Per- and polyfluoroalkyl substances (PFAS) are a family of widespread, environmentally persistent compounds that have been used in numerous industrial, household and agricultural applications or products since their discovery in the 1930s (Fremlin et al., 2023; Ricolfi et al., 2024). Because they integrate both industrial and residential sewage, wastewater treatment plant (WWTP) effluents and residual biosolids have been identified as important sources of PFAS to the environment (Ankley et al., 2021; Custer et al., 2024). Aqueous film-forming foams (AFFF), used as fire extinguishing agents, primarily at military and commercial airports and firefighting training facilities, are another major source of PFAS contamination throughout the United States and elsewhere (Ruyle et al., 2021). Perfluoroalkyl carboxylic acids (PFCAs) and perfluoroalkane sulfonic acids (PFSAs) are the most widely monitored PFAS sub-groups and are the primary PFAS that are targeted by regulatory agencies worldwide (Ankley et al., 2021). Despite efforts to phase out the manufacturing and use of medium- and long-chain PFAS (e.g., C6 – C12 PFSA), many PFAS are still commonly used (Buck et al., 2011). The persistence and solubility of PFAS, particularly long-chain PFAS, have resulted in their detection in wildlife on a global scale (Ankley et al., 2021; Giesy & Kannan, 2001; Houde et al., 2011). Furthermore, studies in humans and wildlife report associations between PFAS and disturbances to the endocrine and immune systems (DeWitt et al., 2019; Fenton et al., 2021; Woodlief et al., 2021).
Chesapeake and Delaware Bays are located in the mid-Atlantic region of the eastern United States and provide important habitat for wildlife, including birds of prey such as osprey (
Osprey are long-lived piscivorous birds, and in view of their high trophic feeding level, widespread distribution and ability to adapt to human landscapes, they have been used as indicators of aquatic ecosystem health (Grove et al., 2009). Osprey exhibit nest site fidelity, their nests are highly visible and often easy to access, and they tolerate short-term disturbance during incubation and brooding of young. Osprey eggs, and blood and feather samples from nestlings, have been used extensively to monitor spatial and temporal variation of environmental contaminant exposure and adverse effects (Golden and Rattner, 2003; Henny et al., 2010). Research suggests that PFAS, especially long-chain forms, may bioaccumulate and biomagnify within food webs, thus potentially affecting the health of apex consumers such as raptors (Fremlin et al., 2023; Miranda et al., 2022; Munoz et al., 2022). In birds, PFAS are known to be transferred from hens to their eggs, with long chain PFAS being preferentially transferred (Jouanneau et al., 2022; Ricolfi et al., 2024). Avian offspring have been demonstrated to average 41% higher concentrations of PFAS than their parents (Ricolfi et al., 2024), likely integrating
In 2011 and 2015, two large scale studies in Chesapeake and Delaware Bays examined temporal trends of a variety of legacy contaminants, including OCs, polybrominated diphenyl ethers (PBDEs), and PCBs in osprey eggs, and investigated potential relationships between oxidative DNA damage in blood of nestlings and contaminant exposure (Lazarus et al., 2015; Rattner et al., 2018). In the present study, we quantified plasma PFAS in these same osprey samples collected from both bays, as well as health indicators (plasma clinical chemistry, thyroid hormones and several immune function measurements) to investigate if exposure to PFAS may be associated with adverse effects. We predicted that PFAS concentrations would be higher in osprey nestlings sampled near major urban areas and health indicators in nestlings would differ based on PFAS exposure profiles.
