Introduction

Kidney transplantation (KT) stands as a life-altering intervention for patients with end-stage renal disease. It offers significant improvement in quality of life and survival rate when compared to long-term dialysis1,2. Despite these benefits, the post-transplantation period is laden with challenges, including the risk of allograft dysfunction, the complex management of immunosuppressive therapy, and the increased risk of cardiovascular disease (CVD), which is the leading cause of death among transplant recipients2,3. To our knowledge, this is the first assessment of the interaction between fatty acid (FA) profiles and early alterations in the level of microcirculation performed in KT recipients, which might provide significant insights into mechanisms underlying CVD. Considering that microvascular changes might significantly precede the onset of CVD4, the results of the present investigation have the potential to be translated from bench to bedside, modifying CV risk and improving KT recipients’ outcomes.

Cardiovascular health is directly related to the endothelium, which lines the inner surface of blood vessels and is responsible for multiple vasoprotective mechanisms5. The endothelial cells modulate vascular tone by releasing vasoactive substances, regulate inflammatory response, platelet function, and vascular smooth muscle cell growth5. Therefore, the endothelial injury and loss of its beneficial actions are the first steps in the pathogenesis of CVD6,7. Previous studies have revealed FAs to have a multifaceted impact on the endothelial function8,9,10. Some of them, like saturated FAs induce endothelial damage, by reducing endothelial-derived nitric oxide, while others, like n-3 poly-unsaturated FAs (PUFAs) improve vascular health by anti-inflammatory properties and activation of eNOS phosphorylation8,9,10. Substantial majority of these studies are based on the animal models or cultured human endothelial cells11,12,13,14, whereas the novelty of our investigation benefits from in- vivo, non- invasive and ultra- precise methods like adaptive optics retinal Rtx1 camera and flow mediated skin fluorescence (FMSF) performed in KT recipients.

Post-transplant studies of FAs profiling are crucial for understanding vascular health in transplant recipients, as chronic kidney disease (CKD) and transplantation alter FA profiles and influence cardiovascular risk15,16,17,18. The recent investigations focused on FA profiles in KT recipients are very scarce and reported a reduction in n-3 PUFAs and an increase in ultra-long chain FAs among this group of patients, speculating only on the potential impact of these alterations on the circulatory system16,17,18. However, there is a gap in research showing not only imbalances in FAs, but also their connection to the cardiovascular system in this distinctive population of KT recipients. Since microcirculatory structural and functional derangements might contribute to reduced capillary blood flow, acute and chronic rejection, and finally to allograft dysfunction and failure6, there is an increasing need for transplant physicians to identify mechanisms contributing to microvascular changes. Thus, our study, covering this gap by pointing out FAs linked to microvascular condition, might be of great clinical importance and has potential translational implications for nutritional management.

We aimed to evaluate the association between serum FA profiles and microvascular structure and function examined in retinal and cutaneous regions among stable KT recipients. To evaluate the function and structure of microcirculation we used innovative methods in easy-accessible regions: FMSF in cutaneous area and adaptive optics camera in retina, which were demonstrated to correlate with other microcirculation districts of the human body and might be considered to be an indicator of general microvascular status19,20.

Materials and methods

Study design

This is an observational and single-center study. The present investigation adheres to the principles established in the Declaration of Helsinki. The study protocol of the study received approval from the Local Bioethics Committee at the Medical University of Gdansk (protocol no. NKEBN/614/2013–2014). Before participation, all subjects provided written informed consent.

Study population

The study included 52 KT patients admitted to the Nephrology and Transplantology outpatient clinic in University Clinical Centre in Gdansk (Poland). The inclusion criteria were aged above 18 years and minimum time since KT of 3 months. The exclusion criteria were inability to give informed consent, cognitive impairment, active oncological diseases, severe heart failure with left ventricular ejection fraction < 35%, cardiomyopathies, severe valvular disease, and pregnancy.

