What Does the Evidence Say About Craniosacral Therapy?

What systematic reviews have found

The most recent reviews have not found clear statistical evidence that CST works across the conditions it's commonly used for. A 2024 systematic review in Healthcare (PubMed PMID 38540643) looked at 15 randomised controlled trials and concluded that the evidence did not show statistically significant or clinically meaningful changes in pain or disability for headache disorders, neck pain, low back pain, pelvic girdle pain, or fibromyalgia. A separate 2024 meta-analysis in Frontiers in Medicine examined 24 RCTs with 1,613 participants and also found no significant effects in the primary analysis, with only limited signals for neonatal structure and chronic somatic pain among secondary outcomes.

These findings matter and shouldn't be brushed aside. Anyone practising or recommending CST should know about them.

Why CST trials are hard to design

That said, the limits of the existing trials are worth understanding. The standard tool for testing a treatment is the double-blind randomised controlled trial — neither patient nor clinician knows who got the real treatment. That works well for drugs. It's much harder to apply to a hands-on therapy like CST.

Designing a credible sham CST — something that feels like real CST but contains none of the active elements — is genuinely difficult. In most CST trials, the sham is a therapist making light contact without intentional technique. It's not clear how different that is from what an inexperienced or half-present real therapist does. When the control resembles a relaxed version of the real thing, the study will underestimate any real effect.

There are also questions about dose. Most CST trials have used short courses — four to eight sessions — which practitioners often say is too few for longer-standing or complex conditions. The mix of training backgrounds among trial therapists is another problem. A study that lumps together weekend-course practitioners and 700-hour biodynamic graduates isn't measuring the same thing twice.

The fact that 14 of the 24 studies in the Frontiers meta-analysis were rated 'High Risk' of bias tells you something important: most of the trials we have aren't yet good enough to be definitive either way. That isn't pro-CST spin. It's a methodological reality, and it cuts both ways.

Where positive evidence does exist

Despite the mixed picture in systematic reviews, some areas have produced more encouraging results. Chronic neck pain has shown positive findings in several individual trials. Tension-type headache has been a focus of positive studies, and some practitioners report particularly good results with migraine, though the meta-analytic picture is less clear. Stress-related conditions and general nervous system dysregulation are areas where practitioners consistently report benefit. Trial evidence here is limited, but the biological plausibility of gentle touch affecting the autonomic nervous system is more straightforward than some of CST's other theoretical claims.

For infant work — unsettled babies, birth trauma, colic — CST is widely used and parents often report significant improvement. The 2024 reviews were not encouraging about the formal trial evidence here, but trial quality for infant studies has been particularly variable, and many practitioners have deep experience in this area. For premature infant care, some hospital-based studies have shown positive effects on specific neonatal outcomes, which the Frontiers meta-analysis recorded as a signal, though with wide confidence intervals.

The gap between trials and clinic

There's a familiar problem in complementary therapy research: the gap between what RCTs show and what practitioners and clients experience. CST is a clear example. The therapy has been practised in some form for over a century. Hundreds of thousands of people have received it. The amount of positive qualitative evidence — testimonials, practitioners' notes, healthcare professionals who use CST in broader treatment plans — is substantial, even if it doesn't meet the standards of randomised evidence.

None of that proves the mechanism. The theoretical basis of CST — the model of craniosacral rhythm and cranial bone mobility — is contested by mainstream anatomy, and that's a real problem for the field. But the gap between contested mechanism and reported benefit isn't unusual in medicine. Acupuncture's mechanisms aren't fully understood. Many physiotherapy techniques have limited RCT support. The question isn't whether we can perfectly explain the effect. It's whether the effect is real and worth investigating with better trials.

Many practitioners and researchers take the view that what CST reliably produces — a deep settling of the nervous system, a quality of restful contact most people rarely experience — may have real clinical value even if Sutherland's original theory of cranial bone movement doesn't survive scrutiny. The biodynamic tradition in particular has largely moved away from precise mechanistic claims and toward a more phenomenological framing.

Safety, side effects, and what the evidence can actually tell you

Evidence is not only about whether a therapy outperforms a control group. It is also about safety, dose, who was included in trials, who was excluded, and what kind of practitioner delivered the care. On safety, CST generally appears low risk when it is delivered gently by a trained practitioner who respects medical red flags. Reported side effects in trials and clinical practice are usually mild and temporary: tiredness, emotional release, temporary soreness, light-headedness, or a short-term symptom flare. That safety profile is one reason many people try CST alongside physiotherapy, psychotherapy, osteopathy, massage, or conventional medical care.

Low risk does not mean no risk. The evidence base is too small to prove safety for every situation, and trials often exclude the people where caution matters most: recent head or spinal trauma, suspected fracture, raised intracranial pressure, severe neurological symptoms, acute infection, cancer-related bone pain, unstable pregnancy complications, recent surgery, shunts or implanted devices, anticoagulant use, or infants with fever, poor feeding, breathing difficulty, seizures, persistent vomiting, or poor weight gain. These situations require medical assessment first. A CST practitioner who understands the evidence will say that clearly.

The most useful way to read the research is not “CST is proven” versus “CST is disproven.” A better question is: for this person, this problem, this practitioner, this dose, and this outcome, how strong is the evidence, how plausible is the goal, and what would count as success? CST has the strongest practical rationale when the goal is downshifting an overactive nervous system, improving comfort, supporting body awareness, or complementing a broader care plan. It has a weaker evidence basis when it is presented as a stand-alone cure for a named disease or as a diagnostic system that can replace medical evaluation.

That distinction matters for informed consent. A client can reasonably choose CST because the touch is gentle, the practitioner is well trained, the goal is realistic, and the downside is low. A client should not be told that CST has already been proven to reverse complex disease, cure neurological conditions, or make standard care unnecessary. The evidence supports curiosity and careful use; it does not support hype.

A practical checklist for reading CST research

When you read a CST study, start with five questions. First, what condition was studied? Evidence for chronic neck pain, premature infant outcomes, or general wellbeing does not automatically transfer to migraine, autism, long COVID, fertility, concussion, or trauma. Second, what was the comparison group? A wait-list control, usual care, light-touch sham, massage-like contact, and another active manual therapy all answer different questions. Third, how many sessions were delivered? A four-session protocol is not the same as a three-month course with a practitioner who specialises in the problem. Fourth, who delivered the treatment? Training route, experience, supervision, and scope of practice can affect results in hands-on therapies. Fifth, what outcome was measured: pain score, disability, sleep, anxiety, parent report, autonomic marker, medication use, or quality of life?

Those details help explain why reviews can sound negative while individual clients and practitioners still report meaningful change. A systematic review averages across heterogeneous studies. A person in a clinic experiences one practitioner, one relationship, one problem, and one course of care. Neither level should erase the other. The field needs larger, better-described, independently replicated trials with credible control groups, transparent adverse-event reporting, and clearer practitioner-training descriptions. It also needs pragmatic studies that ask real-world questions: when CST is added to usual care, which people improve, by how much, and at what cost?

For now, the most responsible interpretation is measured. CST is not a magic cure, and some of its older anatomical explanations remain contested. At the same time, it is a coherent hands-on practice with a long clinical history, many trained practitioners, a generally gentle safety profile, and enough promising signals to justify better research rather than dismissal. If you are considering it, use the evidence to set expectations: choose a qualified practitioner, define one or two outcomes you care about, keep appropriate medical care in place, and reassess after a limited number of sessions.