In-depth descriptions of study sites can be found in previously published papers (Lazarus et al., 2015; Rattner et al., 2018). Briefly, ospreys nesting on navigational markers, platforms, duck blinds, and other structures were sampled during nesting seasons (March-August) in 2011 (Chesapeake Bay) and 2015 (Delaware Bay). In the Chesapeake Bay, samples were collected from nests found within three regions, encompassing parts of Washington, D.C. (DC), northern Virginia (VA) and Maryland (MD) including (i) the Patapsco and Back Rivers in Baltimore, MD (
Nests in the Delaware Bay watershed were spread across three segments within Delaware (DE) and Pennsylvania (PA). These included: (i) South (Delaware Inland Bays: Indian River Bay north through Rehoboth Bay;
Blood samples were collected from osprey nestlings (40-45 d old) at each surveyed study nest. Nestlings (one/nest) were removed for processing as described in Lazarus et al. (2015) and Rattner et al. (2018). One nestling from each nest was selected at random: in several cases, only one nestling was present. Although egg laying order may affect deposition of contaminants (Jouanneau et al., 2022), we were unable to link egg laying order to individual nestlings. A 5-7 mL brachial blood sample was collected using a 23-gauge, 25.4 mm needle and a heparinized syringe (Sarstedt International). Samples were then centrifuged at 1060 ´
Plasma was analyzed for 40 PFAS congeners (see online supplementary material, Table S1) by SGS AXYS Analytical Services (Sidney, British Columbia, Canada) using SGS AXYS method MLA-110, which is based on USEPA method 1633 (SGS AXYS Analytical Services, 2021; US Environmental Protection Agency, 2021). Sample concentrations were determined by isotope dilution/internal standard quantification. Briefly, after thawing, 13C-labelled surrogate standards were added to the samples, which were then extracted via solid phase extraction using weak anion exchange cartridges (Waters, Milford, MA, USA). Samples were analyzed by liquid chromatography-mass spectrometry (LC-MS/MS) on an ultrahigh performance liquid chromatography (UPLC-MS/MS) reversed phase C18 column coupled to a triple quadrupole LC-MS/MS. Samples were analyzed in batches (20 samples), with each batch containing lab blanks and spiked reference matrices for quality assurance and control. The limit of quantification (LOQ) for the PFAS analytes ranged from 0.01 to 59.5 ng/g (see online supplementary material, Table S1). Spike recoveries averaged 99.0% and ranged from 67.5% to 177.0% for all analytes except 3:3 FTCA (mea
All PFAS chemical family terminology used in this manuscript follow the guidelines of Buck et al. (2011). Total PFAS (∑PFAS) are the sum of all PFAS congeners detected in a sample. Total perfluoroalkyl sulfonic acids (∑PFSA) and total perfluoroalkyl carboxylic acids (∑PFCA) refer to the sum of congeners within each category. Total precursors (∑Prec) refer to the sum of the following congeners: fluorotelomer sulfonic acids (FTSs), fluorotelomer carboxylic acids (FTCAs), and perfluoroalkane sulfoamido derivatives (methyl perfluorooctane sulfonamidoacetic acid (MeFOSAA), ethyl perfluorooctane sulfonamidoacetic acid (EtFOSAA), N‑methyl perfluorooctane sulfonamidoethanol (N-MeFOSE), N‑ethyl perfluorooctane sulfonamidoethanol (N-EtFOSE), perfluorooctane sulfonamide (PFOSA)). If an individual PFAS concentration was <LOQ, it was assumed to be 0 when calculating the final summation categories. Thus, only those PFAS that were found above the LOQ were included in the totals.
