Links Between Mediterranean Diet and Blood Vessel Health in People with Long COVID

The present study is a cross-sectional investigation, and the data were collected from subjects from the BioICOPER study. The work was conducted at the Primary Care Research Unit of Salamanca (APISAL). The protocol was published in 2023 [28]. This project was registered at Clinical Trials.gov in April 2023 (register number: NCT05819840).

Data from 304 subjects diagnosed with long COVID (LC) were analysed, obtained from primary care records and the LC consultation data of the Internal Medicine Department at Salamanca University Hospital. The diagnostic criteria for LC defined by the World Health Organization (WHO) were utilised as inclusion criteria [29]: subjects are considered to have LC if they have symptoms for more than 2 months, which appear within 3 months of COVID-19 infection and cannot be explained by any other cause. Symptoms may persist from the acute phase, reappear after initial improvement, or fluctuate during the course of the disease [29]. Exclusion criteria were subjects being unable to attend APISAL due to their health status, or having established cardiovascular disease, or an estimated glomerular filtration rate of less than 30 mL/min/1.73 m² [28]. The flow chart (Figure 1) details the subjects who participated in this study, as well as the causes for exclusion.

Four trained healthcare professionals conducted tests, examinations, and analysis of the variables, following a standardised protocol [28]. The information was recorded in the REDCAP system and data quality control was performed by an independent researcher.

The mean values of the quantitative variables were compared with two categories by utilising Student's t-test. The Chi-square test was employed to compare categorical variables. Logistic regression models were used to analyse the association between MD and its components with the parameters of vascular structure, function, and vascular aging. The dependent variables were cIMT, cfPWV, baPWV, CAVI, and VAI, classified into two categories using the median value (encoded as 0 = value less than the median; 1 = value greater than the median). The independent variables were adherence to MD and compliance or non-compliance with each of its components (encoded as 0 = value less than the median; 1 = value greater than the median). Regarding the adjustment variables, two models were utilised. The first model was used in the global analyses, which included age in years, and encoded sex (0 = male and 1 = female); duration of LC in months; use of antihypertensive, lipid-lowering, and/or antidiabetic drugs, coded as (0 = no, 1 = yes); cardiovascular risk factors that define MetS (encoded as the number of MetS components); and the following lifestyle factors: alcohol consumption in grams, average steps per day, and tobacco, coded as (0 = no, 1 = yes). The second model was used for the sex-specific analysis and contained the same adjustment variables, with the exception of sex. SPSS for Windows, v28.0 (IBM Corp, Armonk, NY, USA) was used for the analyses, and the cut-off point for statistical significance was p < 0.05.

The study was approved on 27 June 2022 (CEIm reference code: Ref. PI 2022 06 1048) by the "Ethics Committee for Research with medicines in the Salamanca Health Area". The standards of good practice in observational studies established by the Declaration of Helsinki [40] and by the WHO were followed throughout the study. The confidentiality of participants was always guaranteed in accordance with Organic Law 3/2018, European Regulation 2016/679, and Council Directive 27/04/2016 on Data Protection. All participants filled out an informed consent form before being recruited for the study, after receiving information about the procedures that would be performed on them.

Table 1 shows the general features of the participants included in this study, both global and by sex. More women than men were included in the study (207 versus 97). Men had a higher mean age than women (55.70 ± 12.28 years versus 51.32 ± 11.54 years). Men had higher alcohol consumption, blood pressure, glucose in blood, triglycerides, BMI, and waste circumference, and a higher percentage of diagnoses of MetS. Women demonstrated higher HDL cholesterol levels than men. The time span from the beginning of SARS-CoV-2 acute infection diagnosis until study inclusion was 38.66 ± 9.58 months, with no differences between the sexes.

Table 2 shows the mean values for vascular structure, function, and aging in the global analysis and by sex. Vascular structure and aging values were higher in men than in women.

Table 3 presents a comparative analysis of the variables examined in relation to adherence to the MD. Participants with high adherence to MD were older and had higher DBP and lower triglycerides.

Table 4 shows the adherence to each MD component in the global analysis and by sex. Three components demonstrated a compliance percentage higher than 90%: use of olive oil as the principal source of fat for cooking; consumption of <1 serving of butter, margarine, or cream per day; and consumption of <1 carbonated and/or sugar-sweetened beverage per day. Men had a higher compliance percentage for weekly consumption of ≥7 cups of wine (100 mL); ≥3 portions of pulses; and ≥2 portions of boiled vegetables, pasta, rice, or other dishes with a sauce of tomato, garlic, onion, or leeks sautéed in olive oil. Women had a higher compliance percentage for intake of ≥2 servings of vegetables per day and eating white meat instead of red meat.

Figure 2 shows the compliance percentages for the 14 items of the MD adherence questionnaire according to the degree of adherence. All components had a higher percentage in the high-adherence-to-MD group (p < 0.005).

