Functional Medicine for Type 2 Diabetes: What a New Randomized Trial Shows

This article is for educational purposes intended for healthcare practitioners and informed readers. It does not constitute medical advice. Anyone on insulin therapy should make changes to diet, supplements, or insulin dosing only under the supervision of their prescribing clinician.

Why This Trial Matters

Naturopathic and functional medicine practice rests on a testable premise: that layering diagnostic testing, dietary elimination, targeted supplementation, and intensive lifestyle coaching onto standard care produces better outcomes than standard care alone. That premise is rarely tested with a randomized controlled design, because full functional medicine programs are expensive, labour-intensive, and hard to standardise for a trial protocol. Most of the evidence base for "functional medicine" as a system, rather than for its individual components such as a single nutrient or herb, is observational or comes from single-arm pilot programs with no comparison group.

In March 2026, a team from the Cleveland Clinic Center for Functional Medicine published one of the first randomized trials to test a comprehensive functional medicine program against usual care in a genuinely high-stakes population: adults with insulin-treated type 2 diabetes. The result was a clear null finding on the primary endpoint, a signal of increased hypoglycemia risk in the intervention arm, and a candid discussion from the authors, several of whom are prominent functional medicine clinicians, about why the program did not deliver on its premise. This is exactly the kind of evidence a site built on honest evidence-strength claims should cover, including when the result is unfavourable to the model it is reviewing.

The Trial: Design and Population

The study, "A functional medicine approach to diabetes control and quality of life: an open-label randomized pilot study," was published in the Journal of Clinical & Translational Endocrinology (volume 44, article 100437) by Eldib, Hatipoglu, Beidelschies, Alejandro-Rodriguez, Xiao, Rothberg, Hyman, and Zhou, and funded by the Cleveland Clinic Foundation.

It was an open-label (unblinded) randomized pilot trial run over 24 months. Seventy-one adults with type 2 diabetes who had been on insulin therapy for at least 12 months were randomised, stratified by total daily insulin dose (above or below 75 units), to one of two arms:

• Functional medicine (FM) plus usual care: 34 participants
• Usual care (UC) alone: 37 participants

Eligible participants were 18 or older, on insulin for a year or more, with no positive GAD antibody (excluding likely type 1 or LADA), HbA1c of 12% or below, no history of diabetic ketoacidosis, no pregnancy, no cognitive impairment or dementia, and an eGFR of at least 45 mL/min/1.73m².

What the Functional Medicine Arm Actually Received

The FM arm was not a single supplement or a single dietary change. It was a full program: diagnostic testing and symptom-based treatment planning, an elimination-style dietary protocol, and a tailored supplement regimen that could include a multivitamin, omega-3 fatty acids, alpha-lipoic acid, magnesium, vitamin D3, and berberine, alongside lifestyle modification covering exercise, stress reduction, and sleep hygiene. Visit frequency was substantially higher than usual care: physician visits at months 1, 2, 3, and 9, nutrition visits at months 1, 2, 3, 6, and 9, plus optional health coaching and dietitian sessions. This is a reasonable real-world approximation of what a well-resourced functional medicine clinic delivers, which is precisely what makes the trial informative: it tested the model as practised, not a stripped-down version of it.

The Primary Result: No Significant Difference

The primary endpoint was a composite: discontinuation of insulin with no increase in HbA1c, or an HbA1c below 7%, at 12 months. In the FM arm, 3 of 20 participants who completed the 12-month assessment (15.8%) met this endpoint, compared with 3 of 27 in the usual care arm (11.1%). The difference was not statistically significant (p = 0.68).

The gap between the enrolled numbers (34 and 37) and the numbers analysed at 12 months (20 and 27) reflects substantial attrition: 24 of the 71 randomised participants, 33.8%, were lost to follow-up by 12 months, and the dropout was uneven. Only 58.8% of the FM arm remained in the study at 12 months, versus 72.9% of the usual care arm. A more demanding, higher-visit-frequency intervention arm losing participants faster than a lower-burden control arm is itself a finding, not just a limitation, and the authors treat it as central to interpreting the result.

Secondary Outcomes: A Consistent Pattern of No Benefit, and One Safety Signal

Across the secondary outcomes, the pattern was consistent with the primary result. BMI fell more in the FM arm at 12 months (−2.4 kg/m²) than in usual care (−0.4 kg/m²), but the difference was not statistically significant (p = 0.38), and the authors report that this early weight loss in the FM arm was not sustained through the full 24 months. HbA1c change and insulin dose reduction at 12 months were not significantly different between arms. Diabetes distress, measured with the DDS17 scale, improved in both arms with no significant between-group difference (FM −5.16 points versus UC −7.36 points). Quality-of-life measures (the MSQ symptom questionnaire, SPADE pain and mood screening, and PHQ-9 depression screening) showed no statistically significant difference between groups at any time point.

