Lidocaine resistance and EDS

In a 2019 survey of 980 adults with Ehlers-Danlos syndromes who had undergone a dental procedure with local anesthetic, 88% recalled inadequate pain prevention. The comparable figure in respondents without EDS was 33%. The most effective single agent in the EDS group was articaine, and it worked for 30% of the people who tried it. Lidocaine, the standard amide on a dentist's tray in the United States, was reported as effective far less often.

The phenomenon is older than the survey. In 2005, Alan Hakim, Rodney Grahame, Paul Norris, and Colin Hopper published a short paper in the Journal of the Royal Society of Medicine titled "Local anaesthetic failure in joint hypermobility syndrome." They surveyed 172 women with joint hypermobility syndrome (the term then used for what is now hypermobile EDS), 53 non-hypermobile age-matched controls, and 28 women with hypermobility who did not meet full JHS criteria. The question was whether a previous local anesthetic, given for dentistry, minor surgery, or epidural, had worked as well as it should have. Fifty-eight percent of the JHS group said no. Twenty-one percent of controls said no. The control rate is not zero because anesthesia sometimes fails in everyone. The signal in the JHS group is what separates the conditions.

Twenty-one years separate Hakim 2005 from the present. The mainstream dental literature has not closed the loop. AAOMS has no position statement on local anesthetic resistance in connective tissue disorders. The American Dental Association's pain management resources do not flag EDS as a population in which standard dosing is unreliable. Most general dentists in the United States have never been taught to ask about Ehlers-Danlos syndrome before drawing up a carpule of 2% lidocaine with epinephrine. Most dental anesthesia textbooks do not name the condition.

The clinical effect of that gap is direct. Adults and children with EDS undergo dental fillings, extractions, root canals, suturing, biopsies, and minor surgical procedures with anesthesia that does not fully take. When the person says "I can still feel that," the proceduralist often responds with the same script: anxiety, low pain tolerance, a need to relax. The procedure continues. The person remembers it.

What the surveys show

The Hakim paper was a clinical observation written up as a letter, not a definitive epidemiological study. It came out of a Danish skin biopsy program in which clinicians took punch biopsies from people with EDS type III to measure tissue strength, noticed that the local anesthetic kept failing, and started asking each person about their broader anesthetic history. Every person they asked reported a prior partial or complete failure in dental or obstetric anesthesia. The Hakim survey followed.

Schubart and colleagues published the larger follow-up in 2019 in the Journal of Dental Anesthesia and Pain Medicine, titled "Resistance to local anesthesia in people with the Ehlers-Danlos Syndromes presenting for dental surgery." The cohort was 980 adults with EDS and a non-EDS comparison group, surveyed about adequacy of dental anesthesia. The headline number was the 88% inadequate-pain-prevention rate in the EDS group versus 33% in the non-EDS group, an almost three-fold difference. The most effective agents in the EDS group were articaine, bupivacaine, and mepivacaine, in that order. Lidocaine and procaine were less reliable. Even the best agent, articaine, succeeded in only 30% of the EDS respondents who had received it.

Other survey work and clinical case series have produced failure-rate estimates across the EDS population in the 50% to 90% range, depending on which subtype is studied, which agent is used, which procedure is involved, and how failure is defined. The variability is wide. The direction is consistent. Adults with EDS do not numb the way the dosing tables predict.

What might be happening

The mechanism is not settled. Several hypotheses live in the literature.

The diffusion hypothesis starts with the connective tissue. Local anesthetics injected into subcutaneous tissue or near a nerve must diffuse through the surrounding matrix to reach the sodium channels on the axon membrane. If the collagen scaffold and the ground substance are different in EDS, the drug may distribute differently, clear faster, or sit in a tissue compartment that the nerve does not see. The Schubart authors note articaine's higher lipid solubility as a possible reason it performs better than lidocaine in this group. Articaine crosses membranes more readily, which could compensate for an unfavorable diffusion environment.

The fascial hypothesis is related. Fascia in EDS is mechanically different. A nerve block depends on the anesthetic reaching the right tissue plane and staying there long enough to soak the nerve. Looser fascia and altered tissue compliance could redirect the bolus or accelerate its clearance through capillary uptake.

The vascular clearance hypothesis points to the vessels. Many adults with hEDS have abnormally distensible blood vessels and altered autonomic regulation. If local capillary uptake of the drug is faster than typical, the effective tissue concentration around the nerve drops below the blocking threshold sooner.

