The Raine Study recruited 2900 pregnant women from the general population at King Edward Memorial Hospital (Subiaco, Western Australia) and nearby private practices from 1989 to 1991 [18]. The women gave birth to 2868 live babies (Generation 2, Gen2). This study focuses on the 27-year follow-up of Gen2 (Gen2-27) in which 1084 participants volunteered, with 978 providing blood samples for measurement of SPM. The Human Research Ethics Committee of the University of Western Australia approved the study. Participation was voluntary and informed written consent was obtained.

Body weight (to the nearest 100 g) and height (to the nearest 0.1 cm) were measured using a Wedderburn Chair Scale and a Holtain Stadiometer, respectively. BMI was calculated as weight (kg)/height2 (m). Waist circumference was measured at the umbilicus level with a tape measure (to the nearest 0.1 cm). Skinfold thickness measurements were obtained from the anterior abdominal wall, subscapular, and suprailiac skinfolds with a skinfold caliper (Holtain Tanner/ Whitehouse skinfold caliper, Holtain, Crosswell, United Kingdom). Resting systolic blood pressure and diastolic blood pressure were measured after 5 min supine rest using an oscillometric sphygmomanometer (DINAMAP ProCare 100 vital signs monitor; GE Healthcare, USA). Six readings were recorded, each 2 min apart, with the last five readings averaged.

Fasting bloods were analyzed at PathWest Laboratory, Royal Perth Hospital, for serum insulin, glucose, triglycerides, total cholesterol, HDL-cholesterol, high-sensitivity C-reactive protein (hs-CRP), the liver enzymes gamma-glutamyl transferase, alanine transaminase, aspartate transaminase and alkaline phosphatase, and leukocyte count. LDL-cholesterol was calculated using the Friedewald equation [19]. Insulin resistance was estimated using HOMA-IR calculated as fasting insulin [µU/ml] × fasting glucose [mmol/L]/22.5). Fatty acids were measured in erythrocytes that were collected and stored at −80 oC until assay [20]. Plasma leptin and adiponectin were measured in plasma stored at −80 oC using Quantitkine ELISA kits for Human Leptin and Human Total Adiponectin/ Arcp30, respectively (R&D systems Inc, Minniapolis USA). The metabolic syndrome (MetS) was defined using the criteria of the International Diabetes Foundation [21].

Self-reported measures of smoking were dichotomized as a ‘yes/no’ response. Alcohol consumption and physical activity were calculated from self-reported diaries from the previous week [22]. Alcohol consumption was further dichotomized to <100grams/week or >100grams/week based on the ‘Australia guidelines to reduce health risks from drinking alcohol’ [23]. The use of hormonal contraception in females was assessed from a questionnaire that inquired about the current use of the oral contraceptive pill, implant, injection or any intrauterine hormonal contraceptive devices.

Subcutaneous and visceral abdominal fat depots were measured with a Siemens Magnetom Espree 1.5 T (Siemens AG, Erlangen, Germany) (76 × 18 Channels, Max Slew Rate 170 T/m/s) magnetic resonance imager (MRI) under the supervision of Dr Brendan Adler at Envision Medical Imaging, Perth. Participants were screened to ensure they had no contraindications to MRI prior to their scan. They were studied in the supine position with radiofrequency coils placed in the area being imaged. Images were acquired using a Multiecho Gradient Echo sequence, with a breath-hold technique. Data were analyzed at the Center for Advanced Imaging, University of Queensland. The DICOM images were converted to NIfTI files to facilitate data processing. After registering and histogram-matching the upper and lower abdomen segments, the two scans were merged to one full abdomen image. The full abdomen image was segmented into subcutaneous and visceral fat compartments using the method “vatsatseg”, a python implementation of the matlab segmentation tool “SAT_VAT_segmentation” as described by Shen et al. [24].

SPM pathway intermediates (18-HEPE, 17-HDHA and 14-HDHA), SPM and LTB4 were extracted from plasma collected into EDTA and analyzed by liquid chromatography-tandem mass spectrometry (LCMSMS) as previously described [15, 25]. Standards 14-HDHA; 18-HEPE; 17S-HDHA; RvE1; RvD1; 17R-RvD1; RvD2; 10S,17S-DiHDHA; MaR1; LXA4; LTB4; and RvD2-d5 were purchased from Cayman Chemicals (Ann Arbor, MI). PD1; RvE2; RvD3; and RvD5 standards were a gift from Professor Charles N. Serhan (Harvard Medical School, Boston, MA). RvE3 and 18R-RvE3 standards were a gift from Professor Makoto Arita (Department of Health Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo, Japan).

