Risks of CCI Treatment

Why this page exists

Fusion surgery and ligament injections help many people with craniocervical instability, and some of the clearest recovery stories in the ME/CFS world come from them. They are also permanent or repeated, expensive, mostly paid for out of pocket, and nobody can tell you your odds: every published result is a clinic reporting on its own patients with no comparison group. This page collects what has actually been published about the downside, in plain language, with every figure linked to its source. Read the harm accounts as carefully as the recovery accounts.

What the numbers can and cannot tell you. All of the surgical figures below come from "retrospective case series": a clinic looking back at its own patients, scored by the clinic, with nobody to compare against and no control for the placebo effect. The surgeons say this themselves. One of the largest series notes that "placebo and expectancy effects cannot be excluded" (Bloom et al. 2026), and another that subjects "may have exaggerated the degree of improvement" (Henderson et al. 2019). No controlled study of fusion against non-surgical care in hypermobility-type CCI was found; a 2022 systematic review found only low-level evidence across 16 studies (Lohkamp et al. 2022), and one controlled trial is reported to be under way (Bloom et al. 2026).

Fusion: what you give up, and what can go wrong

What fusion is. Occipitocervical fusion fixes the skull to the top one or two neck bones with titanium plates and screws so that bone grows across and the joint never moves again. Atlantoaxial (C1-C2) fusion does the same one level down.

It is permanent, and it costs movement. Skull-to-C2 fusion removes about 20 to 30 degrees of nodding and about 35 degrees of turning to each side (Henderson et al. 2019). The joints below then do the work the fused joint used to do, and the surgeons who do this operation write that "legitimate concerns exist regarding increased adjacent segment degeneration and the need for further fusions at the subaxial levels" (Henderson et al. 2024).

What the surgeons' own series of EDS patients report. Five surgical series, 273 EDS patients in total, published by the operating surgeons (one series, Rock et al. 2023, mixes fusion with other operations such as decompression and tethered-cord release):

Wound problems. 4 of 53 patients (7.5%) went back to the operating room within four weeks for a wound that reopened, which the authors call "disappointing" and attribute to EDS skin fragility (Henderson et al. 2024). In a separate study of wound closure technique, 9 of 18 craniocervical patients (50%) closed without a muscle flap needed an unplanned reoperation, against 5 of 39 (15%) with one (Ann Plast Surg 2020).
Painful hardware. In a 5-year series of 20 patients, 8 (40%) asked for the hardware to be removed one to four years later because of pain over the screws (Henderson et al. 2019). That series used an older technique.
Revision and further surgery. At one academic centre, 20 of 115 operations on EDS patients (17.4%) needed a revision, three quarters of them for continued or returning symptoms, and the revision rate for occipitocervical fusion alone was 5 of 18 (27.8%) (Rock et al. 2023). In the 5-year series, 12 of 20 (60%) had "multiple surgeries" within five years, including further neck surgery (Henderson et al. 2019). In the 53-patient series, 3 (5.7%) needed a further fusion lower in the neck within about 15 months (Henderson et al. 2024).
Other complications. In the largest series (107 fused), the authors list one deep infection needing hardware revision, one painful hardware removal, one adjacent-level instability needing the construct extended, five wound reactions to a bone substitute needing the graft removed and reinserted, two seromas and one wound reopening with readmission; that is 11 complication events in 107 patients (10.3%) by our count, with no implant failures or non-unions (Bloom et al. 2026). In a 26-patient series, 3 (11.5%) had an operative complication, including one vertebral artery occlusion without symptoms (Spiessberger et al. 2020).
No benefit, or worse. 13 of 53 (25%) reported no improvement or worsening overall, 9 (17%) reported being worse, which the authors attribute to co-morbid conditions, and 3 (6%) would not have the surgery again (Henderson et al. 2024). In the 107-patient series, 98 answered a follow-up questionnaire; of those, 12 of 88 (13.6%) who met full criteria and 3 of 10 (30%) borderline cases were dissatisfied (Bloom et al. 2026). At the academic centre, 6 of 64 (9.4%) had no improvement and 15 (23.4%) only temporary or mild improvement (Rock et al. 2023).
Deaths. None reported across the five EDS series (273 patients).

What happens in other populations. In adults fused for all causes (average age 57), a 2025 meta-analysis of 42 studies and 1,274 patients found surgical site infection in 7.0%, reoperation in 4.7%, spinal fluid leak in 2.9%, hardware failure in 2.6% and vertebral artery injury in 1.6% (Aguirre et al. 2025). Trouble swallowing or breathing after occipitocervical fusion affected 5 of 29 patients (17%) in one Japanese series, driven by the angle at which the head is fixed (Miyata et al. 2009); the EDS surgeons say they set that angle to avoid it and report none. Serious breathing problems after the operation, some fatal, are documented in mixed populations (Khan et al. 2025). About 1 in 5 people have a vertebral artery that runs where a screw would normally go, a figure the surgeons cite from the anatomy literature, which is why pre-operative CT imaging of the artery matters (Henderson et al. 2024).

