This cross-sectional study encompassed 112 WMH patients. All patients fulfilled the ethical standards set by the Ethics Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University. Furthermore, all participants voluntarily signed the Informed Consent Form. The subjects comprised adult patients admitted to the Department of Neurology, Shi Zhong District People’s Hospital, Jinan, Shandong Province, between January and December 2023. Upon admission, fundamental subject data was gathered, encompassing name, age, gender, body mass index (BMI), blood pressure (BP), and a comprehensive history of prior health conditions, such as hypertension (HP), diabetes mellitus (DM), coronary heart disease (CHD), hyperlipidemia (HLP), and glomerular filtration rate (eGFR). Additionally, the smoking and drinking patterns of the patients were meticulously evaluated. The inclusion criteria stringently stipulated an age threshold of 18 years, confirmation of WMH via FLAIR, and consistent Fazekas scale scores. Conversely, exclusion criteria excluded individuals with severe organ insufficiency, alternative forms of cognitive decline, and a history of severe cardiovascular or cerebrovascular disorders. From eligible patients, blood plasma samples were collected early in the morning after admission, and subsequently stored in a −80°C refrigerator for further analysis.

All participants were scanned using a 3.0 Tesla brain MR scanner (Philips, Achieva) with a 32-channel head coil, encompassing axial T1-weighted imaging (T1WI), axial T2-weighted imaging (T2WI), and FLAIR sequence imaging with a slice thickness of 3 mm. The FLAIR imaging is a 3D sequence with multiplanar reformations. All MR scanners and Fazekas score assessment were done by 2–3 neuroradiologists, ensuring the evaluators were blinded to the participants’ basic information. The diagnostic criteria for WMH serve as a reference and involve the identification of scattered dot-like, block-like, or large-scale high-signal intensities located around or within the bilateral regions of the brain, based on intracranial MRI examination findings.

The categorization of WMH has been previously detailed in other clinical trials (Yu et al., 2021), our study employed comparable assessment techniques to categorize overall Fazekas scores into three groups: non-mild WMH (Fazekas score 0–2), moderate WMH (Fazekas score 3–4), and severe WMH (Fazekas score 5–6). Based on the severity of P-WMH and D-WMH, they were further classified as non-mild (Fazekas score 0–1), moderate (Fazekas score 2), and severe (Fazekas score 3). The Fazekas scores of P-WMH and D-WMH were summed up to derive an overall Fazekas score, ranging from 0 to 6.

The concentrations of plasma TMAO and MMP-9 were quantitatively determined utilizing the Enzyme-linked immunosorbent assay (ELISA) technique. To summarize, the process entailed pre-coating a specific antibody, subsequently adding the sample and horseradish peroxidase (HRP), with 3,3′, 5,5′-tetramethylbenzidine (TMB) serving as the substrate. The absorbance (OD value) was then measured at 450 nm, utilizing a standard curve for comparison.

Categorical variables were represented using proportions, while continuous variables with a normal distribution were characterized by the mean ± standard deviation. For continuous variables with a non-normal distribution, the median (interquartile range) was utilized. The statistical evaluations were executed utilizing SPSS 25.0, GraphPad Prism 10, and R Studio. Comparisons among continuous variables were facilitated by the ANOVA, Kruskal-Wallis test, or Mann–Whitney U test, as applicable. Categorical variables were scrutinized using the Pearson Chi-square test. Furthermore, three logistic regression models were devised to gauge the risk of WMH in relation to TMAO. The basic model entailed no adjustments, model a adjusted for gender and pivotal cardiovascular risk factors (HP, DM, CHD, HLP), and model b further considered BMI > 24 kg/m² and eGFR<90 mL/min. In the regression analysis, the OR value and its 95% CI were computed. The Spearman rank correlation was employed to ascertain the link between plasma TMAO levels and Fazekas scores. The receiver operating characteristic (ROC) curves for non-mild, moderate, and severe WMH, P-WMH, and D-WMH were generated using R Studio software. The area under the ROC curve (AUC) was calculated to forecast the severity of WMH by evaluating plasma TMAO levels. All tests were conducted as two-sided, and statistical significance was deemed at p < 0.05.