The median time since KT was 115.5 (4-348) months. A pre-emptive transplant was performed in 17.3% of patients. Among all subjects, 23.1% had been diagnosed with diabetes mellitus (including 7.7% with diabetes type 1), however, the primary cause of renal failure in the study group was glomerulonephritis, affecting 28.8% of participants. The research encompassed KT recipients in stable clinical conditions. Post-transplant, the participants encountered no surgical or infectious setbacks linked to KT, and instances of allograft rejection were absent. Comprehensive immunosuppressive management, including tacrolimus/cyclosporin, mycophenolate mofetil, and glucocorticosteroids, was administered to KT patients.

Biochemistry

Blood samples were obtained following an overnight fasting period. Standard assessments were conducted as part of routine procedures, including complete blood count, glucose, creatinine, blood urea nitrogen (BUN), urea, calcium, phosphorus, sodium, potassium, and iron. After centrifugation at 3,000 × g, the serum sample was coded and stored at − 80 °C until analyzed for FA, high-sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), and lipid profile. Levels of hs-CRP and IL-6 were determined using the enzyme-linked immunosorbent assay (ELISA) method. Serum lipid profile and albumin were estimated using a laboratory analyzer (Erba Diagnostics Mannheim Gmbh, XL-100, Mannheim, Germany).

Fatty acids analysis

Total lipids were extracted from serum samples with a chloroform-methanol mixture (2:1, v/v) according to Folch et al.21 and after dried under a nitrogen stream were subjected to 3 h of hydrolysis with 0.5 M KOH at 90 °C. After incubation, the mixtures were acidified with 6 M HCl. 1 mL of water was added, unesterified FA were extracted thrice with 1 mL of n-hexane and the organic phase was evaporated under a nitrogen stream. The extracts were then derivatized into fatty acid methyl esters (FAME) with 10% boron trifluoride in methanol solution at 55 °C for 1.5 h. Then, 1 mL of water was added and the FAME were extracted with 3 × 1 mL n-hexane, dried under a nitrogen stream, and stored at − 20 °C until analysis. The FAME were analyzed with a GC-EI-MS QP-2020 NX (Shimadzu, Kyoto, Japan) with chromatographic separation on a Zebron ZB-5MSi capillary column, 30 m × 0.25 mm i.d. × 0.25 μm film thickness, (Phenomenex, Torrance, CA, USA). The samples were injected into dichloromethane. 1 µl of the sample was injected at a split mode. The column temperature was set in a range of 60 °C to 300 °C (4 °C/min), with helium as the carrier gas at the column head pressure of 60 kPa. The temperature of injection, ion source, and transfer line were 300 °C, 200 °C, and 300 °C, respectively. The electron energy used for FAME ionization was 70 eV; 19-methylarachidic acid was used as an internal standard. Full scan mode was used with a mass scan range of m/z 45 to 700. Accurate identification of the FA profile was possible based on FAME mixture standards (Larodan, Michigan, USA, and Merck, Darmstadt, Germany)22.

Blood pressure measurement

Blood pressure was acquired through a validated oscillometric device applied to the non-dominant arm or, in the instance of individuals with a previously utilized arteriovenous fistula for hemodialysis, on the dominant arm. The measurement was conducted three times with the participant in a supine position, and the average blood pressure was considered for subsequent analysis.

Microvascular assessment

Functional parameters

Microvascular function was assessed using a non-invasive flow-mediated skin fluorescence (FMSF) technique23. Quantification of FMSF parameters was performed using AngioExpert created by Angionica Ltd. (Lodz, Poland). FMSF involves monitoring the fluorescence of the reduced form of the nicotinamide adenine dinucleotide (NADH) intensity in epidermidis, in response to the occlusion of the brachial artery (achieved through a 3-minute occlusion using a blood pressure cuff on the arm). This technique enables the evaluation of microvascular function during the post-occlusion period and the observation of biochemical processes during controlled ischemia24.