Total immunoglobulin Y (IgY) analysis was performed according to Fassbinder-Orth et al. (2016) with modifications. Standards were prepared from IgY extracted using the Pierce Chicken IgY Purification Kit (ThermoFisher, Waltham, MA, USA) from addled American kestrel (
For the IgY assay, standards ranging from 25 to 300 ng/mL were prepared from the purified kestrel IgY by serial dilution in coating buffer (0.006 M Na2CO3, 0.044 M NaHCO3, pH 9.3). Osprey plasma samples were serially diluted 1:15,000 with coating buffer and 75 μL aliquots of each sample were added in triplicate to a flat-bottomed 96-well plate, and incubated overnight at 4°C. The following day, plates were brought to room temperature, and the coating solution was removed. The plates were washed three times with wash buffer (50 nM Tris buffered saline, 0.05% Tween, pH 8.0), blocked with 200 μL/well blocking buffer (50 mM Tris buffered saline, 1% bovine serum albumin, 0.05% Tween, pH 8.0), covered, and incubated at room temperature for 30 min with gentle shaking. Plates were washed again four times with wash buffer. Polyclonal goat anti-bird IgY-HRP conjugated antibody (Bethyl Laboratories, Montgomery, Texas, USA; A140-110P) diluted 1:10,000 with blocking buffer was then added (100 µl) to each well, incubated for 1 h at 37°C, and washed four times. Tetramethylbenzidine (TMB)-peroxidase substrate (Bethyl Laboratories) was then added (100 µl) to each well and incubated in the dark at room temperature for 9.5 minutes. The reaction was stopped by adding 100 μL of 0.18 M H2SO4. Plates were read immediately at 450 and 630 nm using a BMG FLUOstar Omega Microplate Reader (BMG Labtech, Ortenburg, Germany). Absorbance at 630 nm was subtracted from that at 450 nm and the difference used for calculating IgY concentration using a 4-Parameter logistic curve fit.
Quality assurance/quality control included analysis of blanks, assessment of linearity and intra- and inter-assay variability. Four osprey plasma samples were used as reference samples across all plates. The mean intra-assay variability (%CV) was 6.6% and mean inter-assay variability was 10.3%. The assay LOQ was 25 ng/mL, and the limit of detection (LOD) was 3.43 ng/mL.
Haptoglobin is an acute-phase protein induced by inflammatory cytokines and produced by the liver (Matson et al., 2012). Haptoglobin-like activity was measured in osprey plasma to determine baseline concentrations and evaluate potential variation associated with sites or contaminant exposure. Haptoglobin-like activity was quantified using a commercially available functional assay (TP801; Tri-Delta Development Ltd., Ireland) that colorimetrically quantifies the heme-binding capacity of the plasma (Matson et al., 2012). Manufacturer instructions were followed with half-volume modifications. Briefly, for all samples, 5 μL of diluent (PBS, pH 7.6) were added to 20 μL of plasma. Next, 6 μL of each plasma sample or standard was deposited on a flat-bottomed half volume 96-well plate (Costar) in duplicate. Fifty microliters of a stabilized hemoglobin solution (Reagent 1) and 70 μL of a chromogen solution (Reagent 2) were added to each well. Plates were then agitated and incubated for 5 min at 22.5°C and then read at 580 nm on a BMG FLUOstar Omega Microplate Reader. Standard curves were obtained from serial dilutions and included seven standards ranging from 0.039 to 0.625 mg/mL.
Quality assurance/quality control was as described for IgY except that two osprey plasma samples were used to assess inter-assay variability. The mean intra-assay variability (%CV) was 4.2% and mean inter-assay variability was 1.6%. The assay LOQ was 0.039 mg/mL and the LOD was 0.031 mg/mL.
Plasma samples were analyzed using the Avian-Reptile Profile Plus rotor on a VetScan VS2 (Zoetis, Parsippany, New Jersey, USA) for 12 clinical chemistry analytes, including albumin in g/dL (ALB), aspartate aminotransferase in units/L (AST), bile acids in μmol/L (BA), calcium in mg/dL (Ca++), creatine kinase in units/L (CK), globulin in g/dL (GLOB), glucose in mg/dL (GLU), potassium in nmol/L (K+), sodium in nmol/L (Na+), phosphorus in mg/dL (PHOS), total protein in g/dL (TP), and uric acid in mg/dL (UA). Frozen archived samples (stored at -80°C) were thawed in a refrigerator and processed within the same day. The rotors were loaded with 100 μL of plasma and analyzed per the manufacturer’s instructions. The analyzer includes internal quality checks of the instrument and controls within each individual rotor. Samples were checked for physical interferences (e.g., hemolysis, lipemia) and no samples were excluded due to quality check failures.