The sodium channel hypothesis is the most speculative. Voltage-gated sodium channels, particularly NaV1.7 encoded by SCN9A, are expressed at high levels in pain-sensing neurons. Local anesthetics work by binding sodium channels and preventing the propagation of action potentials. Specific gain-of-function variants in SCN9A, including I136V, I848T, and V1316A, have been shown in laboratory studies to reduce lidocaine binding affinity, with significantly higher IC50 values than wild-type channels. These variants are described in primary erythromelalgia and small fiber neuropathy populations. Whether they are enriched in EDS is an open question. The 2022 review on resistance to local anesthetics in the British Journal of Anaesthesia treats the SCN9A connection as a plausible contributor to congenital local anesthetic resistance, distinct from the EDS phenotype. The honest read is that a subgroup of people with EDS may also carry sodium channel variants that worsen the response, but no study has yet stratified an EDS cohort by SCN9A genotype.

These hypotheses are not exclusive. Diffusion, clearance, and channel pharmacology can all contribute to a single person's experience.

What is on the dental tray

Standard dental local anesthetic in the United States is 2% lidocaine with 1:100,000 epinephrine. It is the cheapest, most widely stocked amide, and the one most general dentists reach for first. The Schubart data put it near the bottom of the EDS efficacy ranking.

Articaine 4% with epinephrine is the next most common choice in North American practice and the leading choice in much of Europe. It performed best of the agents surveyed in the EDS group, with the caveat that 30% efficacy is still a 70% failure rate. Mepivacaine 3% without vasoconstrictor and bupivacaine 0.5% are typically used for longer procedures or for people who tolerate epinephrine poorly. Both performed better than lidocaine in the EDS survey.

The dosing implications are not in any official guideline. Practitioners who treat EDS populations report using higher initial volumes, supplementing buccal infiltration with intraligamentary or intraosseous injection, allowing longer onset windows of ten to fifteen minutes rather than the standard three to five, and re-dosing earlier and more often than the textbook recipe suggests. Nerve blocks are sometimes preferred over infiltration because they target a single nerve trunk rather than depending on diffusion through a wide tissue field. None of this is in the AAOMS or ADA guidance because the AAOMS and ADA guidance does not address EDS.

A clinical trial registered as NCT05603741, "Local Anesthetic Response in Ehlers-Danlos Syndrome," is one of the first prospective studies to compare anesthetic agents in a controlled EDS cohort. As of May 2026, the cohort is small and the results are pending. A single trial will not establish standard of care. It is the start of the prospective evidence base that should have been built two decades ago.

The "your symptoms are anxiety" pattern

The most common response to a person who reports inadequate anesthesia mid-procedure is to question whether the person is feeling pain or feeling fear. The framing is gendered. EDS is diagnosed in women at ratios reaching 10 to 1 in some clinical series. Women who report pain are more likely than men to have that pain attributed to a psychological cause. A person in a dental chair who has lived with under-treated pain for years has a calibrated sense of what numb feels like. When she says "I can still feel that," she is reporting an instrument reading, not a panic attack.

The pattern repeats across procedural medicine. Skin biopsies done with insufficient anesthetic in dermatology. IUD insertions done with no anesthetic at all. Endometrial biopsies. Joint injections. Suturing in the emergency department. Each of these procedures has its own literature on under-anesthesia in women, and EDS sits as an unstudied confounder underneath all of them.

A person with EDS arriving for a procedure has a short, practical script that helps. State the diagnosis by name. Ask whether the practitioner has used articaine in EDS. Ask for the longer onset window before testing the field. Agree in advance on a stop signal that is not "I can still feel that." Bring a printed copy of the Schubart 2019 paper. None of these steps fix the system. They shift the conversation from "is this real" to "what do we do."

What the data infrastructure would have done

The Hakim survey is from 2005. The Schubart survey is from 2019. Between those two dates, roughly fifteen years passed in which the phenomenon was reported in the literature, repeated in EDS communities, validated in a larger cohort, and still not absorbed into mainstream dental training. As of 2026, a person with EDS who shows up at a general dental practice in the United States cannot assume the dentist knows that lidocaine often does not work in this population.

The question is why a finding survives twenty years of community reporting and two formal surveys without entering the standard textbook. Part of the answer is that dentistry's evidence base is built around the average dental patient, and EDS is treated as a curiosity at the edge of the curriculum. Part of the answer is that the relevant data, the procedure-by-procedure record of how each agent performed in each person, sits in private chart notes that no one aggregates.

Longitudinal patient-reported outcomes, structured at the level of agent, dose, onset time, procedure, and adequacy of pain control, would have closed this loop in a fraction of the time. Ten thousand adults with EDS reporting their experiences with each anesthetic across each procedure type, contributed once and analyzed continuously, would generate the dose-response and agent-comparison data that no single academic survey can produce. The infrastructure for that aggregation did not exist when Hakim wrote the JRSM letter, and it still does not exist in any standardized form across dental practice.

The Schubart paper is a survey of recall. It depends on what people remember about procedures they had years before. A prospective community-contributed registry, capturing each procedure as it happens, would replace recall with measurement. The phenomenon would then be a number, not a clinical impression. Numbers move into textbooks. Clinical impressions wait twenty years.

Read about the diagnostic odyssey in EDS as a systems failure.