Figure 3 reveals global OR results for both sexes concerning MD score and its components in relation to vascular structure, assessed with cIMT; the parameters of vascular function evaluated with cfPWV, baPWV and CAVI; and VAI. Figure 4 and Figure 5 show these results separated by sex, with male and female results displayed separately. MD score and 14 MEDAS components were used as independent variables. cIMT, cfPWV, baPWV, CAVI, and VAI were used as dependent variables, with the results displayed in the subplots corresponding to each variable in the figures: (a) cIMT, (b) cfPWV, (c) baPWV, (d) CAVI, and (e) VAI. Adjustment variables were age; sex; LC development time; antihypertensive, antidiabetic, and/or hypolipidemic drugs; and cardiovascular risk factors that define MetS and lifestyles: alcohol consumption (g), mean steps per day, and tobacco use.

As we can see in Figure 3a and Figure 4a, an association was found between cIMT < 0.625 mm and using olive oil as the main cooking fat in both global (OR = 0.148; 95%CI 0.043–0.509) and men (OR = 0.120; 95%CI 0.021–0.697). In addition, an association was identified between <2 commercial pastries consumption per week and cIMT < 0.625 mm in the global analysis (OR = 0.536; 95%CI 0.294–0.978).

Referring to Figure 3b and Figure 5b, cfPWV < 7.400 m/s was associated with a MD score ≥ 8 in global (OR = 0.444; 95%CI 0.253–0.780) and women (OR = 0.363; 95%CI 0.181–0.726). Likewise, cfPWV < 7.400 m/s was associated with consumption of <2 commercial pastries per week in women (OR = 0.405; 95%CI 0.198–0.829).

Furthermore, Figure 3c and Figure 5c show a correlation between baPWV < 12.315 m/s and ≥3 servings of pulses per week in both global (OR = 0.481; 95%CI 0.250–0.940), and women (OR = 0.428; 95%CI 0.187–0.977). In addition, baPWV < 12.315 m/s was associated with <2 commercial pastries consumption per week in global (OR = 0.471; 95%CI 0.246–0.902), as well as among women (OR = 0.322; 95%CI 0.140–0.742).

CAVI < 7.450 showed association with ≥4 tablespoons of olive oil per day in men (OR = 0.231; 95%CI 0.047–0.970), which can be observed in Figure 4d).

Finally, Figure 3e and Figure 5e display an association between VAI < 63.693 and MD ≥ 8 in global (OR = 0.458; 95%CI 0.243–0.863) and among women (OR = 0.243; 95%CI 0.107–0.550). Additionally, VAI < 63.693 was associated with <2 commercial pastries consumption per week in global (OR = 0.392; 95%CI 0.308–0.972).

The results of this study showed that a higher adherence to the MD (score ≥ 8) was associated with lower carotid-femoral pulse wave velocity (cfPWV) and lower Vascular Aging Index (VAI), both in the global analysis and among women, which could indicate a higher sensitivity of the female vascular system to the benefits of diet in people with LC. We also found an OR < 1 with the other vascular parameters analysed, although these differences were not statistically significant. These findings are consistent with those published in cross-sectional studies by other authors in both the general population and in populations with cardiovascular risk factors.

Several authors have shown a negative association between adherence to the MD and vascular function parameters such as cfPWV and baPWV [22,41]. Regression analysis was used to evaluate the association between different parameters of vascular structure and arterial stiffness (AS) and adherence to MD, finding negative associations; however, it should be noted that in several cases, these associations did not reach statistical significance.

Regarding longitudinal studies, some studies have analysed patterns between adherence to MD in different populations (Spanish [42] and Dutch [43] populations) and parameters of vascular function and AS. All these studies were compiled in a recent meta-analysis published in 2025 [27], and they revealed that people adhering to a MD plan may exhibit reduced AS and, therefore, a diminished cardiovascular risk. This systematic review demonstrated an association between adherence to a MD pattern, evaluated using validated adherence scores or scales, and central AS (assessed with cfPWV), peripheral AS (evaluated with baPWV), and central and peripheral AS (measured with CAVI). It also showed an inverse relationship, which remained constant across all the included studies regardless of the outcome measure used to assess AS [27]. All these studies suggest that the beneficial effects of MD on AS are the result of a combination of mechanisms, including anti-inflammatory effects, antioxidant properties, improved lipid profiles, blood pressure regulation, and endothelial function [27].

Likewise, VAI, a vascular aging indicator that integrates cIMT and cfPWV, showed a negative association with adherence to the MD, particularly in women. These results are in line with previous publications that have analysed this relationship in subjects from populations without previous diseases [44,45]. These studies showed that subjects classified as having healthy vascular aging were more likely to adhere to the MD than subjects classified as having accelerated vascular aging. This reinforces the idea that an anti-inflammatory diet can mitigate the process of accelerated vascular aging that has been observed in individuals with LC [2].

