Sex-specific associations of perioperative Resolvin D1 and D2 dynamics with postoperative atrial fibrillation in cardiac surgery: a secondary analysis of a prospective observational study Authors: Melina Heine, Felix Ulbrich, Samuel Justice, Babak Saravi, Simon A Amacher, Julian Hubrich, Leonard Simeth, Asishana Osho, Charles N Serhan, Jochen D Muehlschlegel, Jakob Wollborn Published: 2026 ABSTRACT AbstractAimsPostoperative atrial fibrillation (POAF) is a prevalent and clinically significant complication that occurs after cardiac surgery. While inflammation has been linked to its onset, the role of active inflammatory resolution—mediated by a group of mediators named specialized pro-resolving mediators (SPMs)—remains unclear. Since SPMs, such as resolvins, actively terminate inflammation and restore tissue homeostasis, impaired resolution could contribute to persistent postoperative inflammation and the development of POAF.MethodsThis study is a secondary analysis of a single-centre prospective cohort of 319 adult patients who underwent cardiac surgery. The occurrence of POAF was identified. Resolvin D1 (RvD1) and D2 (RvD2) levels were measured during the perioperative period. The association between resolvin levels and POAF was assessed using a mixed-effects Tobit regression model. The interaction between gender, POAF, and resolvin levels was furthermore investigated.ResultsPatients who developed POAF exhibited significantly higher predicted perioperative levels of RvD1 [mean ratio (MR—interpreted analogously to an odds ratio): 1.19, P < 0.001] and significantly lower predicted levels of RvD2 (MR: 0.80, P < 0.001) compared to those without POAF. The association between POAF and RvD1 was amplified in women compared to men (P = 0.012). A similar pattern was noted for RvD2, with women tending to show a more pronounced decrease in RvD2 levels associated with POAF. However, this finding did not reach statistical significance (P = 0.054).ConclusionPOAF is associated with elevated RvD1 and reduced RvD2 levels, suggesting a differential response and potentially distinct roles of the two pro-resolving mediators in patients with POAF. Additionally, these associations differ by sex, indicating that inflammatory resolution mechanisms may vary between males and females.RegistrationGerman Clinical Trials Registry (DRKS No. 00017057), Date of registration 5 April 2019, www.drks.de.Keywords: Postoperative atrial fibrillation, Cardiac surgery, Inflammation, Inflammatory Resolution, Resolvins What is Known Postoperative atrial fibrillation (POAF) is a prevalent complication following cardiac surgery, and its pathogenesis has been linked to inflammation. Postoperative atrial fibrillation (POAF) is a prevalent complication following cardiac surgery, and its pathogenesis has been linked to inflammation. Specialized pro-resolving mediators (SPMs), such as resolvins, actively terminate inflammation, but their role in POAF development remains to be elucidated. Specialized pro-resolving mediators (SPMs), such as resolvins, actively terminate inflammation, but their role in POAF development remains to be elucidated. Patients who developed POAF showed higher perioperative levels of Resolvin D1 and lower levels of Resolvin D2, suggesting that these mediators may play divergent roles in inflammatory resolution following surgery. Patients who developed POAF showed higher perioperative levels of Resolvin D1 and lower levels of Resolvin D2, suggesting that these mediators may play divergent roles in inflammatory resolution following surgery. The association between POAF and Resolvin D1 was more pronounced in women, suggesting that sex-specific inflammatory resolution mechanisms may exist. The association between POAF and Resolvin D1 was more pronounced in women, suggesting that sex-specific inflammatory resolution mechanisms may exist. Distinct and sex-dependent SPM dynamics may contribute to the development of POAF, thereby offering potential targets for personalized anti-inflammatory strategies. Distinct and sex-dependent SPM dynamics may contribute to the development of POAF, thereby offering potential targets for personalized anti-inflammatory strategies. Introduction Postoperative atrial fibrillation (POAF) is the most frequent complication, occurring in 30–50% of patients undergoing cardiac surgery.1,2 POAF is associated with prolonged intensive care unit (ICU) and hospital stays, thereby placing a significant burden on the healthcare system.2–4 Baseline patient risk factors for POAF include age over 75 years, chronic kidney disease (CKD), congestive heart failure (CHF), mitral valve disease, and a history of atrial fibrillation (AF).1 When paired with perioperative risk factors, such as surgery