Thyroid function of osprey nestlings was assessed by examining plasma hormone levels. Plasma concentrations of total triiodothyronine (T3) and thyroxine (T4) were determined via enzyme-linked immunosorbent assay (ELISA) using the AccuBind® ELISA T3 Kit (Monobind Inc., Lake Forest, CA, USA) and the NT4 AccuBind ELISA kit (Monobind Inc.) as previously described in Eng et al. (2019). Intra-assay CVs were 5.7 ± 3.9% (mean ± SD) for T3 and 4.5 ± 3.9% for T4. Inter-assay variation was 2.9 ± 2.2% for T3 and 3.8 ± 3.1% for T4, based on two reference samples. The LOD was 0.09 ng/mL and 0.29 ng/mL for T3 and T4. The LOQ was 0.16 ng/mL and 0.78 ng/mL for T3 and T4.
Oxidative DNA damage as indicated by the concentrations of 8-hydroxy-2’deoxyguanosine (8-OHdG) was assayed in whole blood as described and originally reported in Lazarus et al. (2016) and Rattner et al. (2018). These previously published values were used in some PFAS exposure comparisons in the present study.
The complete dataset is available as a U.S. Geological Survey data release (Karouna-Renier et al., 2025). All data analyses were performed in PRIMER / PERMANOVA+ Ver. 7.0.24 (Anderson et al., 2008) or R version 4.2.1 or 4.4.1 (R Core Team, 2021) and α was set to 0.05. For PFAS congeners that were detected in >50% but <100% of samples, descriptive statistics were estimated using the enparCensored procedure (Kaplan-Meier method) in the ‘EnvStats’ package in R (Helsel, 2012; Millard, 2013). Because sex of the nestlings was not equally distributed among sampling locations (see online supplementary material, Table S2), we used nestling weight as a surrogate covariate for sex in models described below. Analysis by Permutational Multivariate Analysis of Variance (PERMANOVA) indicated that female nestlings (1693.6 ± 99.8 g; mean ± SD) weighed more than male nestlings (1405.8 ± 93.1 g) (Pseudo-F1: 113.43, p= 0.0001).
Multivariate analyses of PFAS patterns were performed as described by Helsel (2012) on uscores to account for multiple detection limits within each PFAS congener. Only PFAS that were detected in at least two samples were included in the analysis (Custer et al., 2024). Uscores were generated using the NADA2 package in R (Julian & Helsel, 2024) and imported into PRIMER / PERMANOVA+, where Euclidean distance matrices were constructed on the uscores. The respective distance matrices were then used for overall comparison of PFAS patterns between bays and within bays by nonmetric multidimensional scaling (NMDS) and analysis of similarity (ANOSIM). Additional evaluation of potential relationships between PFAS and spatial factors was conducted using the uscores-based Euclidean Distance matrices and analyzed by PERMANOVA in PRIMER / PERMANOVA+ (Anderson, 2001). For all PERMANOVA tests, a total of 9999 permutations of residuals under a reduced model and Type 1 Sums of Squares were used. Nestling weight was included as a covariate but removed if it was not significant in the model. Because multivariate analyses found associations among PFAS and spatial factors, univariate analyses were also used to examine relationships. We used the non-parametric Peto-Peto test (cen1way) in R package NADA2 (Julian & Helsel, 2024) to determine if concentrations of individual PFAS (with ≥50% detection) and concentrations of ∑PFAS, ∑PFCA, ∑PFSA, and ∑Prec differed among sampling regions within each bay.