Injections: what is known about the risks

Why the top of the neck is different. The needle path into the C1-C2 joint passes, on average, 7 millimetres from the vertebral artery and 6 millimetres from the sac of spinal fluid, and in a 488-scan anatomy study across 488 scans it would have passed through a vertebral artery 7 times and through the sac 11 times (McNabney et al. 2019). That artery supplies the brainstem. This is why this guide tells you to ask what imaging guidance a clinic uses and who trained them.

What has been published.

A fatal stroke after a steroid injection into the C1-C2 joint (Edlow et al. 2010). The drug was a steroid, not PRP or bone marrow, but the needle position is the same.
A spinal cord injury with one-sided numbness and pain, still present a month later, after prolotherapy near the C5 nerve root in the neck at a local clinic (Yun et al. 2011). That is the mid-neck, not the C1-C2 region, but it is the only published prolotherapy cord injury in the neck.
A survey of 171 prolotherapy practitioners (back and neck combined) in which they recalled 472 adverse events including 69 hospitalisations, 5 permanent nerve injuries, 164 spinal headaches and 9 events the authors class as "nonsevere spinal cord insult" (meningitis, paralysis, spinal cord injury) (Dagenais et al. 2006). No rate can be calculated from a recall survey.
An outbreak of 12 hospitalisations for bloodstream, joint and spinal infections after contaminated "stem cell" products were injected at outpatient clinics, including into the cervical spine (CDC 2018). Those were donor cord-blood products, not the patient's own blood or marrow; the point is that "stem cell" marketing covers products with very different oversight.
The FDA's position: stem cell products generally require FDA approval, the only approved ones are blood-forming cells for blood disorders, and unapproved products "may have significant safety issues that put patients at risk" (FDA consumer alert). Courts have upheld FDA action against clinics selling unapproved cell products, including a 2014 ruling that culture-expanded cells are a drug (US v. Regenerative Sciences LLC, 741 F.3d 1314, D.C. Cir. 2014; a 2024 ruling in the same line is summarised here).

What has not been published. We searched PubMed on 30 October 2026 for artery injury, stroke, brainstem or spinal cord injury, or serious infection after PRP, prolotherapy or bone marrow injections to the top of the neck in CCI patients and found no case reports and no series. That is not evidence of safety. These injections are done in private clinics that do not publish their complications, and the only numbers that exist are the clinics' own statements.

Transoral injections (PICL) in particular. No peer-reviewed outcome or complication series has been published after more than a decade of use. A registered case series (NCT03557606) is listed as withdrawn, with its last update in April 2019, and a registered sham-controlled trial (NCT03517761) has no published results. The clinic that performs the procedure states on its own blog that about 180 procedures had been done by March 2020 with no permanent harm, and that about 3 in 10 patients get little or no relief from the first procedure (clinic blog, 2020; clinic blog, 2022). Those are the clinic's figures, not independently verified.

Diagnosis is the weak link

Some people are treated for CCI whose symptoms were not caused by it. The surgeons say so themselves.

No agreed criteria. A 2022 systematic review found "a lack of consensus on diagnostic imaging parameters, indications, and outcomes of surgical treatment", ten different measurements in use, and "no method described to date which allows pre-operative assessment of the most probable neurological outcome" (Lohkamp et al. 2022). Surgeons who do the operation write that "there remains difficulty in the determination of the point at which craniocervical hypermobility becomes CCI" (Henderson et al. 2024).
The cut-offs move. The key angle (clivo-axial angle) is called abnormal below 130, 135 or 150 degrees depending on the author (Henderson et al. 2019; Lohkamp et al. 2022).
Experts disagree on the same scan. When a radiologist and a neurosurgeon measured 50 upright MRIs, they first disagreed unacceptably on two of the four measures, and even after agreeing a protocol the reliability of one measure was 0.69 on a scale where 1.0 is perfect agreement (Nicholson et al. 2023).
Healthy-looking necks meet the cut-offs. In those 50 people, scanned for ordinary neck problems, 2% to 8% met the angle cut-off and 6% to 14% met the Grabb-Oakes cut-off; the authors write that a CCI diagnosis "can be a serious condition with significant financial and psychological ramifications" (Nicholson et al. 2023).
A third are screened out. At the busiest centre, 63 of 190 patients fully evaluated for suspected CCI (33%) did not meet surgical criteria; of six people who did not improve after fusion, two probably had another cause in the authors' retrospective judgement (Bloom et al. 2026). The same authors write that flexion-extension imaging "would predictably generate false positives in asymptomatic individuals who are merely hypermobile", that their typical patient "appeared as a false negative on DMX", and that they abandoned collar trials for several reasons, including that prolonged wear caused "severe cervical muscle deconditioning".
The opposite error exists too. At the academic centre, all 13 patients who had a Chiari decompression elsewhere without fusion later needed a fusion (Rock et al. 2023).