involving the mitral valve and the use of cardiopulmonary bypass (CPB), patients are at an even higher risk of developing POAF and its subsequent consequences.2 Patients diagnosed with POAF have an elevated risk for adverse events, including major bleeding, acute kidney injury, prolonged mechanical ventilation, reoperation, strokes, and an increased mortality rate.3–5 Furthermore, the risk for developing recurrent AF over time remains increased, presenting a significant risk beyond the immediate postoperative period.5–7 While women are less likely to develop POAF, those who do develop POAF have an elevated long-term mortality rate compared to men.8 The pathogenesis of POAF is multifactorial, with local and systemic inflammation playing key roles.9,10 Cardiac surgery frequently elicits a robust inflammatory response, driven by a ‘quadruple hit’ of surgical trauma, haemorrhage or haemodilution, CPB, and ischaemia-reperfusion injury.11,12 A number of inflammatory mediators have been associated with the development of POAF, including C-reactive protein (CRP) and various interleukins.13,14 Inflammation can trigger structural remodelling and electrophysiological disturbances in the atria (Figure 1).10 Emerging evidence underscores the importance of the resolution phase of inflammation in the pathophysiology of AF.15 Inflammatory resolution is defined as an active process that aims to terminate the inflammatory response and restore tissue function. This process is mediated by specialized pro-resolving mediators (SPMs), including resolvins (Rvs), a superfamily of bioactive lipid molecules derived from polyunsaturated fatty acids (PUFAs).16,17 However, the role of Rvs in the perioperative cardiac surgery setting and their potential contribution to POAF development remain to be fully elucidated. Pathogenesis of postoperative atrial fibrillation (POAF). Cardiac surgery triggers local and systemic inflammation, leading to conduction heterogeneity in the atria and may result in transient atrial arrhythmias, such as POAF. If inflammation remains unresolved, fibrotic remodelling and sustained conduction abnormalities can develop on a tissue level, which, in turn, increases the risk of long-term atrial arrhythmias. The process of inflammatory resolution, actively mediated by specialized pro-resolving mediators such as resolvins, may protect against arrhythmia progression. The objective of this study is to investigate the association between POAF and perioperative levels of the main SPMs Resolvin D1 (RvD1) and Resolvin D2 (RvD2). We hypothesize that inflammatory resolution is dysregulated in patients with POAF, characterized by lower levels of SPMs. Methods Design and setting A retrospective analysis was conducted in a cohort of 319 patients who underwent open-heart surgery with CPB. Patients were enrolled at the University Medical Center Freiburg between May 2019 and October 2020 to participate in a prospective observational clinical study to characterize the capillary leak syndrome after cardiac surgery.18 The University Medical Center Freiburg is a highly specialized centre for advanced adult and paediatric cardiac surgery in southwestern Germany. The initial study screened 918 patients for study inclusion and 410 of these patients were enrolled (Figure 2). The patients or their designated healthcare proxies provided consent for the study. The study was approved by the institutional review board (Freiburg, Germany; Chairperson: Prof. Dr. R. Korinthenberg; EK-Nr. 405/18) on 10 January 2019 and subsequently registered in the German Clinical Trials Register (DRKS No. 00017057). The study was conducted and reported according to the STROBE guidelines. Study flow chart. *Permanent atrial fibrillation: ‘AF for which no further attempts at restoration of sinus rhythm are planned, after a shared decision between the patient and physician.’ #Persistent atrial fibrillation: ‘AF episodes which are not self-terminating. Many intervention trials have used 7 days as a cut-off for defining persistent AF. Long-standing persistent AF is arbitrarily defined as continuous AF of at least 12 months’ duration but where rhythm control is still a treatment option in selected patients, distinguishing it from permanent AF.’ Classification of AF according to the 2024 ESC/EACTS guidelines for the management of atrial fibrillation.19. AF, atrial fibrillation; ICU, intensive care unit. Study population The study initially aimed to investigate the occurrence of capillary leak syndrome in adult patients after cardiac surgery.18 During the study period, all patients scheduled for cardiac surgery involving CPB were evaluated for eligibility. Following the exclusion of patients with longstanding AF, patients undergoing orthotopic heart transplantation, and those with incomplete resolvin measurements, 319 patients were included in the present analysis. The patients underwent a variety of cardiac surgical procedures. Surgical procedures were grouped into seven categories. In cases of combined procedures, group assignment was primarily determined by the most invasive procedure. For instance, a patient undergoing combined aortic surgery and coronary artery bypass grafting (CABG) was assigned to the aortic surgery group. Clinical care Study enrolment occurred prior to cardiac surgery after informed consent. Further measurements and blood samples were obtained; however, clinical care provided to the patients remained unchanged throughout the study. The patients received routine care during the surgical procedure as well as during the subsequent postoperative hospitalization. Study measurements A comprehensive review of the medical records was conducted for POAF diagnosis. The reviewed data included anaesthesia protocols, daily notes, discharge summaries, 12-channel ECG recordings, telemetry, and Holter monitoring. The presence of at least one episode of POAF, with a duration exceeding 30 s, was considered sufficient for diagnosis, according to current guidelines.19 RvD1 and RvD2 serum levels were measured in blood samples collected preoperatively (after induction of general anaesthesia), postoperatively at ICU admission, and daily in the morning in the ICU with the routine laboratory workup. A follow-up blood sample was collected from patients who were still hospitalized (in the ICU or step-down unit) 7 days after surgery. Blood samples were collected in serum tubes and centrifuged according to a standardized protocol. The serum was aliquoted immediately and stored at −80°C until analysis. All samples were handled uniformly. Enzyme-linked immunosorbent assay (ELISA) was utilized to quantify RvD1 (Resolvin D1 ELISA Kit, Cayman Chemical) and RvD2 (Resolvin D2 ELISA Kit, Cayman Chemical) levels. A total of 1358 samples for both RvD1 and RvD2 were analysed over the perioperative course according to the manufacturer’s protocol. Statistical analysis Analysis was performed using the R statistical software (R Project for Statistical Computing, Version 4.4.1). Descriptive statistics were applied for the analysis of baseline, intra-, and postoperative characteristics. Continuous variables were summarized using either the mean and standard deviation or the median and interquartile range, depending on their distribution. Comparisons between patients with and without POAF were conducted using the t-test for normally distributed continuous variables and the Mann–Whitney U test for non-normally distributed variables. Categorical variables were compared using the chi-square test. Due to repeated patient measurements and the presence of censored resolvin data due to measurements below the detection limit, mixed-effects Tobit regression (with a gamma link function to account for nonnegativity) was employed to assess the association of POAF with both RvD1 and RvD2 levels. The lower limits of detection for the test assays were 3.3 pg/mL for RvD1 and 1.6 pg/mL for RvD2. The regressions were adjusted for the POAF risk factors age > 75 years, history of AF, history of CKD, history of chronic heart failure, and history of mitral valve disease (echocardiographic findings of mitral valve stenosis or regurgitation) or mitral valve surgery.1 Postoperative day (POD) was additionally included as a covariate in the regression models to adjust for time trends. Although 319 patients were enrolled, sample availability varied across time points due to missing or unavailable serum samples, resulting in 282 measurements at baseline. The higher number of samples at POD7 reflects inclusion of both ICU and step-down unit patients still hospitalized at that time point, rather than ICU readmission. As sampling availability varied across time points, temporal trends should be interpreted with caution. Due to the limited number of measurements after POD 7, all subsequent measurements were combined into a single additional time category (‘POD8+’). The results of the Tobit regression models were summarized as mean ratios, which are interpreted analogously to an odds ratio. While the regression model estimates overall longitudinal differences using all available measurements across all time points, the graphical representations of RvD1 and RvD2 are intended for descriptive visualization of temporal trends. Only measurements up to POD7 are shown in the figures, although all timepoints were included in the regression analyses. Further models with interaction terms were fit to analyse whether the association of POAF with RvD1 and D2 levels