A total of 112 patients were enrolled in our study, with a male predominance of 77.0% and a median age of 65 years. During the period spanning from December 2023 to January 2024, comprehensive tests were conducted, encompassing plasma MMP-9 levels, BMI, BP measurements, and detailed blood analyses. The latter encompassed various blood components such as white blood cells, red blood cells, hemoglobin, platelets, and lymphocytes, as well as assessments of renal function (specifically, blood urea nitrogen and eGFR), lipid profiles (including cholesterol, triglycerides, HDL, LDL, homocysteine), glycemic markers (fasting blood glucose, glycated hemoglobin), and thyroid function indicators (free triiodothyronine, free thyroxine, and thyroid stimulating hormone).

The mean concentrations of plasma MMP-9 and TMAO among the participants were 20.88 ng/mL and 25.03 ppm, respectively. For TMAO levels, the cohort was stratified into three quartiles: a low group with TMAO concentrations ≤21.56 ppm (n = 38), a middle group with concentrations ranging from 21.56 ppm to 25.16 ppm (n = 37), and a high group with concentrations ≥25.16 ppm (n = 37). Notably, no statistically significant differences were observed among these groups in terms of gender distribution, HP status, DM prevalence, or the presence of CVD.

There were 36 patients with non-mild overall WMH, 50 patients with moderate overall WMH, and 26 patients with severe overall WMH. The severity of overall WMH was notably correlated with plasma TMAO levels in patients, with statistical significance at p < 0.001. Furthermore, the severity of overall WMH was linked to elevated plasma MMP-9 levels and age, with a significance level of p < 0.05. The plasma TMAO levels were significantly higher in patients with moderate WMH compared to those with non-mild WMH, as indicated by the median values of 23.48 (IQR 21.71–25.25) ppm versus 20.41 (IQR 17.55–22.51) ppm, respectively, at p < 0.001. Similarly, patients with severe WMH exhibited higher plasma TMAO levels than those with non-mild WMH, with median values of 27.7 (IQR 25.33–30.66) ppm versus 20.41 (IQR 17.55–22.51) ppm, respectively, at p < 0.001. Additionally, patients with severe WMH had significantly higher plasma TMAO levels compared to those with moderate WMH, with median values of 27.78 (IQR 25.33–30.66) ppm versus 23.48 (IQR 21.71–25.25) ppm, respectively, at p < 0.001 (Figure 1).

Furthermore, a significant positive correlation was observed between the plasma TMAO levels and the overall Fazekas score, with a correlation coefficient of r = 0.9318 and a significance level of p = 0.002 (Figure 2). There was also a notable positive correlation between plasma TMAO levels and MMP-9 levels, with a correlation coefficient of r = 0.5012 and a significance level of p < 0.001 (Figure 3).

The findings of the logistic regression analysis, examining the link between plasma TMAO levels and the severity of overall WMH, are presented in Table 1 and Figures 4, 5. Essentially, in the unadjusted model (comprising plasma TMAO levels, plasma MMP-9 levels, TG, and age), a marked association was observed between plasma TMAO levels and moderate to severe WMH, as opposed to non-mild WMH (moderate WMH: OR = 1.373, 95% CI: 1.183–1.594; severe WMH: OR = 1.384, 95% CI: 1.192–1.607). Similarly, age independently correlated with WMH severity (moderate WMH: OR = 1.072, 95% CI: 1.026–1.120; severe WMH: OR = 1.116, 95% CI: 1.062–1.173).