The FMSF method facilitates the analysis of various parameters of the NADH fluorescence curve, with particular emphasis on the ischemic response (IRmax) and hyperemic response (HRmax). The ischemic response (IRmax) is defined as the ratio (in %) of relative to baseline maximal increase in NADH fluorescence intensity observed during cuff occlusion and the hyperemic response (HRmax) expressed (in %) as relative to baseline maximal decrease in NADH fluorescence intensity after cuff release. Additionally, IRindex and HRindex are defined as the area under the curve (AUC) of ischemic and hyperemic response respectively in relation to the baseline. The reactive hyperemia response (RHR) parameter reflects the combined response from both the ischemic and hyperemic FMSF trace. Flowmotion (FM) reflects blood flow oscillations. Frequency analysis reveals that blood flow oscillations fit into several periodic activities, classified as: endothelial NO-independent endothelial NO-dependent; neurogenic myogenic respiratory and cardiac25,26. Normoxia oscillatory index (NOI) represents the proportion of endothelial and neurogenic oscillations in relation to all oscillations detected in the low-frequency range (< 0.15 Hz)23,27. NOI was calculated using the formula:23

$$\begin{aligned}\text{NOI} & = \text{[Power Spectral Densities (PSD)(endothelial)} \\ & + \text{PSD (neurogenic)]/[PSD(endothelial) + PSD (neurogenic) + PSD(myogenic)] } \times 100\%.\end{aligned}$$

The direct measurements of oscillations during the reperfusion stage enable the assessment of the hypoxia sensitivity (HS), which reflects the intensity of myogenic oscillations on the reperfusion line23,28.

Structural parameters

The evaluation of structural parameters of retinal microcirculation was performed using a non-invasive tool- Adaptive Optics Retinal Camera Rtx 1 (Rtx1™; Imagine Eyes, Orsay, France). The camera enables the precise assessment of microvasculature due to the innovative adaptive optics technology based on highly developed optical systems. It consists of 3 main components: a high-resolution fundus camera, a Shack-Hartmann wavefront sensor, and a deformable mirror for real-time correction of the aberrations of the ocular wavefront29. The camera provides software for precise assessment of the arteriole’s structure: lumen diameter (LD), wall thickness (WT), wall-to-lumen ratio (WLR), and wall cross-sectional area (WCSA). The structural arteriolar diameters of the retinal microcirculation were determined using the following equations30:

$${\text{WLR}}\,{\text{=}}\,{\text{(OD – LD)/LD}}.$$
$${\text{WT}}\,{\text{=}}\,{\text{OD-LD/2}}.$$
$${\text{WCSA}}=\,\pi /4 \times\left[(OD)^{2}-(LD)^{2}\right].$$

The examinations were conducted in a dark room in a sitting position following a 15-minute rest. All measurements were taken in the supertemporal artery of the right eye, free of the presence of focal arterial nicking or arteriovenous crossings. Pupil dilatation was not required30.

Echocardiographic analysis

An echocardiographic examination was performed to exclude severe heart failure, cardiomyopathies, or severe valvular diseases. Vivid E95 (GE Healthcare, Horten, Norway) was used to examine patients. Two and three-dimensional images were obtained in the left decubitus position along the parasternal long and short axes and from apical long-axis views in accordance with principles described in Recommendations for Chamber Quantification by Echocardiography31. All recordings included three cardiac cycles and were analyzed offline using echocardiographic quantification software (EchoPac 201, GE Healthcare, Norway).

Statistical analysis

Statistical analyses were performed using R version 4.1.2 (cran.r-project.org). A p-value < 0.05 denoted statistical significance. Descriptive statistics were reported as means ± standard deviations (SD) for normally distributed variables or medians (interquartile ranges, IQRs) for non-normal variables, determined by the Shapiro-Wilk test. Correlations between serum fatty acid (FA) levels and microvascular parameters were examined using Pearson’s correlation for normally distributed variables or Spearman’s rank correlation for non-normal variables, with R-values indicating effect sizes. Given the exploratory design and small sample size (n = 52), p-values were unadjusted for multiple comparisons to prioritize sensitivity and minimize type II errors, with potential type I error risks addressed in the Discussion. This choice reflects the hypothesis-generating nature of the study, where false positives are preferable to missing novel signals in an understudied area like fatty acid-microcirculation links in KT recipients32,33; however, it may inflate the family-wise error rate, necessitating cautious interpretation and replication in larger cohorts. Additionally, post-hoc power analyses, while provided for contextualization, do not inform prospective study design and should be interpreted cautiously.