varied by gender (male, female) or CRP decline group (slow vs. fast decline). For each patient, the trajectory of CRP decline between POD 3 and 10 was quantified as a slope derived from a linear regression model. POD 3 was chosen as the starting point, as CRP typically began to fall after its postoperative peak, whereas POD 10 corresponded to the plateau phase. Based on the median slope, patients were stratified into two subgroups representing a fast and slow CRP decline. Results For the original study, 918 patients undergoing cardiac surgery with CPB at the University Medical Center Freiburg were screened for inclusion. Of those, 410 patients were recruited.18 After applying the additional exclusion criteria for the current study, 319 patients were included in the secondary analysis (Figure 2). Baseline, intraoperative, and postoperative characteristics The male-to-female ratio in this cohort was 3:1 (Table 1). The mean age was 61.8 ± 14.1 years. Based on the predefined surgical groups, 27% of the patients (n = 85) underwent coronary artery bypass grafting (CABG), 22% underwent thoracic aortic aneurysm surgery (n = 71), 17% underwent aortic valve surgery (n = 54), 12% underwent mitral valve surgery (n = 38), 8% (n = 24) underwent multivalve repair or replacement, and 5% (n = 15) underwent left ventricular assist device (LVAD) implantation. Thirty-two patients (10%) could not be assigned to either group and were thus categorized as ‘other type of cardiac surgery’. These patients underwent atrial septal defect or ventricular septal defect repair, for example. The mean EuroScore I was 5 ± 4. The median CPB and aortic cross clamp times were 126 (95–179) and 92 (65–123) min, respectively. The median length of stay in the ICU and the step-down unit was 4 (2–7) and 8 (6–11) days, respectively. Table 1Patient demographics, clinical characteristics, and intra- and postoperative factorsTotal (n = 319)POAF (n = 161, 50.5%)No POAF (n = 158, 49.5%) P-value Demographics Age (mean ± SD), years61.8 ± 14.166.4 ± 12.157.0 ± 14.4<0.001 Sex, No. (%)0.873 Male238 (74.6)119 (73.9)119 (75.3) Clinical characteristics, No. (%) CHF60 (18.81)43 (26.7)17 (10.8)<0.001 History of atrial fibrillation or flutter71 (22.26)61 (37.9)10 (6.3)<0.001 Coronary artery disease175 (54.86)97 (60.2)78 (49.2)0.066 Valvular heart disease206 (64.58)116 (72.0)90 (57.0)0.007 COPD15 (4.7)4 (2.5)11 (7.0)0.104 CKD40 (12.54)28 (17.4)12 (7.6)0.013 Intraoperative factors EuroScore I (mean ± SD)5 ± 4.176.56 ± 4.285.11 ± 3.930.002 CPB (median ± IQR), min126 (95.3–178.8)139 (102–188)119 (88–167)0.008 Cross clamp time (median ± IQR), min92 (65–123)95 (67–132)90 (59–117)0.077 Ventilation time (median ± IQR), h7 (6–8)7 (6.06–8.5)6.75 (6–8)0.038 Surgical groups, No. (%)0.001 CABG85 (26.65)39 (24.2)46 (29.1) Aortic valve54 (16.93)28 (17.4)26 (16.5) Mitral valve38 (11.91)24 (14.9)14 (8.9) Multivalve24 (7.52)20 (12.42)4 (2.53) Thoracic aortic aneurysm surgery71 (22.26)29 (18.0)42 (26.6) LVAD implantation15 (4.7)11 (6.8)4 (2.5) Other type of surgery32 (10.03)10 (6.2)22 (13.9) Occurrence of intraoperative AF, No. (%)19 (5.96)14 (8.7)5 (3.16)<0.001 Already in AF when arriving in the OR4 (1.25)3 (1.9)1 (0.6) New-onset intraoperative AF15 (4.7)11 (6.8)4 (2.5) Postoperative characteristics LOS ICU (median ± IQR), days4 (2–7)5 (3–9)3 (2–5)<0.001 LOS step-down unit (median ± IQR), days8 (6–11)8 (5–12)7 (6–10)0.61 Need for cardiac reoperation, No. (%)44 (13.79)32 (19.9)12 (7.6)0.003 Duration until extubation (median ± IQR), h8 (6–16)9.5 (6–21)7 (5–10)<0.001AF, atrial fibrillation; CABG, coronary artery bypass grafting; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CPB, cardiopulmonary bypass; ICU, intensive care unit; IQR, interquartile range; LOS, length of stay; LVAD, left ventricular assist device; OR, operation room; POAF, postoperative atrial fibrillation; SD, standard deviation. Patient demographics, clinical characteristics, and intra- and postoperative factors AF, atrial fibrillation; CABG, coronary artery bypass grafting; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CPB, cardiopulmonary bypass; ICU, intensive care unit; IQR, interquartile range; LOS, length of stay; LVAD, left ventricular assist device; OR, operation room; POAF, postoperative atrial fibrillation; SD, standard deviation. In this cohort, 51% of the patients developed POAF during hospitalization, with a comparable incidence observed in male and female patients. The peak incidence of AF was observed on POD 2 and POD 3. The highest incidence of POAF was observed in the multivalve surgical group. Patients diagnosed with POAF had longer (median ± IQR) times until extubation of 10 (6–21) h, in comparison to the median time of 7 (5–10) h observed in patients