What it costs

There is no published or official price for CCI fusion or for an injection series. What follows is what patients and their families have reported on public fundraising pages and in local press, with dates; none of it is verified by us.

Skull-to-C2 fusion in Barcelona: about 60,000 to 70,000 euros or pounds for the operation (fundraiser; fundraiser, 2020); 100,000 euros for a child's case including post-operative care (The Clare Echo, 2020); an estimated 160,000 euros for a longer fusion (Westmeath Independent, 2021).
Private US surgeon, out of network: about 50,000 dollars out of pocket, plus travel and post-operative physiotherapy not covered by insurance (fundraiser, 2025; fundraiser, 2025). Academic centres that take insurance exist; see the surgical directory.
Transoral injections: no published price; patients on one forum report roughly 12,000 to 15,000 dollars per procedure and two to five procedures, which we could not verify (dated figures in the guide).
Who pays: in the cases above, UK families describe the NHS as unable or unwilling to fund the operation, Irish families were ineligible for the HSE Treatment Abroad Scheme (one because the hospital is private, one for lack of an Irish consultant referral), and a German statutory insurer paid nothing for diagnostics or surgery. One family seeking a redo of a failed fusion in the US set a 750,000 pound target (fundraiser, 2020).

What patients report

Public, first-person accounts, summarised in our words with a link to the original. Clinicians are not named here. These are individual experiences, not rates.

Harm or no benefit after fusion

A UK woman with EDS, CCI and 20 years of severe ME had a fusion and decompression in January 2020. Her family describes the result as worse than before: a more unstable neck, worse ME, still bed-bound and with new incontinence, and a redo fusion abroad being sought with "no guarantees" (fundraiser, 2020).
A German woman with EDS, CCI and ME/CFS had a fusion in Barcelona in March 2025, improved at first, then deteriorated sharply in late 2025 and has been fully bedridden since December, with further surgery being discussed. Her insurer paid for nothing (fundraiser, 2026).
Three months after a skull-to-C2 fusion, a forum member sees no benefit, has worse POTS and new shoulder, back and leg pain, and spends most of the day in bed (Phoenix Rising, 2024).
A UK woman describes lifesaving neurosurgery in Barcelona; her neck fusion later broke after a routine knee operation, leaving her two years in a halo brace before a second, larger fusion (fundraiser, 2025).
A parent writes that their daughter's fusion and decompression helped her pain and some symptoms but did not cure her ME/CFS; she decided against tethered-cord surgery (Health Rising, 2019).

Harm or no benefit after injections

A person with 30 years of neck instability had bone marrow injections five times for loose upper-neck ligaments, says it was extremely expensive, lost the early progress, and now lives with pain they rate 10 out of 10 while asking about a full fusion (Mayo Clinic Connect, 2023; follow-up).
A 66-year-old with EDS reports ten years of prolotherapy that only ever helped for short periods (Living with Eagle forum, 2022).
A young woman lists what she tried before being offered a 50,000-dollar fusion: trigger-point injections that the fundraiser says progressively worsened her condition over time, Botox that weakened her neck, nerve blocks that did nothing and prolotherapy that did not address the laxity (fundraiser, 2025).

Clear benefit

Jeff Wood, a US man with severe bed-bound ME, had a skull-to-C2 fusion in January 2018 and reports that his PEM, POTS, sensitivities and brain fog resolved and that he now goes to the gym and runs. He also describes a wound infection at about three weeks with a week in hospital on IV antibiotics. The 2026 series cited above lists a Jeffrey D. Wood of the operating centre among its authors (Phoenix Rising, 2018; podcast, 2019).
Jennifer Brea, the US filmmaker, had a fusion in late 2018 followed by tethered-cord surgery and reports that her POTS and ME went into remission (Health Rising, 2019; The Sick Times, 2025).
A UK woman ill since a virus at 23 had a cervical fusion in Spain at 51 and reports that her PEM is gone and she has walked four miles; one shoulder now partly slips out of joint (Science for ME, 2019).
Positive accounts of injections are common in patient communities; the guide carries several alongside the clinic listings.

Before you commit

Ask for the clinician's own complication, revision and non-response rates, in writing, and compare them with the published series above.
Ask who treats you if something goes wrong after you fly home.
Get an independent second opinion from someone who does not sell the treatment. The questions to ask any injector or surgeon are in the guide.
If you are considering this because you are desperate, read the harm accounts as carefully as the recovery accounts. People in both groups faced the same decision you are facing.