Multivariable regression modelled FA associations with microvascular parameters, adjusted for age, sex, diabetes, eGFR (CKD-EPI), hs-CRP, and statin use. Linear regression analyzed continuous outcomes (wall-to-lumen ratio [WLR], wall cross-sectional area [WCSA], hyperaemic response index [HRindex]), reporting coefficients and R² for significant FAs (p < 0.05) in Table 4, with full model results in Supplemental Table S2. Logistic regression modelled microvascular status (above/below median for WLR, WCSA, lumen diameter [LD], normoxia oscillatory index [NOI]), using median cutoffs owing to the absence of established reference ranges in KT recipients and the clinical relevance of eGFR and hs-CRP as confounders. For significant FAs (p < 0.05, AUC > 0.70), odds ratios (ORs), 95% confidence intervals (CIs), AUC, sensitivity, and specificity are reported in Table 5. Extremely wide CIs (upper/lower bound ratio > 1,000) are included to capture potential associations for hypothesis generation despite imprecision from small sample size or sparse data and are flagged in footnotes to urge cautious interpretation. Full logistic model results are provided in Supplemental Table S3. Logistic model performance was assessed via leave-one-out cross-validation (LOOCV), suitable for small samples, with AUC, sensitivity, and specificity averaged across iterations. Only models with AUC > 70% and p < 0.05 were reported, ensuring clinically relevant predictive performance. The Youden index determined optimal decision thresholds, and complete-case analysis was used for each model.

Results

Patients characteristics

The basic characteristic of the study group (n = 52) is presented in Table 1. The average age of KT patients was 50.4 ± 11.6 years. Women constituted 44.2% of the study population. Estimated glomerular filtration rate (eGFR) assessed by the Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI) less than 60 ml/min/1.73 m2 was observed in 61.5% of the study group. Total cholesterol was elevated (> 190 mg/dL) in 44.2% of KT recipients, and 32.7% of patients were prescribed statins. High-sensitivity C-reactive protein (hs-CRP) was elevated in 17.3% of study participants.

Table 1 Characteristics of the study group.
Full size table

Microcirculation analysis

The results of the microcirculation examination are presented in Table 2.

Table 2 Characteristics of microvascular structure and function.
Full size table

Figure 1 presents the comparison of microvascular structural characteristics between KT recipient (a) and a healthy volunteer (b). LD is evidently narrower, while WT, WCSA, and WLR are much greater in KT patient than in a healthy volunteer.

Fig. 1
figure 1

Comparison of microvascular structural characteristics between KT recipient (a) and healthy volunteer (b). LD - lumen diameter; WT - wall thickness; WCSA - wall cross-sectional area; WLR – wall-to-lumen ratio.

Full size image
Fig. 2
figure 2

Comparison of microvascular functional characteristics evaluated after post-occlusion reactive hyperemia test between a kidney transplant recipient (a) and a healthy volunteer (b). IR - ischemic response; HR - hyperemic response; FM - flowmotion; HS - hypoxia sensitivity; RHR - reactive hyperaemia response.

Full size image

Figure 2 depicts microvascular functional characteristics in KT recipient and healthy volunteer. Compared to the healthy volunteer, KT recipient presented a severely impaired ischemic response to brachial artery occlusion; the hyperemic response was weaker as well.

Serum fatty acid analysis

The full FA profile in KT patient sera is presented in Supplemental Table 1. The significant results of the analysis of the correlation between serum FA levels and microvascular parameters are depicted in Table 3. Adrenic acid (22:4n-6) and eicosatetraenoic acid (20:4n-3) were significantly related to WLR. Some branched-chain fatty acids (BCFAs) were positively associated with LD and HR index, pentadecanoic acid (15:0) correlated positively with HRindex, whereas linoleic acid (18:2n-6) was positively correlated with WCSA. To contextualize the robustness of these associations given the small sample, post-hoc power calculations were performed for the observed effect sizes (assuming two-sided alpha = 0.05), with caveats on their limitations noted above. Observed powers ranged from 0.518 to 0.852, indicating moderate-to-high detectability for these effects but underscoring the need for larger studies. These findings are hypothesis-generating and require independent validation.