without POAF (P < 0.001). Furthermore, ICU length of stay (median ± IQR) was longer in POAF patients [5 (3–9) days vs. 3 (2–4.7) days, P < 0.001]. Association between POAF and resolvin levels Patients who developed POAF had 19% higher mean predicted RvD1 levels over the perioperative period in the adjusted mixed-effects Tobit regression model [mean ratio (MR): 1.19, 95% confidence interval (CI): 1.09–1.31, P < 0.001; Figure 3A & C]. Conversely, patients with POAF exhibited 20% lower predicted RvD2 levels on average when compared to patients without POAF (MR: 0.80, 95% CI: 0.72–0.90, P < 0.001; Figure 3B & D). In a sensitivity analysis restricted to early postoperative timepoints (≤POD 2), the results remained consistent with the main analysis, with POAF being associated with higher RvD1 levels (MR: 1.15, 95% CI: 1.05–1.27, P = 0.004) and lower RvD2 levels (MR: 0.78, 95% CI: 0.70–0.86, P < 0.001). Perioperative dynamics of Resolvin D1 and D2 in patients with and without postoperative atrial fibrillation. Predicted (A) RvD1 and (B) RvD2 levels (mean ± SD) over time in patients with (red) and without POAF (blue), based on a mixed-effects Tobit regression. Results of the mixed-effects Tobit regression identifying independent associations between POAF and (C) Resolvin D1 and (D) D2 levels, adjusted for relevant clinical covariates (AF history, CKD, CHF, age, and mitral valve disease/surgery). POAF is associated with higher levels of RvD1 and lower levels of RvD2. Number of measurements: RvD1: Preop 282 (207*), Postop 273 (214*), POD1 274 (197*), POD2 117 (91*), POD3 82 (64*), POD4 57 (43*), POD5 44 (39*), POD6 35 (32*), POD7 110 (72*), POD8+ (POD8-POD21) 85 (71*). RvD2: Preop 282 (206*), Postop 273 (205*), POD1 274 (196*), POD2 117 (71*), POD3 82 (48*), POD4 57 (38*), POD5 44 (31*), POD6 35 (20*), POD7 110 (99*), POD8+ (POD8-POD21) 85 (60*). *Number of measurements within test kit specification. AF, atrial fibrillation; CHF, congestive heart failure; CI, confidence interval; CKD, chronic kidney disease; MR, mean ratio; MV, mitral valve; POAF, postoperative atrial fibrillation; POD, postoperative day; Rv, resolvin; SD, standard deviation. Associations with clinical risk factors Following adjustment for established POAF risk factors, several notable associations were revealed. A history of CHF was significantly associated with higher RvD1 levels (MR: 1.33, 95% CI: 1.20–1.48, P < 0.001). A similar association was observed for CKD (MR: 1.22, 95% CI: 1.08–1.39, P = 0.002; Figure 3C). Patients older than 75 years exhibited elevated RvD1 levels (MR: 1.17, 95% CI 1.05–1.31, P = 0.005). However, a history of AF showed a trend toward lower levels (MR 0.90, 95% CI 0.81–1.00, P = 0.05). Mitral valve disease did not have a significant impact on RvD1 serum concentrations. For RvD2, patients with a documented history of AF had significantly higher RvD2 levels (MR: 1.38, 95% CI: 1.21–1.57, P < 0.001) (Figure 3D), while those older than 75 years had significantly lower levels of RvD2 (MR: 0.86. 95% CI: 0.75–0.98. P = 0.03). However, the study found no statistically significant associations between CHF, CKD, and MV disease and RvD2 levels. Interaction with postoperative CRP trajectory Furthermore, an analysis of interaction showed that the association between POAF and perioperative SPM levels was strongly modified by the postoperative CRP trajectory. In patients with a slow CRP decline, POAF was linked to a more pronounced increase in RvD1 (P < 0.001) and a greater reduction in RvD2 (P < 0.001) compared with those with a faster CRP decline (Figure 4). Perioperative dynamics of Resolvin D1 and D2 stratified by POAF status and CRP decline group (mixed-effects Tobit regression; adjusted for AF history, CKD, CHF, age > 75 years, and MV disease/surgery). Predicted (A) RvD1 and (C) RvD2 levels (mean ± SD) in the fast CRP decline group in patients with (pink) and without POAF (rose). Predicted (B) RvD1 and (D) RvD2 levels (mean ± SD) in the slow CRP decline group in patients with (dark blue) and without POAF (light blue). AF, atrial fibrillation; CHF, congestive heart failure; CKD, chronic kidney disease; CRP, C-reactive protein; MV, mitral valve; POAF, postoperative atrial fibrillation; POD, postoperative day; Rv, resolvin; SD, standard deviation. Sex differences in the POAF–resolvin relationship Sex significantly modified the association between POAF and RvD1 levels, with the increase in RvD1 in patients with POAF being smaller in males compared with females (P = 0.012; Figure 5). For RvD2, a trend towards a similar sex interaction was observed, with females showing a more pronounced decrease in RvD2 levels associated with POAF, whereas this association was attenuated in males. However, the observed