The research team, led by Bertin Magamba of Region Halland with colleagues from the University of Gothenburg and Lund University, tapped an unusually complete data source. The Regional Healthcare Information Platform in Halland integrates clinical, demographic, and administrative records across every level of care in a region of roughly 340,000 inhabitants served by three acute hospitals and 48 primary care units. Because the platform captures laboratory measurements regardless of where patients are treated, the investigators could classify kidney function and glycemic control for a genuinely community-dwelling cohort rather than a hospital-selected sample. Every participant was aged 75 or older, carried a documented type 2 diabetes diagnosis, and had at least one measurement of both glycated hemoglobin (HbA1c) and estimated glomerular filtration rate (eGFR) during 2019.

The technical picture that emerged was striking. Kidney function was classified by eGFR into preserved function (60 mL/min/1.73 m² or above), moderate impairment corresponding to CKD stage 3 (30–59), and severe impairment at stages 4–5 (below 30). More than half of the cohort—56 percent—already sat in stage 3, and one in ten had reached stage 4–5. Glycemic control, measured as HbA1c, was below 52 mmol/mol in 43 percent of participants, between 52 and 70 mmol/mol in 45 percent, and above 70 mmol/mol in 12 percent. Notably, kidney function declined in lockstep with blood sugar: mean eGFR fell from 53.9 to 45.7 mL/min/1.73 m² across HbA1c categories, and the share of patients with advanced kidney disease climbed from 6 percent in the best-controlled group to 21 percent in the worst.

When the researchers counted hospital days during 2019, the gradients were unmistakable. Mean inpatient days rose from 1.5 in the lowest HbA1c group to 2.9 in the highest, and from 1.6 in those with preserved kidney function to 4.6 in those at CKD stages 4–5. To quantify these associations independently, the team used Poisson regression, a statistical approach suited to count outcomes, adjusting for age, sex, comorbidities, laboratory values, and treatment variables. The results were decisive: patients with CKD stages 4–5 accumulated nearly twice as many hospital days as those with preserved kidney function (relative rate 1.93, 95 percent confidence interval 1.71–2.17), while even stage 3 disease carried a modest elevation (RR 1.15). Poor glycemic control above 70 mmol/mol independently increased hospital days by 44 percent (RR 1.44), and atherosclerotic cardiovascular disease added a further 23 percent.

These numbers tell a coherent biological story. Chronic hyperglycemia is well established as a driver of microvascular damage, progressively injuring the glomeruli of the kidneys and the vessels of the retina. As eGFR declines, cardiovascular risk compounds: fluid balance deteriorates, drug clearance becomes unpredictable, and the margin for metabolic error narrows. Large registry studies have long shown that declining kidney function amplifies mortality in diabetes, but the Swedish analysis adds a health-services dimension—each step down the renal ladder translates directly into more nights in a hospital bed, even among patients already receiving intensive outpatient follow-up.

Treatment patterns in the cohort revealed both clinical caution and a persistent gap between guidelines and practice. Insulin use rose steeply with worsening glycemic control, from 18 percent of the best-controlled patients to 75 percent of the worst, while metformin peaked in the intermediate group. Incretin-based therapies, including GLP-1 receptor agonists, increased from 14 to 51 percent across HbA1c categories, and SGLT-2 inhibitors from 2 to 12 percent. Yet despite robust trial evidence that SGLT-2 inhibitors and GLP-1 receptor agonists protect both the heart and the kidneys, their absolute use among these oldest patients remained low—particularly in those with advanced kidney disease, where insulin dominated and metformin and SGLT-2 inhibitors were scaled back. DPP-4 inhibitors, delivered as oral tablets, were comparatively popular, suggesting a prescriber preference for oral over injectable therapy in frail older adults.

The regression results on medications demand careful interpretation, and the authors are explicit about this. Insulin use was associated with more hospital days (RR 1.18), while dietary treatment, metformin, and incretin-based therapy were associated with fewer (RR 0.46, 0.62, and 0.76 respectively). But these are associations, not causal effects. Insulin-treated patients almost certainly represent a sicker subgroup with more advanced diabetes, greater comorbidity, and higher baseline healthcare needs—a textbook case of confounding by indication. Similarly, the favorable signals for metformin and incretin agents may partly reflect that these drugs are prescribed to patients well enough to tolerate them. What the data do show clearly is that treatment allocation shifts dramatically with kidney stage, and that therapeutic inertia—the reluctance to initiate proven cardiorenal therapies in the very old—remains a real phenomenon in routine care.

Perhaps the most sobering finding concerns primary care. Patients in the cohort were far from neglected: on average they saw a primary care physician 4.6 times and a nurse 9.0 times during the single study year, and nurse visits climbed to 13.3 among those with stage 4–5 kidney disease. Yet hospitalizations remained high despite this intensive outpatient engagement. The authors interpret this not as a failure of follow-up but as a signature of multimorbidity and frailty—mobility limitations, polypharmacy, cognitive impairment, and care coordination needs that extend well beyond glucose management. Primary care encounters were associated with kidney disease but not with glycemic dysregulation, hinting that the care system is responding to the complexity of the whole patient rather than the diabetes numbers alone. Evidence from structured post-discharge programs suggests that better coordination and systematic follow-up could still trim avoidable readmissions in this group.

Mortality during the year was comparatively low at 6 percent—88 individuals—but deaths clustered among those with advanced kidney disease and higher HbA1c, consistent with the direction of risk seen in longer-term studies where one-year mortality reaches 10 to 15 percent in primary-care populations and exceeds 20 percent among patients with new cardiovascular diagnoses. The study has honest limitations: requiring recorded HbA1c and eGFR may have excluded the most severely ill, laboratory values and outcomes were measured in the same calendar year so reverse causality cannot be excluded, and no data on frailty or functional status were available. Even so, the message is hard to ignore. In adults over 75 with type 2 diabetes, the kidney is the organ that most reliably forecasts the hospital ward, and age-specific evidence—long overdue—will be essential if a rapidly growing population of older adults is to avoid the admissions that currently define their care.

Subject of Research: Associations between kidney function, glycemic control, and hospitalization in adults aged 75 years and older with type 2 diabetes

Article Title: Hospitalization in adults aged 75 years and older with type 2 diabetes: the role of kidney impairment and glycemic control

Article References: Magamba, B., Jonasson, J. M., Pettersson, L., Halling, A., & Agvall, B. (2026). Hospitalization in adults aged 75 years and older with type 2 diabetes: the role of kidney impairment and glycemic control. BMC Geriatrics, 26(1), Article 1254. https://doi.org/10.1186/s12877-026-08386-7

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08386-7

Keywords: type 2 diabetes, chronic kidney disease, hospitalization, HbA1c, older adults, geriatrics, primary care, insulin, metformin, SGLT2 inhibitors, GLP-1 receptor agonists, healthcare utilization