Table 3 Correlation analysis between fatty acids and microvascular parameters.
Full size table

Post-hoc power calculations (two-sided α = 0.05) are provided for contextualization, assuming observed effect sizes reflect population values; these are interpretive aids only and underscore the need for replication (see Statistical analysis for details).

Table 4 presents a multivariable analysis between FA and structural and functional parameters of microcirculation, adjusted for confounding factors such as age, sex, diabetes, eGFR, hs-CRP, and statin therapy. The analysis confirmed that WLR was independently associated with PUFA: 22:4n-6 and 20:4n-3. A significant and strong relationship was found between WCSA and 18:2n-6. Among functional microvascular measures, only HRindex was independently linked to iso C15. R² represents the proportion of variance explained by the full model. Confounder results are provided in Supplemental Table S2.

Table 4 Multivariable analysis between fatty acids and microvascular parameters.
Full size table

Logistic regression analysis was performed to obtain determinants of microvascular status (Table 5). Because of the lack of established reference ranges for microvascular parameters, using median cut-offs was necessary in this logistic regression analysis. Models predicted microvascular status (parameters below median) and were adjusted for age, sex, diabetes, eGFR CKD-EPI, hsCRP and statin therapy. Only fatty acids with p < 0.05 and AUC > 70% are reported. Extremely wide CIs (upper/lower bound ratio > 1,000, denoted by **) reflected imprecision due to small sample size (n = 52) or sparse data and should be interpreted cautiously as hypothesis-generating signals rather than precise effect estimates. Large ORs may be overestimated; such estimates primarily indicate direction rather than precise magnitude and theirs interpretation requires caution. Microcirculatory variables expressed as values below the median reflect positive or negative change in microvascular status (WLR and WCSA below median - positive, whereas LD and NOI below median - negative). 22:4n-6 and 20:4n-3 were revealed as significant variables determining a decrease in WLR, which is associated with a beneficial change. 18:2n-6 was related to increased WCSA, which shows poor condition of the vascular wall, whereas isoC17 was positively related to lumen diameter. Docosapentaenoic acid (22:5 n-6) correlated positively with NOI indicating reflecting blood flow oscillations. Confounder results are provided in Supplemental Table S3.

Table 5 Logistic regression analysis evaluation the relationships between microcirculation and fatty acids.
Full size table

Obesity was not included in the primary models due to the limited sample size (n = 52) and the risk of overfitting with an additional confounder, given the already large number of variables adjusted for. However, to assess its potential influence, supplementary linear regression models were prepared incorporating obesity as a binary variable (BMI ≥ 30 kg/m² vs. <30 kg/m²) alongside the original confounders (Supplemental Table S4). Logistic regression models including obesity are presented in Supplemental Table S6.

Figure 3 summarizes the clinical relevance (positive or negative) of the associations between FAs and microcirculation.

Fig. 3
figure 3

The associations between FAs and microcirculation in KT recipients.

Full size image

Discussion

To the best of our knowledge, this is the first study on the association between FA profile and microcirculation in patients after KT. CKD is associated with increased cardiovascular risk, even after KT. Cardiovascular disorders in KT patients comprise impaired structure and function of microcirculation, which is the result of numerous factors, including metabolic, inflammatory, and pharmacological ones34. Lipid alterations are also associated with different kidney disorders. Our recent studies showed many alterations in FA profiles in patients after KT including considerable PUFA deficiencies16,17,18,35.