trend did not reach statistical significance (P = 0.054). Perioperative dynamics of Resolvin D1 and D2 stratified by POAF status and sex (mixed-effects Tobit regression; adjusted for AF history, CKD, CHF, age > 75 years, and MV disease/surgery). Predicted (A) RvD1 and (C) RvD2 levels (mean ± SD) in female patients with (dark red) and without POAF (light red). Predicted (B) RvD1 and (D) RvD2 levels (mean ± SD) in male patients with (dark blue) and without POAF (light blue). AF, atrial fibrillation; CHF, congestive heart failure; CKD, chronic kidney disease; MV, mitral valve; POAF, postoperative atrial fibrillation; POD, postoperative day; Rv, resolvin; SD, standard deviation. Discussion After adjusting for known clinical risk factors,1 POAF was found to be significantly associated with altered trajectories of SPMs. Patients who developed POAF had persistently elevated RvD1 levels and lower Rv2 levels, indicating an imbalanced or dysregulated inflammatory resolution response. These associations remained robust even after accounting for comorbid conditions, such as CHF and CKD, mitral valve disease and history of AF, as well as demographic factors, such as age and type of surgical procedure. A number of risk factors, including CHF and advanced age, have been identified as independently associated with variations in resolvin profiles. Our initial hypothesis that patients with POAF would exhibit lower levels of both RvD1 and RvD2 was only partially supported. Lower mediator levels were observed only for RvD2 in POAF patients. The opposite was proved for RvD1, suggesting that RvD1 and RvD2 may play distinct, possibly opposing, roles in perioperative inflammation and arrhythmia susceptibility. Despite both being derived from the same intermediate 7,8(S,S)-epoxytetraene, RvD1 and RvD2 are produced by different cell types20, act through different receptors, target different tissues,21–24 and are subject to differential inactivation.25 These mechanistic differences might explain the divergent response of the two D-series resolvins observed in POAF. Our findings are consistent with previous human studies indicating differential resolvin behaviour. For instance, elevated levels of plasma RvD1 were observed in women with breast cancer compared to women without, while no significant differences were found in RvD2 levels.26 These results underscore the necessity of future studies to elucidate the specific contributions of individual SPMs to POAF risk. Given the demonstrated importance of chronic inflammation in the development of numerous cardiovascular diseases, consideration of delayed or failed inflammatory resolution is imperative.27 In the context of cardiac surgery, transient inflammation of the atria has been observed to disrupt normal conduction pathways and facilitate re-entry mechanisms. Unresolved inflammation can lead to fibrotic remodelling of atrial tissue (Figure 1).9,10,28 The inflammatory response can be divided into three functional phases: initiation, lipid-mediator class switching, and resolution.15 SPMs have been identified as critical mediators in the active process of inflammatory resolution. Examples of SPMs include resolvins (Rvs), protectins (PDs), maresins (MaRs), and lipoxins (LXs).16 The shift from inflammation to resolution is characterized by the biosynthesis of prostaglandin E2 and LX, which are secreted by apoptotic neutrophilic granulocytes. The production of additional SPMs, including resolvins, is derived from circulating n-3 PUFAs (lipid-mediator class switching). SPMs have been shown to halt continued neutrophil infiltration and counter regulate pro-inflammatory cytokine signalling, while enhancing the differentiation of monocytes into M2 macrophages (anti-inflammatory), which in turn produce additional SPMs. The resolution phase aims to restore tissue homeostasis, and failure of this phase can lead to a persistent inflammatory state.15–17 Results from preclinical animal studies have implicated resolvins in cardiac arrhythmias.15 In the mixed-effects Tobit regression, which was adjusted for established POAF risk factors, CHF and CKD were associated with higher RvD1 levels (MR: 1.33 and 1.22, respectively). This finding may be indicative of a compensatory upregulation of early-phase resolution mediators in response to chronic systemic inflammation. A study by Szczuko et al. similarly suggests that resolution is counteractively upregulated in chronic inflammatory states. The researchers found a positive correlation between CRP and increased levels of the SPMs RvE1 (R: 0.395, P = 0.025) and RvD1 (R: 0.336, P = 0.057) in Hashimoto’s thyroiditis.29 In the present study, we observed that the postoperative decline in CRP trajectory modified the association