Background Environmental metal exposure is closely associated with the onset and progression of mild cognitive impairment (MCI) in the elderly. Effectively identifying hazardous metal exposure and assessing their interaction effects have significant public health implications.

Objective To explore the relationship between urinary single metal and metal mixture exposure and MCI in elderly compound residents.

Methods This study included 391 elderly individuals aged 60 and above from residential compounds in Zhongshan City, Guangdong Province. Concentrations of iron (Fe), copper (Cu), selenium (Se), arsenic (As), cadmium (Cd), manganese (Mn), chromium (Cr), nickel (Ni), vanadium (V), cobalt (Co), antimony (Sb), thallium (Tl), zinc (Zn), calcium (Ca), and magnesium (Mg) in urine were measured using inductively coupled plasma mass spectrometry. Cognitive function in the elderly was assessed using the Chinese version of the Mini-Mental State Examination (MMSE). Logistic regression was used to explore the relationship between single metal exposure level and MCI. LASSO regression and multi-metal logistic regression models were used to identify key metal ions associated with MCI. Bayesian kernel machine regression (BKMR) was employed to analyze the relationship between key metal ion mixtures and MCI, as well as the interactions between metals. Age, gender, education level, occupation, and body mass index were adjusted as covariates.

Results A total of 78 among the 391 elderly individuals surveyed (19.94%) were diagnosed with MCI (MCI group), and the other 313 individuals were controls. The levels of Se, Cd, Mn, and As in the urine of the MCI group were significantly higher than those in the control group (P < 0.05). In the single-metal model, after adjusting for covariates and using the first quartile (Q₁) of each metal concentration as the reference, the OR for MCI in the elderly in the Q₄ group of Se was 2.190 (95%CI: 1.017, 4.716); for Cd, the OR was 2.345 (95%CI: 1.041, 5.283) in the Q₃ group and 2.371 (95%CI: 1.043, 5.393) in the Q₄ group; for Mn, the OR was 2.355 (95%CI: 1.038, 5.344) in the Q₂ group; for As, the OR was 3.377 (95%CI: 1.442, 7.908) in the Q₃ group and 2.886 (95%CI: 1.227, 6.788) in the Q₄ group; for Sb, the OR was 2.779 (95%CI: 1.234, 6.257) in the Q₂ group. When urinary metal concentrations were ln-transformed and included as continuous variables in the single-metal model, Cd concentration was positively correlated with MCI (OR=1.377; 95%CI: 1.008, 1.882; P=0.044). Cd, Se, Mg, Ca, Mn, As, Cr, Co, Tl, and Sb were selected by the LASSO regression model and included in the multi-metal model. In the multi-metal model, compared with Q₁, the OR for MCI in the elderly was 0.395 (95%CI: 0.164, 0.953) in the Q₂ group of Co and 0.390(95%CI: 0.167, 0.911) in the Q₃ group of Co; for Mn, the OR in the Q₂ group was 2.636 (95%CI: 1.053, 6.596); for Sb, the OR in the Q₂ group was 2.640 (95%CI: 1.047, 6.658). As continuous variables, Mg (OR=0.472; 95%CI: 0.248, 0.899; P=0.022) and Co (OR=0.857; 95%CI: 0.737, 0.996; P=0.044) concentrations were negatively correlated with MCI. The BKMR mixture analysis suggested that Mg and Co exhibited a synergistic negative correlation with MCI, while Mn and Sb exhibited a synergistic positive correlation with MCI. Mg and Co attenuated the positive correlation of Mn and Sb with MCI, whereas Mn weakened the protective effects of Mg and Co.

Conclusion Elevated levels of Se, Cd, As, Mn, and Sb in urine may increase the risk of MCI in the elderly, while Mg and Co have protective effects. Potential synergistic or antagonistic interactions may be found among Mn, Sb, Mg, and Co, which should not be overlooked in terms of their impact on the cognitive function of the elderly.