In the adjusted model a, where gender and cerebrovascular risk factors (HP, DM, CHD, HLP) were taken into account, plasma TMAO levels retained its significant association with WMH severity (moderate WMH: OR = 1.436, 95% CI: 1.214–1.669; severe WMH: OR = 1.446, 95% CI: 1.222–1.711), and the correlation with age persisted (moderate WMH: OR = 1.075, 95% CI: 1.023–1.129; severe WMH: OR = 1.123, 95% CI: 1.063–1.187).

Furthermore, in model b, which additionally considered BMI > 24 kg/m² and eGFR<90 mL/min, the independent relationship between plasma TMAO levels and moderate to severe WMH remained statistically significant (moderate WMH: OR = 1.415, 95% CI: 1.171–1.711; severe WMH: OR = 1.422, 95% CI: 1.175–1.721). However, the association of age with severe WMH in this model was noted with a minor adjustment in the CI range (OR = 1.016, 95% CI: 1.020–1.199).

To delve deeper into the correlation between plasma TMAO levels and the severity of WMH in distinct regions, we segregated the study participants into two groups: P-WMH and D-WMH. Specifically, the P-WMH group comprised 30 non-mild, 56 moderate, and 26 severe subjects. Notably, individuals with moderate and severe P-WMH exhibited a greater tendency to have elevated levels of age and SP in comparison to those with non-mild P-WMH, with statistical significance (p < 0.05). Shifting focus to the D-WMH classification, we identified 56 non-mild, 46 moderate, and 10 severe subjects. Our analysis revealed a marked association between plasma MMP-9 levels and D-WMH severity, again with statistical significance (p < 0.05). Furthermore, patients with moderate and severe D-WMH were significantly more likely to be of advanced age compared to those with non-mild D-WMH, achieving a higher level of statistical significance (p < 0.005).

The severity of WMH, both P-WMH and D-WMH, positively correlates with the levels of TMAO in plasma. Specifically, the plasma TMAO levels were significantly higher in patients with moderate P-WMH compared to those in the non-mild P-WMH group [median 23.04 (IQR 21.23–25.86) ppm versus median 20.57 (IQR 17.29–22.58) ppm, p < 0.05]. Furthermore, patients with severe P-WMH exhibited even higher plasma TMAO levels than those with non-mild P-WMH [median 27.78 (IQR 25.33–30.66) ppm versus median 20.57 (IQR 17.29–22.58) ppm, p < 0.0001], and the same trend was observed when comparing severe P-WMH patients to those with moderate P-WMH [median 27.78 (IQR 25.33–30.66) ppm versus median 23.04 (IQR 21.23–25.86) ppm, p < 0.01] (Figure 6). Similarly, the plasma TMAO levels were significantly elevated in patients with moderate D-WMH compared to the non-mild D-WMH group [median 25.00 (IQR 22.13–28.19) ppm versus median 21.46 (IQR 18.62–22.91) ppm, p < 0.0001]. This trend continued in patients with severe D-WMH, who showed even higher plasma TMAO levels than those with non-mild D-WMH [median 27.78 (IQR 25.33–30.97) ppm versus median 21.46 (IQR 18.62–22.91) ppm, p < 0.0001]. However, no statistically significant correlation was found between the plasma TMAO levels in patients with severe D-WMH and those with moderate WMH (Figure 7).

The results of the logistic regression analysis examining the relationship between plasma TMAO levels and the severity of P-WMH are depicted in Figures 8, 9. In the unadjusted model, plasma TMAO levels were significantly and independently associated with both moderate (OR = 1.295, 95% CI: 1.095–1.531) and severe P-WMH (OR = 1.307, 95% CI: 1.104–1.547). Furthermore, age was exclusively and independently associated with severe P-WMH (OR = 1.092, 95% CI: 1.024–1.164). Upon adjustment in model a, taking into account sex (male) and cerebrovascular risk factors including HP, DM, CHD, and HLP, plasma TMAO levels retained their independent association with moderate (OR = 1.330, 95% CI: 1.107–1.538) and severe P-WMH (OR = 1.343, 95% CI: 1.118–1.614). Age also remained independently associated with severe P-WMH (OR = 1.106, 95% CI: 1.030–1.188). In the second adjusted model b, similar findings were observed, with plasma TMAO levels independently associated with moderate (OR = 1.336, 95% CI: 1.109–1.609) and severe P-WMH (OR = 1.349, 95% CI: 1.119–1.626) in patients. Age, once again, was independently associated with severe P-WMH in this group (OR = 1.088, 95% CI: 1.006–1.177).