NMDS and ANOSIM (PRIMER) were used to preliminarily explore whether the overall pattern of health indicators differed between bays. All health indicator data were standardized to a common scale with mean 0 and variance 1, and these values were used to construct a Euclidean distance matrix. Similarly, for within bay comparisons, we compared patterns among sampling regions using the same methods. One outlier sample from Chesapeake Bay was excluded due to analytical concerns with plasma biochemistry. Although samples in the two bays were collected 4 years apart (Chesapeake in 2011 and Delaware in 2015), they were analyzed simultaneously for all endpoints. The exploratory multivariate procedures identified differences in health indicators between the two bays (see online supplementary material, Figure S1), but we could not rule out the effects of sampling year driving these differences. Therefore, additional analyses were performed separately within each bay. Additionally, PERMANOVA, based on the health indicator Euclidean distance matrices for each bay, was used to investigate relationships between health indicators and spatial factors. Nestling weight was included as a covariate. Each global PERMANOVA was followed by pair-wise comparisons among sampling regions in PRIMER / PERMANOVA+ (Anderson, 2001).
Further multivariate analyses were used to evaluate potential relationships within each bay among global PFAS concentrations and health indicators obtained for each osprey nestling. A principal coordinate analysis (PCoA) was used to extract summary ordination axes from the Euclidean distance matrix created from PFAS uscores for each bay separately. Principal coordinates (PCs) that contributed to 94% of the total variation in the PFAS dataset were considered in downstream analyses. Next, a stepwise distance-based linear model (DISTLM) (McArdle & Anderson, 2001) was used to determine the most parsimonious combination of PCs to explain variation in overall health indicator patterns (Euclidean Distance matrix). Each predictor variable (PC) was entered into the model one at a time, while adding or removing other variables and the final model selected based on AICc (Akaike’s Information Criteria corrected for small sample sizes). The corresponding statistical results indicate the increase in proportion of explained variation attributable to each variable that is retained. If no model with significant predictors was identified in DISTLM, the predictor variables collectively explained little to no variation in the response data. If significant relationships were observed among multiple PCs and the health indicator Euclidean distance matrices, additional PERMANOVAs (Type 1 Sum of Squares, 9999 permutations) were employed to determine (1) if variation in health indicators among individuals within sampling regions were related to PFAS and (2) if spatial variation in health indicator profiles could be attributed to PFAS concentrations. Similarly, we examined relationships between specific health indicators (Euclidean Distance matrix) and individual PFAS detected in >70% of samples or summary PFAS categories along with nestling weight in each bay using DISTLM. Marginal tests examined individual relationships and were followed by sequential tests using a stepwise DISTLM procedure to identify (based on the AICc) the most parsimonious combination of PFAS (if any) or separately, summed PFAS categories, which may account for the variability observed in each individual health indicator.
All samples were evaluated for 40 PFAS analytes, with 22 of the 40 PFAS analytes detected in at least one sample. Descriptive statistics for these 22 PFAS, including detection frequencies, concentration ranges, and means are reported in Tables 1 and 2. Total PFAS (∑PFAS) ranged from 88.9 to 1332 ng/mL in osprey nestlings (Tables 1 and 2). Detection frequencies of PFAS congeners varied by bay and site (nest), with perfluorohexane sulfonic acid (PFHxS), perfluoroheptane sulfonic acid (PFHpS), PFOS, perfluorodecane sulfonic acid (PFDS), perfluorononanoic acid (PFNA), PFDA, PFUnA, perfluorododecanoic acid (PFDoA), perfluorotridecanoic acid (PFTrDA), and perfluorotetradecanoic acid (PFTeDA) present in at least 94% of the plasma samples. PFOS was the most commonly detected PFAS, accounting for 50.8 to 72.5% of ∑PFAS concentration in the osprey samples (Figure 2). The second and third largest contributors to osprey PFAS profiles were PFUnA (6.5 to 15.7%) and PFDA (7.0 to 11.9%), depending on sampling location. In osprey nestlings from Chesapeake Bay, PFOS tended to be a higher contributor to PFAS profiles (65.9%) compared to nestlings from Delaware Bay (55.4%). In contrast, long-chain PFCAs such as PFUnA and PFDA comprised larger portions of ∑PFAS in osprey from Delaware Bay relative to birds from the Chesapeake Bay. Total precursors (∑Prec) constituted more than 1.0% of the PFAS profile only in nestlings from PBR, Poplar Island, and DE North.
