A large-scale population-based study revealed a small but significant association between cumulative surgical procedures, cognitive decline, and neurodegeneration in older adults. While the observed effects were small, the findings underscored the importance of considering brain health in surgical planning among older patients undergoing multiple procedures.

In the study, published in The Lancet Healthy Longevity, investigators used UK Biobank data to analyze the outcomes of 492,802 participants aged 40 to 69 years at recruitment, with follow-up periods spanning up to 19 years.

The study linked UK Biobank data to the National Health Service Hospital Episode Statistics database. Participants with dementia diagnoses were excluded. Fully adjusted models included age, time, sex, Lancet Commission dementia risk factors, stroke, and hospital admissions, with a participant random effect.

The primary outcomes of the study were hippocampal volume and white matter hyperintensities, measured using magnetic resonance imaging (MRI). Cognitive outcomes included reaction time, fluid intelligence, prospective memory, and numeric memory. Surgeries were calculated cumulatively starting from 8 years prior to the baseline evaluation.

Linear mixed effects regression models were constructed for cognitive outcomes and imaging variables. The foundational model included surgeries, age, age², time, and sex. The fully adjusted model added stroke admissions, other medical admissions, deprivation, education, body mass index (BMI), smoking, depression, alcohol consumption, physical activity, hypertension, diabetes, and solid tumor history.

Among 502,412 UK Biobank participants, 492,802 were included in the study, with 46,706 undergoing MRI. The mean age at baseline was 56 years (standard deviation [SD] = 8), and 54.5% were female. The ethnic background was predominantly White (94.1%).

Of the initial 492,802 participants, 71,873 (14.6%) returned at instance 2, and 6,324 (1.3%) returned at instance 3. Among instance 2 participants, 46,706 (65.0%) had a brain MRI scan, with 4,779 (10.2%) having a follow-up MRI scan at instance 3.

Baseline mean cognitive scores were:

• Reaction time: 559 ms (SD = 117).
• Fluid intelligence score: 5.99 (SD = 2.16).
• Numeric memory score: 6.49 (SD = 1.82).
• Prospective memory (correct first time): 76.6%.

Cognitive impact:

• Per surgical procedure, reaction time increased by 0.273 ms (95% confidence interval [CI] = 0.022–0.523, P = .033).
• Fluid intelligence score decreased by 0.057 correct responses (95% CI = –0.064 to –0.049, P < .0001).
• Odds of correct first-attempt matches in prospective memory tests decreased (odds ratio [OR] = 0.96, 95% CI = 0.95–0.97, P < .0001).
• Numeric memory maximum correct matches decreased by 0.025 (95% CI = –0.034 to –0.016, P < .0001).

Neuroimaging results:

• Surgeries were associated with smaller hippocampal volume (β = –5.76 mm³, 95% CI = –7.89 to –3.64, P < .0001).
• Greater white matter hyperintensities volume was observed (β = 100.02 mm³, 95% CI = 66.17–133.87, P < .0001).
• Neurodegeneration of the insula and superior temporal cortex was noted in exploratory analyses.

Mean brain volumes at instance 2 were:

• White matter hyperintensities: 5,216 mm³ (SD = 6,877).
• Hippocampus: 3,818 mm³ (SD = 447).

Cumulative and complexity effects:

• Increased number of surgeries showed stronger associations (β = 2.53 for 1 to 2 surgeries, β = 7.63 for 3 to 5 surgeries, and β = 12.46 for ≥ 6 surgeries; all P < .0001).
• High-complexity surgeries had larger effects (e.g., β = 2.97 for reaction time, P < .0001).

Surgical exposure:

• Mean number of cumulative surgeries at baseline: 0.84 (SD = 1.30).
• 55.5% had 0 surgeries, 34.8% had 1 to 2 surgeries, 8.5% had 3 to 5 surgeries, and 1.2% had ≥ 6 surgeries at baseline.

High-complexity surgeries were associated with greater changes:

• White matter hyperintensities increased by 247.73 mm³ (95% CI = 161.70–333.76, P < .0001).
• Average hippocampal volume reduced by 17.05 mm³ (95% CI = –22.46 to –11.64, P < .0001).

Surgery was associated with reductions in mean cortical thickness in 23 (65%) of 31 Desikan-Killiany-Tourville (DKT) regions. After adjustment, significant neurodegeneration was observed in the insula (β = –0.001 mm) and superior temporal (β = –0.001 mm) cortical regions.

Stroke admissions were associated with greater changes in white matter hyperintensities (β = 6,271.29 mm³) and average hippocampal volume reduction (β = –161.29 mm³).

The study had several limitations, including the inability to exclude unmeasured confounders or ascribe causality, potential bias from substantial loss to follow-up, and the introduction of imaging only in later instances of data collection. The classification of surgical severity using health insurance risk codes and main admissions diagnosis may not fully reflect the complexity of procedures and postoperative complications.

Conflict of interest disclosures can be found in the study.