The results of the logistic regression analysis examining the relationship between plasma TMAO levels and the severity of D-WMH are depicted in Figures 10, 11. In the unadjusted model, it was found that plasma TMAO levels were independently correlated with moderate D-WMH (OR = 1.352, 95%CI 1.170–1.562) and severe D-WMH (OR = 1.357, 95%CI 1.170–1.573). Additionally, age emerged as an independent predictor for severe D-WMH (OR = 1.108, 95%CI 1.019–1.204). Upon adjustment in model a, which accounted for gender (male) and cerebrovascular risk factors (HP, DM, CHD, HLP), plasma TMAO levels remained independently associated with moderate D-WMH (OR = 1.419, 95%CI 1.209–1.665) and severe D-WMH (OR = 1.419, 95%CI 1.207–1.672). Age also continued to be independently associated with severe D-WMH (OR = 1.108, 95%CI 1.010–1.216). In the adjusted model b, which further controlled for BMI > 24 kg/m²and eGFR<90 mL/min, plasma TMAO levels were independently associated with both moderate D-WMH (OR = 1.468, 95%CI 1.235–1.744) and severe D-WMH (OR = 1.469, 95%CI 1.233–1.751). Furthermore, BMI > 24 kg/m² and eGFR<90 mL/min were found to be independently associated with moderate D-WMH (BMI: OR = 3.236, 95%CI 1.037–10.096; eGFR: OR = 4.360, 95%CI 1.135–16.752).

The ROC curve was employed to assess the diagnostic efficacy of plasma TMAO levels in discerning the severity of WMH. The ROC analysis for overall WMH revealed an AUC of 0.81 for both non-mild and severe WMH, with a significance level of p < 0.05. Simultaneously, the AUC for moderate cases was 0.55, indicating a lack of statistical significance at p = 0.421 (Figure 12A). For P-WMH, the ROC curve yielded an AUC of 0.77 for non-mild and 0.81 for severe WMH, both with p < 0.05. The AUC for moderate cases was 0.55, which was not statistically significant at p = 0.421 (Figure 12B). In the case of D-WMH, the ROC curve demonstrated AUC values of 0.81, 0.72, and 0.72 for non-mild, moderate, and severe subjects, respectively, all with p < 0.001 (Figure 12C).

Our study has deepened our understanding of the relationship between TMAO and WMH; however, our analysis suggests several areas that warrant further exploration. Initially, although an association between TMAO and WMH has been established, the development and progression of WMH is a complex, multifactorial process that involves BBB dysfunction, neuroinflammatory responses, and other mechanisms that are not yet fully understood. Future studies should adopt a longitudinal design to further investigate whether there is a causal relationship between TMAO and the progression of WMH, and to explore the specific molecular pathways involved. Secondly, the sample size of this study was relatively small and did not include TMAO analysis in cerebrospinal fluid and feces, which limits a comprehensive understanding of TMAO’s role in WMH. Considering CSF samples can directly reflect the biochemical environment within the brain, and fecal samples can provide insights into gut microbial composition, the examination of these specimens may be crucial in understanding the biological effects of TMAO. Additionally, there may be significant inter-individual variations in gut microbial composition and hepatic metabolic capacity, which could introduce greater variability in outcomes when TMA or TMAO is measured independently. The TMA/TMAO ratio might be a better predictor of disease risk development compared to TMAO alone. Pursuing these research directions will be essential in identifying potential intervention targets for WMH and in revealing the mechanisms by which TMAO contributes to the onset and progression of WMH.