Is the link between late diagnosed celiac and autonomic disorders established?

The short version: the retrieved evidence does not directly test a link between late-diagnosed celiac and autonomic disorders, but it shows plausible mechanisms that could connect them.

The records here do not include a study that tracks autonomic neuropathy or dysautonomia in people with late-diagnosed, untreated celiac disease. What they do show is that untreated celiac disease causes malabsorption of vitamin B12, with 41% of one cohort deficient [1]. B12 deficiency is a well-documented cause of peripheral and autonomic nerve damage. The records also show that untreated celiac disease drives a massive inflammatory response in the gut: S100B protein expression increases 24-fold, and nitric oxide production rises 247% above controls [2]. Systemic nitric oxide can affect vascular tone and autonomic signaling.

One study directly connects gluten sensitivity to sensory ganglionopathy, a form of nerve damage that can involve autonomic fibers [3]. In that study, 32% of sensory ganglionopathy patients had serologic evidence of gluten sensitivity, and a strict gluten-free diet stabilized or improved the condition in most who adhered [3]. Another study identified transglutaminase 6 (TG6) as a neuronal autoantigen in gluten ataxia, with antibodies found in 58% of gluten ataxia patients without enteropathy [4]. That same study found TG6 antibodies in 24% of peripheral neuropathy patients [4]. These are not autonomic disorders specifically, but they show gluten sensitivity can target nervous tissue.

The evidence is suggestive but not conclusive for autonomic disorders. The mechanism is plausible: chronic inflammation and nutrient malabsorption from untreated celiac can damage nerves, including autonomic ones. But no study in this retrieval measured heart rate variability, gastric motility, or other autonomic outcomes in celiac patients.

My call: the link is mechanistically plausible and partially supported by related neurological findings, but not directly established by the evidence here.

Confidence: low. The retrieval lacks a study that tests the specific question.