For Chesapeake Bay, seven PCs accounted for >94% of variation in osprey plasma PFAS concentrations among sampling regions, with PC1 and PC2 explaining 75.3% of the variability (Figure 4). PFHxA and 6:2 FTS were closely associated with Poplar Island samples, showing strong positive relationships with PC1 and PC2, whereas PFOSA and 7:3 FTCA were negatively associated with PC2 and drove the separation of the PBR samples from the other locations. For the Delaware samples (Figure 5; see online supplementary material, Table S3), PCs 1 to 6 accounted for 94% of the variation in plasma PFAS, with PC1 and PC2 explaining 79.0% of the variability. Separation among the North, Central and South samples was driven by multiple PFAS that were negatively associated with PC1. PFOA and PFNA were negatively associated with PC2.
Within Chesapeake Bay samples, PFNA, PFDA, PFDoA, PFTrDA, PFUnA, PFOS, PFDS, PFTeDA, PFHxS, PFHpS, ∑PFAS, ∑PFSA, ∑PFCA, and ∑Prec differed among sampling regions (Table 3). Apart from ∑Prec, individual PFAS were highest in samples collected from APR (Table 1). Within the Delaware samples, PFOA, PFDA, PFNA, PFDoA, PFDS, PFHpS, PFOS, PFOSA, PFTeDA, PFTrDA, PFUnA, ∑PFAS, ∑PFSA, ∑PFCA, and ∑Prec, differed among sampling locations (Table 2), with PFAS concentrations generally highest at the northern locations and decreasing downstream to the nesting locations in the southern section (Table 2).
Summary statistics for health indicators are provided in Table 4. Health indicator profiles within Chesapeake Bay did not differ among sampling regions (ANOSIM: r = 0.051,
Examination of marginal tests within the DISTLM indicated that within Chesapeake Bay only PCoA axes 1 (
Although in a multivariate framework, PFAS (as represented by the PCs) were only limited predictors of the overall health indicator profiles, we also examined potential associations at a finer scale between individual PFAS or summary categories and individual health indicators. In Chesapeake Bay, marginal tests identified several predictors of Ca (PFOS, ∑PFSA, ∑PFAS), DNA Damage (∑Prec), GLU (PFOS, ∑PFSA), HAPT (PFDA, PFUnA, ∑PFCA), K (PFTrDA, PFDoA, ∑Prec), Na (∑Prec), TP (PFOS, ∑PFSA, ∑PFAS), and UA (PFDS, PFDA, ∑PFCA) (see online supplementary material, Table S4a). However, the total variance explained by each individual or summed PFAS predictor was always <30%. When combining predictor PFAS variables to explain the highest proportion of variation in each health indicator, more parsimonious models (explaining a higher proportion of variability compared to the single predictors) were identified only for GLOB, GLU, HAPT, and Na (see online supplementary material, Table S4b) and marginal improvement was observed for TP, T3, and ALB (see online supplementary material, Table S4b) within Chesapeake Bay.
In the Delaware samples, associations between individual or summed PFAS and health indicators were also limited. Marginal tests identified individual explanatory variables for AST (PFOS, ∑PFSA), Ca (PFOSA, ∑Prec), DNA Damage (PFOSA), GLU (PFNA), Hapt (PFOSA, ∑Prec), and T3 (PFUnA, PFDoA, PFTeDA, PFTrDA, ∑PFCA) but each of these PFAS predictors explained only 14 to 22% of the variability in the individual health indicators (see online supplementary material, Table S5a). Using a combination of predictors improved the variability explained for AST, Ca, DNA Damage, GLOB, GLU, Hapt, IgY, PHOS, T4, and TP. These combinations accounted for 25 to 53% of variability in these health indicators (see online supplementary material, Table S5b).