A relevant finding of this study is the identification of differences in association according to sex. In women, the associations between adherence to the MD and vascular markers were more consistent and significant. In men, only some individual components of the MD showed relevant relationships (e.g., use of olive oil with lower cIMT or ≥4 tablespoons of olive oil per day with lower CAVI). These findings could be related to the greater vascular vulnerability observed in women after COVID-19 in other studies, characterised by lower arterial elasticity and a more altered vascular response. Some studies have shown that women who had survived COVID-19 presented higher values in AS and inflammatory parameters, as well as higher markers of inflammation. Consequently, these results were associated with greater AS [7,8,46]. In addition, hormonal and non-hormonal factors affect gender differences. It is well established that endogenous oestrogens provide a protective effect in women until the onset of menopause. In contrast, in men, arterial stiffness increases linearly from puberty onwards, indicating that women inherently have increased main AS compared to men. These effects are lessened by sex steroids during the reproductive phase of life. Further factors that may be pertinent to this issue include height, body fat distribution, and inflammatory markers [47,48]. Furthermore, the findings indicate that women with LC appear to be more affected by central arterial stiffness; considerably higher cfPWV values were measured in women with LC compared to the control group. This suggests that, within this demographic, cfPWV serves as a more precise indicator of alterations in arterial distensibility among women [49].

A detailed analysis of the 14 components of the MD adherence questionnaire allows us to identify those with the greatest impact on vascular health, namely, the protective role of using olive oil as the main cooking fat, consuming pulses ≥ 3 times per week, and limiting the consumption of industrial pastries. Olive oil and pulses are foods rich in polyphenols, fibre, and monounsaturated fatty acids, with proven anti-inflammatory and antioxidant effects that can improve endothelial function, reduce AS, and prevent the progression of atherosclerosis [10,12]. MD is characterised by the utilisation of olive oil, especially extra virgin olive oil, as a fundamental ingredient. It is renowned for its high concentration of monounsaturated fats and bioactive compounds, such as polyphenols like oleocanthal and hydroxytyrosol. These are potent antioxidants that have been demonstrated to produce anti-inflammatory, cardioprotective, and neuroprotective actions. Research has demonstrated that olive oil's antioxidant properties can decrease markers of oxidative stress. This enhances the bioavailability of fat-soluble vitamins and polyphenols from other elements of the MD, thereby boosting its health benefits and establishing it as a fundamental component in precision nutrition strategies designed to reduce oxidative stress and inflammation [26]. On the other hand, pulses provide soluble fibre and bioactive compounds that enhance the lipid profile and decrease blood glucose, both of which are relevant in control of metabolic syndrome, which is common in LC [50]. Given that LC is characterised by sustained inflammation and oxidative stress, these potential effects could be particularly beneficial in preventing vascular dysfunction.

Most existing studies on diet and LC have focused on the effect of diet on symptoms or disease severity in the acute phase [51,52]. Alternatively, they have examined the potential of diet and nutrition to support recovery from LC symptoms by targeting inflammation and improving outcomes [2]. Therefore, the paucity of literature analysing the relationship between diet and vascular health in this population renders this study a novel contribution to the field. Among these studies, there is a report conducted in a sub-Saharan adult population that showed that central arterial stiffness measured with cfPWV was higher in adults with LC than in controls who had never been infected with SARS-CoV-2 [49]. AS is intimately linked to the aging process, vascular injury, and the function of endothelial cells. The endothelium helps to regulate blood flow and pressure by releasing substances that affect blood vessels, such as nitric oxide, prostacyclin, endothelin-1, and platelet-activating factor [53]. Furthermore, the findings of this study are consistent with those of other studies conducted on the general population or in patients with chronic diseases, where adherence to MD has shown benefits in AS and other vascular indicators [16]. However, it should be noted that the inflammatory profile of LC could modify the magnitude or direction of these associations. Consequently, it is paramount to investigate this relationship in specific populations, such as the one studied, to ensure a comprehensive understanding of its implications.

The strengths of the study include the sample size, which is larger than most previous studies, the use of validated instruments for measuring vascular structure and function, and the adjustment for multiple confounding factors in the analyses (age, sex, lifestyle, use of medication, and cardiovascular risk factors that define MetS). In addition, a validated questionnaire was used to measure adherence to the MD, which had previously been used in the PREDIMED study.

However, several limitations should also be considered. Firstly, the cross-sectional design does not allow the establishment of causal relationships between the MD and vascular parameters. Secondly, although multiple variables were adjusted, the existence of unmeasured residual confounding factors, such as socioeconomic status or sleep quality, is possible. Thirdly, the diet questionnaire is based on self-reporting, which is susceptible to memory bias. Finally, the population included comes only from Salamanca, which may prevent the generalisation of the results to other areas or healthcare contexts. Therefore, future longitudinal studies or randomised interventions could provide more robust evidence on the impact of the MD on the progression of vascular damage in patients with LC.

The results of this study indicate that encouraging dietary habits like the MD could be a useful non-pharmacological approach to reduce the vascular impact of LC, especially in women. This approach is particularly attractive in primary care, where lifestyle intervention plays a pivotal role. Future research should explore the longitudinal evolution of vascular parameters in patients with LC and whether improved adherence to the MD translates into sustained clinical benefits. It would also be relevant to analyse the role of other dietary components such as probiotics or fermented foods, which could modulate post-COVID-19 intestinal dysbiosis.