Emerging evidence suggests that measuring serum magnesium levels could improve early detection of diabetic retinopathy when combined with traditional markers like HbA1c, offering a simple aid in risk stratification.
A study published in BMC Ophthalmology on 22 June 2026 suggests that serum magnesium could help sharpen how doctors identify diabetic retinopathy, after researchers found that magnesium levels fell in step from people without diabetes to people with diabetes and then to those with retinal disease. The Chinese team, led by Yicang Wang and colleagues, analysed 450 patients split evenly across non-diabetic, diabetic and diabetic retinopathy groups, then paired the clinical work with experiments in diabetic mice.
The main signal was not that magnesium outperformed established diabetes markers on its own, but that it added a little to them. In the hospital cohort, HbA1c rose as magnesium and sodium fell across the three groups, and among people already living with diabetes, longer disease duration, higher systolic blood pressure and higher HbA1c were each linked to greater odds of retinopathy. When the researchers combined HbA1c with magnesium, the model detected diabetic retinopathy with a sensitivity of 75.3% and a specificity of 66.0%, only a modest gain over HbA1c alone, but enough for the authors to argue that the marker deserves attention. In the mouse work, diabetic retinopathy was also accompanied by lower magnesium in both serum and retinal tissue, alongside structural retinal changes.
That pattern has shown up before in smaller clinical studies. A 2021 paper in Therapeutic Advances in Ophthalmology examined 104 people with type 2 diabetes and normal renal function and found lower mean serum magnesium in patients with retinopathy than in those without it, at 1.63 mg/dL versus 1.76 mg/dL. Patients with sight-threatening diabetic retinopathy had the lowest levels, averaging 1.55 mg/dL. The same study also linked retinopathy to younger age, longer diabetes duration and poorer glycaemic control, underscoring that magnesium is appearing within a broader cluster of risk rather than in isolation.
Evidence from a much larger US dataset points in a similar direction, while also suggesting that magnesium may matter most when blood sugar control is poor. In Science Progress, researchers using NHANES data from 2005 to 2018 analysed 4,308 adults with diabetes and found that those with retinopathy reported lower magnesium intake than those without it, 259.1 mg a day against 269.8 mg. Their most striking result came from looking at magnesium depletion together with HbA1c: compared with participants whose HbA1c was below 7% and whose magnesium depletion score was low, those with HbA1c of 7% or more and a depletion score of at least 2 had markedly higher odds of retinopathy. That helps explain why the new BMC Ophthalmology paper focused on a combined magnesium-and-HbA1c approach rather than presenting magnesium as a stand-alone test.
Other mineral studies have reached related conclusions while widening the frame beyond glucose control. In Endocrine Practice, a Shanghai study combining cross-sectional and longitudinal analyses looked at 3,269 patients at baseline and followed 649 people without vision-threatening diabetic retinopathy for a median of 9.7 years. It reported that lower serum magnesium sat alongside other independent risk factors including higher corrected calcium, younger age, longer diabetes duration, albuminuria and impaired renal function. Over follow-up, 95 participants developed vision-threatening disease, and higher glycated haemoglobin was again among the predictors. That is a reminder that retinal damage in diabetes is driven by overlapping metabolic, vascular and renal pathways, not by any single blood test.
The broadest synthesis so far also leans in the same direction, albeit cautiously. An updated systematic review and meta-analysis published in 2026 pooled 17 studies covering 1,100 patients with diabetic retinopathy and 1,132 diabetic controls without it. The review found that circulating magnesium levels were consistently lower in retinopathy cases across Asian and non-Asian settings and in both cross-sectional and case-control designs. But the authors graded the certainty of the evidence as low, largely because the underlying studies were observational and highly heterogeneous.
That caveat matters. The new BMC Ophthalmology paper was retrospective, single-centre and cross-sectional, with no follow-up to show whether low magnesium comes before retinal damage or results from the same disease processes that produce it. It also lacked treatment-status data and blinded assessment. The national US analysis and the meta-analysis cannot settle causation either. So the practical takeaway is narrower than some headlines might suggest: this is not yet proof that magnesium supplements prevent eye disease in diabetes. It is, however, a growing body of evidence that a cheap and familiar biomarker may help refine risk assessment when read alongside HbA1c. As the BMC Ophthalmology authors put it, “Combined detection of serum magnesium and HbA1c can significantly improve the diagnostic performance of DR.”
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