between POAF and resolvin levels, highlighting the link between prolonged inflammatory states and resolution mechanisms. Specifically, patients with a slow CRP resolution exhibited higher RvD1 and lower RvD2 levels in the context of POAF. This finding suggests the possibility that elevated or prolonged systemic inflammation may amplify the regulation, potentially dysregulation, of inflammatory resolution pathways and mediators. The use of CRP as a dynamic marker underscores the potential influence of systemic inflammation on the regulation of SPMs. Further mechanistic studies are necessary to elucidate how the interplay between inflammatory and pro-resolutory pathways might influence the development of POAF. Results from a 2017 study by Rathod et al. demonstrated that females exhibit a faster resolution of inflammation than males, with a significantly higher production of SPMs, particularly those in the D-resolvin series.30 Our data support the existence of sex-specific differences in the resolution of inflammation by demonstrating distinct perioperative resolvin responses associated with POAF. The observed increase in RvD1 among POAF patients was smaller in males than in females (P = 0.012). A marginal sex interaction was also seen for RvD2, with females demonstrating a greater POAF-associated reduction in RvD2 levels (P = 0.054). Considering RvD1 as a promising future potential therapeutic agent, Hiram et al. demonstrated that RvD1 was able to prevent the development of pacing-induced atrial arrhythmias in experimental right heart disease (RHD). While arrhythmia induction was successful in 100% of the control rats with RHD, only 33% of RHD rats pre-treated with RvD1 developed arrhythmias. They furthermore showed that RvD1 attenuates fibrotic atrial remodelling.31 Another possible mechanism by which RvD1 may exert anti-arrhythmogenic effects is via the modulation of potassium currents in atrial myocytes. It has been established that potassium channel activity undergoes alterations in the context of chronic AF, resulting in reduced atrial action potentials and shortened refractory periods in the atria that favour re-entry.32 A study by De la Cruz et al. indicates that RvD1 may influence cardiac potassium currents by directly inhibiting repolarizing currents and prolonging atrial repolarization, thereby modulating atrial electrophysiology.33 This study has several strengths. First, it includes a large, heterogeneous cardiac surgery cohort comprising patients with different surgical procedures and perioperative risk profiles. This heterogeneity enhances the external validity of our findings. Secondly, to the best of our knowledge, this is the first study to examine perioperative serum levels of RvD1 and RvD2 in cardiac surgery patients. By focusing on SPMs, our study addresses a novel, mechanistically relevant aspect of POAF pathophysiology. However, certain limitations are acknowledged. The number of available serum resolvin measurements decreased after the immediate postoperative period, largely due to patient transfers out of the ICU. Consequently, later postoperative serum samples were obtained from a more selected subgroup of patients, who remained hospitalized in the ICU. Conversely, POD7 sampling included follow-up samples from patients in both the ICU and regular floor. Therefore, apparent fluctuations in the descriptive time-course figures, particularly at later postoperative time points should be interpreted with caution, as they may partly reflect changes in the composition of the sampled population rather than true biological shifts over time. As an observational study nested within a larger prospective cohort, the findings are also limited by the availability and quality of pre-existing clinical data, which may vary in documentation completeness. POAF was ascertained retrospectively based on a comprehensive review of the available clinical documentation. It is possible that brief or asymptomatic episodes were missed. Although we restricted the inclusion of covariates to avoid model overfitting, unmeasured confounders, including CPB duration, transfusion volume, and reoperation, may have influenced inflammatory resolution and arrhythmia risk. Finally, the study cohort was predominantly male (yet somewhat typical for a cardiac surgery cohort), which may limit the applicability of our findings to female patients. The identification of sex-specific interactions in resolvin responses to POAF underscores the need for further investigation into how resolvin pathways may vary between men and women. Because follow-up was limited to the in-hospital period, long-term arrhythmic outcomes and inflammatory recovery patterns beyond discharge could not be assessed. Due