Autonomic Nervous System Testing

For physicians

What is measured

Continuous heart rate and photoplethysmographic waveform acquisition at rest and during standardized challenges — the Ewing battery: deep metronomic breathing, active standing or head-up tilt, and a Valsalva maneuver, with positional blood pressure measurement. Analysis covers heart rate variability in both time and frequency domains, and the derived cardiac autonomic reflex indices.

Peripheral oxygen saturation, autonomic function and microvascular arterial function are assessed together from the pulse waveform, which is what allows a single acquisition to serve the autonomic and the vascular questions.

The interpretive discipline

1. Read against the medication list, always. Beta-blockers blunt the heart rate response to standing and to deep breathing; anticholinergics and many antidepressants reduce variability and sudomotor output; diuretics and antihypertensives exaggerate orthostatic change. A result read without the list is not an interpretation.
2. Coach the maneuvers properly. A poorly performed Valsalva or an irregular breathing cadence produces a trace indistinguishable from autonomic failure. This is the single largest source of false abnormality.
3. Do not act on one low variability index. It moves with sleep, alcohol, caffeine, recent exercise and intercurrent illness.
4. Separate autonomic from arterial. Artery and autonomic nerve dysfunction are the most frequent complications of diabetes, kidney disease, hepatitis, thyroid failure and aging, their symptoms imitate each other, and their treatments differ. Measuring both in one sitting is the practical way to distinguish them.

Conditions to record

Fasting state, caffeine, nicotine, alcohol in the preceding 24 hours, sleep duration, recent exercise, time of day, room temperature, medication and timing. These are part of the result, not context for it. Staffing and workflow.

Where it fits

Component detail is on heart rate variability and cardiac autonomic reflex tests. Pair with sudomotor testing for a small-fiber picture, and with balance testing when the presentation is unsteadiness.

Evidence and limitations

Written for a reader who wants the literature rather than the summary. Where the evidence is thin we say so and give both sides, and where a claim is mechanistic rather than demonstrated we label it as such.

Mechanism

Cardiac autonomic control is a closed loop: baroreceptors sense arterial pressure, afferents reach the brainstem, and vagal and sympathetic efferents adjust heart rate and vascular tone within beats. Damage to the small, unmyelinated fibers carrying that traffic degrades the loop’s gain, which is measurable as loss of beat-to-beat variability at rest and as blunted responses to standardized challenges. Cardiovascular autonomic neuropathy is defined as impairment of cardiovascular autonomic control in the setting of diabetes after other causes have been excluded.

The Toronto Consensus criteria are explicitly graded: one abnormal cardiovagal test identifies possible or early involvement; at least two abnormal cardiovagal tests are required for definite or confirmed involvement; and orthostatic hypotension alongside abnormal heart rate testing identifies severe or advanced involvement. Progressive stages carry progressively worse prognosis. That grading is the reason a single abnormal index is treated as a hypothesis rather than a diagnosis.

What the evidence shows

The Toronto Consensus panel reported the prevalence of confirmed cardiovascular autonomic neuropathy at around 20%, rising to as high as 65% with age and diabetes duration, and identified it as a risk marker for mortality and cardiovascular morbidity.

A 2021 systematic review and meta-analysis screened 18,467 records and synthesized 26 studies. Across 16 studies comparing patients with (n=2,875) and without (n=11,722) cardiac autonomic neuropathy, the pooled relative risk of future cardiovascular events was 3.16 (95% CI 2.42 to 4.13). Across 19 studies with (n=3,679) and without (n=12,420), the pooled relative risk of all-cause mortality was 3.17 (95% CI 2.11 to 4.78). Risk was higher in type 1 than type 2 diabetes, and higher for definite than for possible disease — a dose-response relationship that strengthens the case for the graded criteria.

In a 2024 cohort of 211 patients with type 2 diabetes and overweight or obesity without established atherosclerotic disease, 67.2% had cardiac autonomic neuropathy, and 81.5% were in the very-high calculated cardiovascular risk band. Sudomotor scores and Ewing test scores both remained significantly associated with calculated risk on multiple linear regression.

Limitations of that evidence

• Most of the prognostic evidence is in diabetes. Extrapolating the pooled relative risks to a non-diabetic population with autonomic abnormality is not supported by these data.
• Observational design. These are prospective cohorts, not randomized trials. Autonomic neuropathy travels with duration of disease, glycemic exposure, nephropathy and vascular disease, and residual confounding cannot be excluded.
• Risk of bias. The meta-analysis authors judged three of the included studies to carry a risk of serious bias, and heterogeneity in how autonomic function was defined across cohorts is substantial.
• Technique dominates the measurement. Poorly coached deep breathing or a Valsalva performed at the wrong pressure produces a trace indistinguishable from autonomic failure. Between-study differences in acquisition are rarely reported in enough detail to compare.
• Medication. Beta-blockade, anticholinergic burden and antiarrhythmics alter every index in this battery, in different directions.

What remains uncertain

Two things are genuinely unsettled. First, whether screening changes outcomes: the Toronto Consensus stated that evidence on the economic case for testing was lacking, and apart from the preventive role of intensive glycemic control in type 1 diabetes, recommendations could not be made for most therapeutic approaches. The 2021 meta-analysis argues that population-based screening would identify a subgroup at disproportionately higher risk, which is a reasonable inference and not the same thing as a trial.

Second, what an abnormal result means outside diabetes. Autonomic abnormality is documented in Parkinson’s disease, amyloid neuropathy, chronic kidney disease, thyroid disease, long-standing alcohol use and simple deconditioning, and the prognostic weight attached to a given index almost certainly differs between them. Where a page on this site presents autonomic testing as useful in a non-diabetic patient, that is a claim about characterizing physiology and directing further workup, not a claim about a validated risk score.