The one outcome that moved in a direction of concern rather than neutrality was hypoglycemia risk between months 6 and 12, estimated at 52% probability in the FM arm versus 32% in usual care. Read alongside the early, unsustained weight loss in the same arm, this is mechanistically coherent: rapid dietary change and weight loss in an insulin-treated population can lower insulin requirements faster than clinicians or patients adjust dosing, raising hypoglycemia risk during the transition. It is a concrete, clinically actionable reason for caution rather than a statistical curiosity.

By 24 months, the FM arm's insulin dose trend had reversed relative to usual care, increasing by 0.05 units/kg/day where the usual care arm continued to decrease by 0.11 units/kg/day, the opposite of what the intervention was designed to achieve. The authors' own explanation is adherence collapse: participants who lost weight and reduced supplement or dietary discipline over time did not sustain the behaviours that produced the initial, non-significant improvements. In their words, "difficulty with adherence to complex regimens...remains a crucial barrier," and their overall conclusion is direct: "the addition of FM did not lead to significant improvement in glycemic, metabolic or quality-of-life outcomes compared to usual care for individuals with type 2 diabetes."

Reading This Trial Honestly: Strengths, Limits, and What It Does Not Show

This trial deserves to be taken seriously rather than dismissed, in either direction. It is a genuine randomized comparison, run by a functional medicine centre with every incentive to see its own model succeed, published with a null result and a candid discussion of why. That is a mark of scientific integrity worth acknowledging, and it is the kind of self-scrutiny naturopathic and integrative medicine has historically been criticised for lacking.

At the same time, several limits mean this single pilot trial should not be read as a definitive verdict on functional medicine as a whole. It was a single-centre pilot with a small sample (71 randomised, 47 analysed at 12 months), it was unblinded, which can affect both patient-reported outcomes and behaviour, and the population, insulin-dependent type 2 diabetes of at least a year's standing, is a more advanced and metabolically entrenched group than the newly diagnosed or pre-diabetic patients many naturopathic clinics see. A program layering multiple simultaneous elements (diet, several supplements, coaching, more frequent visits) also cannot isolate which component, if any, might work in a better-adhered or better-targeted protocol. Individual components of the stack, berberine in particular, have their own separate trial literature for glycemic control that this study does not overturn; what the trial tests is the effect of bundling many interventions into one intensive, adherence-demanding program, not the standalone evidence for each ingredient. Berberine's evidence base in the specific context of PCOS-related insulin resistance is discussed separately and is not addressed by this trial.

What the trial does show, credibly, is that simply adding more testing, more visits, more supplements, and more dietary restriction on top of standard insulin-dependent diabetes care does not reliably translate into better glycemic control or quality of life at the population level, and that the complexity of such programs carries a real adherence cost that can erode any early gains. For a condition where the underlying metabolic dysfunction is described in more mechanistic detail in our review of insulin resistance and metabolic dysfunction, this is a useful corrective against assuming that more intervention is automatically better intervention.

Practical Implications for Naturopathic Practice

For practitioners working with insulin-treated type 2 diabetes patients, this trial supports several concrete cautions rather than a wholesale change of approach. First, any dietary or supplement protocol expected to produce meaningful weight loss or metabolic change in an insulin-treated patient needs proactive insulin dose review, not a wait-and-adjust approach, given the hypoglycemia signal observed here during a comparable weight-loss period. Second, program complexity should be matched to what a given patient can plausibly sustain for the long term; a protocol that produces gains in month one but collapses by month twelve has not produced a durable clinical benefit, and the trial's own trajectory illustrates that risk concretely. Third, this is one pilot trial, not a closed question, replication in larger and more diverse populations, and dismantling studies that test individual components against the full bundle, would meaningfully sharpen what is actually working within a functional medicine protocol and what is inert or counterproductive scaffolding around it.

Clinical Summary

A 2026 Cleveland Clinic randomized pilot trial found that adding a comprehensive functional medicine program, diagnostic testing, an elimination-style diet, a multi-supplement regimen, and intensive lifestyle coaching, to usual care did not significantly improve the primary composite endpoint of insulin discontinuation with controlled HbA1c in insulin-treated type 2 diabetes, nor did it improve HbA1c, BMI, diabetes distress, or quality-of-life measures at 12 months. Early weight loss in the intervention arm was not sustained, hypoglycemia risk was higher during the period of active weight change, and by 24 months insulin requirements in the intervention arm had increased relative to usual care, a pattern the authors attribute to declining adherence to a complex regimen. The trial is a single-centre pilot with real limitations, but it is a genuine test of the functional medicine model as practised, run and published by a functional medicine centre, and its honest, unfavourable result is worth weighing seriously rather than explaining away.

References

• Eldib M, Hatipoglu B, Beidelschies M, Alejandro-Rodriguez M, Xiao H, Rothberg M, Hyman M, Zhou K. "A functional medicine approach to diabetes control and quality of life: an open-label randomized pilot study." Journal of Clinical & Translational Endocrinology. 2026;44:100437. DOI: 10.1016/j.jcte.2026.100437. PMID: 41953254. Full text: PMC13054257.

This article is intended for educational purposes and professional practice reference. It does not constitute individual medical advice. Clinical decisions, particularly any change to insulin dosing, should be made in consultation with a qualified prescribing clinician.