to the observational design of the study, it is not possible to make causal inferences; however, the results of our study provide novel insights and generate hypotheses for future interventional studies targeting inflammatory resolution agonists in the perioperative setting. Furthermore, given the limitations of our analyses to the humoral response, direct inferences regarding inflammatory and resolutory processes in atrial tissue cannot be made. This remains an important topic for future investigation. Registration German Clinical Trials Registry (DRKS No. 00017057), Date of registration 5 April 2019, www.drks.de. Conclusion Resolvins demonstrate a differential expression in POAF, marked by elevated RvD1 and reduced RvD2 levels. The correlation between POAF and RvD1 is more pronounced in female subjects, suggesting an amplified RvD1 response to POAF. These findings suggest that inflammatory resolution is not only dysregulated in POAF but may also be shaped by sex-specific biological mechanisms. Further research is needed to determine the exact role and impact of RvD1 and RvD2 on POAF development and to examine whether this sex-specific alteration in the resolution pathways contributes to the observed differences in long-term mortality between male and female POAF patients. The interaction with postoperative CRP trajectories further supports the relevance of systemic inflammatory burden and/or failed resolution mechanisms in shaping SPM dynamics. Given their active role in restoring immune homeostasis, specialized pro-resolving mediators may offer a promising avenue for biomarker development and therapeutic intervention in cardiac surgery patients at risk for POAF. Acknowledgements All of the medical illustrations were created with the BioRender software (https://BioRender.com). Contributor Information Melina Heine, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Felix Ulbrich, Department of Anesthesiology and Critical Care, Medical Center, University of Freiburg, Hugstetter Str. 55, 79106 Freiburg im Breisgau, Germany. Samuel Justice, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Babak Saravi, Department of Oral, Maxillofacial and Facial Plastic Surgery, Medical Faculty and University Hospital Düsseldorf, Heinrich-Heine-University, Moorenstraße 5, 40225 Düsseldorf, Germany. Simon A Amacher, Department of Anesthesiology and Critical Care, Medical Center, University of Freiburg, Hugstetter Str. 55, 79106 Freiburg im Breisgau, Germany. Julian Hubrich, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Leonard Simeth, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Asishana Osho, Division of Cardiac Surgery, Mass General Brigham, Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, Boston, MA 02114, USA. Charles N Serhan, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Jochen D Muehlschlegel, Department of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, 1800 Orleans St, Baltimore, MD 21287, USA. Jakob Wollborn, Department of Anesthesiology, Mass General Brigham, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115, USA. Data availability The data underlying this article will be shared on reasonable request to the corresponding author. Author contributions Melina Heine (Data curation, Writing—original draft [lead], Formal analysis [equal], Visualization [lead]), Felix Ulbrich (Data curation [equal], Supervision [supporting]), Samuel Justice (Formal analysis, Methodology [equal], Visualization [supporting]), Babak Saravi (Data curation [equal]), Simon A. Amacher (Data curation [equal]), Julian Hubrich (Formal analysis [equal]), Leonard Simeth (Formal analysis [equal]), Asishana Osho (Project administration [equal], Writing—review & editing [supporting]), Charles N. Serhan (Resources [lead], Supervision, Writing—review & editing [equal]), Jochen D. Muehlschlegel (Project administration [lead], Supervision, Writing—review & editing [equal]), and Jakob Wollborn (Conceptualization, Project administration, Supervision, Writing—review & editing [lead], Formal analysis, Methodology [equal]) Funding M.H. is funded by the Biomedical Education Program (BMEP) and the German Heart Foundation (Deutsche Herzstiftung e.V.). S.A.A. received grants from the Mach-Gaensslen Foundation Switzerland and the Nora van Meeuwen-Haefliger Foundation of the University of Basel, Switzerland. C.N.S. gratefully acknowledges grant support from the NIH, National Institute of General Medical Sciences grant 5R35GM139430. J.D.M. received support from the NIH (grants 7R01HL15040 and 701HL149998). Abiomed, Johnson&Johnson, Vivalink. References Associated Data Data Availability Statement The data underlying this article will be shared on reasonable request to the corresponding author.