SPM were analyzed on a Thermo Scientific TSQ Altis Triple Quadrupole mass analyzer equipped with a HESI source attached to a Vanquish Ultra High Performance Liquid Chromatography (UHPLC) system. UHPLC separation was carried out on a Thermo Accucore C18 column (100 mm×2.1 mm, 2.6 μm particle size) at a flow rate of 400 μl/min with 5 mM ammonium acetate, pH 8.9 (solvent A) and methanol (solvent B) as mobile phases. Buffer was freshly prepared and pH was adjusted using 25% ammonia solution. All chemicals and solvents used were LCMS grade. Gradient conditions were as follows: 50% B at 0 min to 65% B at 8.2 min, 95% B at 9.2 min and held at 95% B to 11.0 min, reduced to 50% B at 12.0 min and held at 50% B to 15 min to equilibrate to starting conditions. The retention times were: RvE1 = 0.95 min, RvD2-d5 = 2.2 min, RvD2 = 2.02 min, RvD3 = 2.17 min, RvD1 = 2.56 min, LXA4 = 2.50 min, 17R-RvD1 = 2.69 min, RvE2 = 3.45 min, RvE3 = 3.53 min, PD1 = 5.26 min, 10 S,17S-DiHDHA=4.77 min, RvD5 = 5.18 min, MaR1 = 5.28, LTB4 = 5.40 min, 18-HEPE = 6.99 min, 17-HDHA = 9.62 min, 14-HDHA = 9.68 min. The total run time was 15 min.

The mass spectrometer was operated in the negative ion multiple reaction monitoring (MRM) mode using argon as collision gas. Nitrogen was used as the sheath, auxillary and sweep gas set to 50, 10 and 1 arbitrary units, and argon as the CID gas at 1.5 mTorr. The vaporizer temperature of the ESI source was 350 °C and the spray voltage was 2.5 KV. The Q1 and Q3 mass resolution of the spectrometer was 0.7 Da at full width at half maximum. SPM pathway intermediates, SPM and LTB4 were identified using all of the following criteria: (i) retention time that matched the authentic standard; (ii) MRM using two or for most SPM three product ions identified from the standards and optimized for collision energy ; and (iii) confirmation of retention time and MRM product ions in stripped plasma with added SPM standard. The MRM transitions were (precursor ion→product ions): 14-HDHA (m/z 343.175 → 281.167, 205.083, 161.083); 17-HDHA (m/z 343.188 → 281.167, 245.083, 201.155); 18-HEPE (m/z 317.200 → 259.137, 255.167, 215.137); LXA4 (m/z 351.188 → 235.137, 217.137, 114.970); LTB4 (m/z 335.175 → 317.167, 195.054, 151.054); RvE1 (m/z 349.125 → 205.000, 195.054, 161.071); RvE2 (m/z 333.175 → 315.208, 271.167, 253.167); RvE3 (m/z 333.225 → 315.167, 245.083, 201.167); 18R-RvE3 (m/z 333.188 → 315.220, 245.155, 201.155); RvD1 (m/z 375.200 → 233.155, 215.083, 141.054); 17R-RvD1 (m/z 375.150 → 233.155, 215.083, 141.071); RvD2 (m/z 375.200 → 215.137, 175.071, 141.000); RvD2-d5 (m/z 380.200 → 277.083, 175.071, 141.000); RvD3 (m/z 375.138 → 180.857, 147.054, 137.071); RvD5 (m/z 359.125 → 341.167, 297.167, 199.155); 10 S,17S-DiHDHA (m/z 359.212 → 206.083, 153.071); PD1 (m/z 359.162 → 341.238, 206.083, 153.054); MaR1 (m/z 359.212 → 341.167, 297.167, 177.054). Full mass spectrometry spectra are provided in Supplementary Figs. 1–6. Instrument control and data acquisition used Tracefinder software, version 4.1. Plasma concentrations expressed in pg/mL were determined from calibration curves constructed by spiking stripped plasma with standards and used linear regression analysis and the ratio of metabolite to internal standard (RvD2-d5) (typically R2 > 0.99). The % CV for all measured metabolites ranged from 5–15%.

Statistical analysis was carried out using SPSS version 27.0 or STATA version 15.1. SPM pathway intermediates, SPM and other variables that were not normally distributed were natural log transformed prior to analysis. Participant characteristics are described as arithmetic means and standard deviations or geometric means and 95% confidence intervals.

Sex differences in participant characteristics were assessed using unpaired t-test or Chi square analysis for categorical variables. Correlation coefficients were used to initially assess the relationship between CVD risk factors and plasma SPM pathway intermediates and SPM.

Power calculations showed that measurements in at least 20% of participants would be required to give 80% power to detect a correlation coefficient of 0.2 at p < 0.05. Consequently, only plasma 18-HEPE, 17-HDHA, 14-HDHA, RvE1 and RvE3 fulfilled these criteria.

Pearson coefficients were calculated separately for each sex for plasma 18-HEPE, 17-HDHA, 14-HDHA, RvE1 and RvE3 and selected omega-3 (EPA, DHA, omega-3 index) and omega-6 (AA) fatty acids, and relevant cardiovascular risk factors (adiposity, alcohol consumption and smoking). Given only RvE1 showed consistent association with cardiovascular risk factors, the effect of BMI (categorized as healthy, overweight, or individuals with obesity), smoking (yes/no) and alcohol consumption (≤100 g/wk v’s >100 g/wk) on plasma RvE1 was assessed using univariate analysis that adjusted for sex differences. This was further explored using multiple regression analysis to assess the significant predictors of plasma RvE1 with a cut-off p > 0.1 applied to covariates that were not significantly associated with the outcome. Covariates considered in the model included sex and smoking as categorical variables; BMI, alcohol intake, AA and DHA, (ln) hs-CRP, (ln) triglycerides, cholesterol, adiponectin, and physical activity as continuous variables.