breath count measurements, at what interval, and against what baseline?
• Was a written escalation threshold recorded before the decline, or does the chart contain only the response to it?
• Where did the immunotherapy date sit relative to the onset of weakness, and does the record show the window was considered?
• Do the dates of onset, of the worst point, and of any subsequent worsening appear in the record with enough precision to separate an acute illness from a chronic one?
• Where no nerve conduction study appears, does the record distinguish a study that was considered and declined from one the facility could not offer?
• Where the patient was discharged from an early visit, does the record show what would bring them back, to whom, and within what interval, or only a diagnosis and a disposition?
• Does the outcome sit inside one of the recurring failure patterns, or outside them?

Causation raises a separate set of questions, particularly where a vaccine or an antecedent infection is alleged as the cause. Part II takes those up.

Most cases resolve once these elements are placed in order. Some that look strong on the complaint dissolve against a chart showing serial measurements, a documented threshold, and a decline that outran a plan which was in fact in place. Others that look strong on the defense collapse against a chart in which the diagnosis was reached promptly and nothing was measured afterward.

The diagnosis names the disease. The measurements show what was known, and when.

References

Footnotes

1. van Doorn PA, Van den Bergh PYK, Hadden RDM, et al. European Academy of Neurology/Peripheral Nerve Society Guideline on diagnosis and treatment of Guillain-Barré syndrome. Eur J Neurol. 2023;30(12):3646–3674. doi:10.1111/ene.16073 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰

2. Doets AY, Verboon C, van den Berg B, et al. Regional variation of Guillain-Barré syndrome. Brain. 2018;141(10):2866–2877. doi:10.1093/brain/awy232 ↩ ↩²

3. Fokke C, van den Berg B, Drenthen J, Walgaard C, van Doorn PA, Jacobs BC. Diagnosis of Guillain-Barré syndrome and validation of Brighton criteria. Brain. 2014;137(Pt 1):33–43. doi:10.1093/brain/awt285 ↩ ↩² ↩³ ↩⁴

4. Ruts L, Drenthen J, Jongen JL, et al. Pain in Guillain-Barré syndrome: a long-term follow-up study. Neurology. 2010;75(16):1439–1447. doi:10.1212/WNL.0b013e3181f88345 ↩

5. Bose S, Loo LK, Rajabally YA. Causes and consequences of diagnostic delay in Guillain-Barré syndrome in a UK tertiary center. Muscle Nerve. 2022;65(5):547–552. doi:10.1002/mus.27506 ↩

6. Walgaard C, Lingsma HF, Ruts L, et al. Prediction of respiratory insufficiency in Guillain-Barré syndrome. Ann Neurol. 2010;67(6):781–787. doi:10.1002/ana.21976 ↩

7. Gray E, Menadue C, Piper A, Wong K, Kiernan M, Yee B. Hypercapnia is not excluded by normoxia in neuromuscular disease. ERJ Open Res. 2024;10(4):00927-2023. doi:10.1183/23120541.00927-2023 ↩

8. Lawn ND, Fletcher DD, Henderson RD, Wolter TD, Wijdicks EF. Anticipating mechanical ventilation in Guillain-Barré syndrome. Arch Neurol. 2001;58(6):893–898. doi:10.1001/archneur.58.6.893 ↩

9. Hughes RAC, Swan AV, van Doorn PA. Intravenous immunoglobulin for Guillain-Barré syndrome. Cochrane Database Syst Rev. 2014;(9):CD002063. doi:10.1002/14651858.CD002063.pub6 ↩

10. Hughes RAC, Brassington R, Gunn AA, van Doorn PA. Corticosteroids for Guillain-Barré syndrome. Cochrane Database Syst Rev. 2016;(10):CD001446. doi:10.1002/14651858.CD001446.pub5 ↩