In the present investigation, we employed various statistical methods to search for associations between wide FA profile and microcirculation parameters obtained from advanced examinations using innovative, precise, and non-invasive tools such as Rtx1 or FMSF. The major finding of this study is the presence of significant associations between specific FA and microvascular structural and functional parameters in KT recipients. The highest number of such relationships was found among the PUFA group. Two representatives of PUFA, 22:4n-6 and 20:4n-3 correlated negatively with WLR indicating positive remodelling reflected by a decrease of WLR (below median). It seems a particularly valuable observation, since 22:4n-6 and 20:4n-3 levels decreased after KT, yet they still played an important role in the microvascular condition16. Generally, PUFA n-6 are considered as acids with controversial effects on the cardiovascular system, especially in the context of inflammation, since arachidonic acid (20:4n-6) is a precursor of proinflammatory eicosanoids36. A high n-6/n-3 PUFA ratio in blood is associated with elevated inflammation, a state that increases the risk of CVD37. However, 22:4n-6 may protect from the production of proinflammatory leukotriene B4 and in contrast to 20:4n-6 it has rather anti-inflammatory effects38. Linoleic acid (18:2n-6) - the most abundant PUFA n-6 in the blood of the study patients (Supplemental Table 1)- is a precursor of proinflammatory oxylipins36. Our results showed that it was positively correlated with WCSA indicating its association with the arteriolar wall thickening and negative structural remodelling. There is substantial evidence that linoleic acid is an atherogenic FA because of its pro-oxidative and proinflammatory properties.

Interestingly, another PUFA n-6 (22:5n-6), formed from 18:2n-6, was associated with NOI, which reflects microvascular function. This highly unsaturated FA has also been revealed as inversely associated with coronary events in the meta-analysis of Chowdhury et al.39 and correlated with lower mortality in men40.

We have also found associations between microcirculation parameters and some BCFA. Based on previous studies, iso C17 potentially downregulates inflammation and decreases liver lipids production41,42. Considering the microcirculation parameters, it correlated positively with LD suggesting its association with the beneficial structural microvessels’ changes. The same association has been found for iso C16. In turn, iso C15 correlated positively with HR-index which reflects microvascular function. BCFA are responsible for beneficial bioactivities, such as anti-inflammatory and lipid-lowering effects, maintenance of insulin sensitivity, and mitigation of cerebral ischemia and reperfusion injury43. It should be taken into consideration that this BCFA was present in a very small amount in patients’ blood (Supplementary Table 1). All the above-mentioned iso BCFA were related to microvascular status, albeit their levels were significantly lower in KT patients than in healthy subjects17.

Supplementary analyses adjusted for obesity showed that the key associations remained largely consistent, with minimal changes in their direction and significance, thereby supporting the primary findings despite the addition of this potential confounder.

Several limitations should be noted. The study includes a small number of patients (n = 52) however, it is a unique investigation addressing clinically relevant issues affecting KT recipients and the graft function: microcirculation derangements and FA concentrations. These topics were not sufficiently explored in the previous studies, and to our knowledge this is the first attempt to evaluate the relationships between them in this patient population. Therefore, we consider this pioneering, albeit small, study to support the need to conduct further research focused on microcirculation and fatty acids. The modest sample size may lead to wide confidence intervals in some logistic regression analyses, reflecting potential imprecision due to limited data. These intervals were retained to identify associations for hypothesis generation, though their interpretation warrants caution. The use of unadjusted p-values to enhance sensitivity in this exploratory study may increase the risk of type I errors with multiple comparisons. Additionally, we dichotomized microvascular parameters at their medians for logistic regression (Table 5) due to the lack of established clinical reference ranges or validated cutoffs in KT recipients. This enabled modeling binary outcomes (below vs. above median) as proxies for unfavorable states, providing clinically interpretable odds ratios, with the median ensuring balanced groups in our small sample. The single-center design may further limit the generalizability of findings to broader populations, necessitating multicentre studies for validation.

In conclusion, our results showed that some circulating PUFA and iso BCFA are significantly associated with microvascular parameters and that most of them are related to favourable structural and functional characteristics of microcirculation. Enriching the diet of KT recipients in the abovementioned FA may ameliorate microvascular status. Microvascular structural and functional changes in KT patients may serve to assess the influence of FA on cardiovascular health. Since there is the possibility that these FA could improve microcirculation parameters, and their levels are decreased in KT recipients, further in vivo studies are warranted.