Analysis of plasma from osprey nestlings demonstrated widespread PFAS presence throughout the Chesapeake and Delaware Bay watersheds. All samples had detectable levels of PFOS, PFNA, PFDA, PFDoA, PFTrDA, PFUnA, and PFDS. As expected, osprey nestlings from Chesapeake Bay and Delaware Bay exhibited PFAS concentration profiles that were dominated by PFOS. PFOS is the most commonly detected PFAS in birds, and in raptors generally occurs at the highest concentrations (e.g., Dykstra et al., 2021; Elliott et al., 2019; Eriksson et al., 2016; Route et al., 2014; Strom et al., 2025; Wu et al., 2020). Previous studies of PFAS in osprey eggs from the approximate locations in this study also found that PFOS was the most commonly detected congener and represented the highest proportion of examined PFAS (Rattner et al., 2004; Toschik et al., 2005). Similarly, recent data from multiple fish species in both Chesapeake and Delaware Bay watersheds also indicated that PFOS comprised the largest portion of PFAS profiles (Blazer et al., 2021; Conkle, 2024; MacGillivray, 2021).
Overall, PFOS concentrations in osprey nestlings from both bays were higher than those reported in raptors from Europe (see review by Gonzalez-Rubio et al., 2021). However, the levels of PFOS and ∑PFAS in the present study were similar to or slightly lower than those reported for bald eagle nestlings from Wisconsin and Minnesota, USA (Dykstra et al., 2021; Strom et al., 2025). While another long-chain PFSA, PFDS, was detected in every plasma sample, it comprised a small fraction of the PFAS profile in our sampled osprey nestlings (0.48 to 3.98% of ∑PFAS across all sites). The highest concentrations of PFDS (11.7 ng/mL) were found in nestlings in the APR region but were much lower than those measured in bald eagle nestlings sampled in Wisconsin and Minnesota, USA (Dykstra et al., 2021).
Osprey PFAS profiles within each bay differed from region to region. The highest concentrations of PFAS were detected in birds collected from highly urbanized and/or industrialized locations (North region of the Delaware, APR in the Chesapeake Bay). Osprey generally forage in areas up to 16 km from their nests (Poole, 2019; see online supplementary material, Figure S4) and their PFAS profiles are likely to reflect dietary contamination in these areas. Our findings are similar to the trend observed in bald eagle and peregrine falcon nestlings in which the highest PFAS concentrations were observed for nests in locations heavily influenced by industry or urban development (Dykstra et al., 2021; Strom et al., 2025; Sun et al., 2020).
In Delaware, there was a distinct decreasing gradient of PFAS concentrations in osprey nestling plasma moving from north to south with the lowest concentrations found in nests along the ocean near the mouth of Delaware Bay. Recent studies of fish also reported greater concentrations and more variable forms of PFAS in the Delaware River near River km 153 to 77.6, which correspond with our Delaware North nests (Conkle, 2024; MacGillivray, 2021). While collected across different years, these fish samples bracket our sampling window and are therefore likely a good indicator of the concentrations and patterns to which our osprey were exposed. The nestling with both the highest PFOS (1070.0 ng/mL) and ∑PFAS (1300.5 ng/mL) concentrations in our study was from a nest found just upstream of the location (river KM 189.56) in Burlington, NJ/ Bristol, PA identified by Conkle (2024) as having the highest ∑PFAS of all water samples collected in 2021, which suggests a possible point source in this section of the Delaware River. The Delaware North nestlings were also the only samples with detectable levels of the fluorotelomers 5:3 FTCA and 7:3 FTCA. These compounds were also reported in channel catfish (
PFAS concentrations in the Chesapeake Bay samples also reflect the likely foraging locations of the osprey parents (see online supplementary material, Figure S4). The APR sampling locations are proximal to or downstream of Washington, D.C. and generally exhibited the highest PFAS concentrations of the three Chesapeake sampling regions. This region includes several potential point and aggregation sources of PFAS including military installations, airports, and wastewater treatment plants. In 2021, the Maryland Department of the Environment examined surface water and whole fish tissue concentrations from Piscataway Creek, a tributary that enters the Potomac River near our sampling locations and which has known point-sources of AFFFs (MDE, 2021). Water and fish tissue collected downstream of the point source exhibited elevated levels of multiple PFAS. Custer et al (2024) reported tree swallow (
While plasma from nestlings sampled at Poplar Island had the lowest ∑PFAS concentrations in Chesapeake Bay, their PFAS profile exhibited the highest concentrations of ∑Prec of the three Chesapeake regions. These samples only contained 6:2 FTS, a component of some AFFFs but also the main transformation product of other perfluoroalkyl acids (PFAA) precursors in AFFF, such as 6:2 fluorotelomer thioether amide sulfonate (6:2 FtTAoS) and 6:2 fluorotelomer sulfonamide alkylbetaine (6:2 FTAB) (Hu et al., 2025). Poplar Island is an environmental restoration project that has been rebuilt using dredged material from channels near the Port of Baltimore (
Several long-chain PFCAs (≥ 8 carbon chain length), including PFNA, PFDA, PFUnA, PFDoA, and PFTrDA, were detected in all osprey nestling plasma samples. Chain length is generally thought to strongly affect bioaccumulative potential of PFAS, where long-chain PFAS will more readily bioaccumulate in a food web (Hopkins et al., 2023; Lewis et al., 2022). Our findings are similar to those reported by Blazer et al. (2021), who found PFDA, PFUnA, and PFDoA in all smallmouth bass samples (
In general, plasma biochemistry parameters fell within previously reported reference intervals for osprey nestlings (Meredith et al., 2012). However, mean Ca and GLOB from both bays were at or below the expected lower limit of the interval (2.5% percentile; Meredith et al., 2012). Reference data for GLU and PHOS are not available in the literature for osprey but our values fell within or just higher than refence intervals obtained for nestling golden eagles (
While health indicator profiles did not differ among sampling regions within each bay, our samples did show limited relationships between several health indicators and individual PFAS. Overall, the proportion of variability in individual health indicators that was explained by each individual PFAS was low (<30%). In Delaware, T3 levels were negatively associated with individual long-chain perfluoroalkyl carboxylic acids (Delaware = PFDoA, PFTrDA, PFTeDA, PFUnA) but no improvement in explained variation was observed with any combination of PFAS analytes in additional models. Furthermore, nearly 50% of the variability in plasma T4 in the Delaware samples was accounted for by the combination of nestling weight, PFOSA and PFUnA. In contrast, in Chesapeake Bay, a combination of PFDoA, PFTeDA, PFDS, and PFNA accounted for 59.8% of the variability in plasma T3 levels. Multiple studies have examined the relationships between thyroid hormones and PFAS in wild birds, and many have noted both positive and negative associations (Aune et al., 2024; Choy et al., 2022; Sebastiano et al., 2023), whereas others, such as Custer et al. (2024), found no association between PFAS and thyroid hormone levels. However, the specific PFAS that are identified as the most important drivers of thyroid hormone changes vary and the direction of the relationships between thyroid hormones and individual PFAS can differ among species (Sebastiano et al., 2023). These studies and others have suggested that alteration of the hypothalamic-pituitary-thyroid (HPT) axis by PFAS could involve binding site competition between PFAS and thyroid hormones, interference with transport proteins, and/or hormone homeostasis via effects on deiodination of T4 to T3 (Coperchini et al., 2021; Sebastiano et al., 2023; Sun et al., 2021). The underlying mechanisms or causes of the complex relationships observed in birds across these studies and ours are unclear but likely reflect complex interactions with other contaminants and stressors to which the nestlings have been exposed (Lazarus et al., 2015; Rattner et al., 2018).
PFAS are known to bind plasma proteins, and in humans, PFAS with carbon chain length ≥ 7 have been shown to preferentially bind globulins, whereas PFAS with < 7 carbons preferentially bind to albumin (Fischer et al., 2024). We did not observe notable relationships between ALB or GLOB and PFAS in the osprey nestlings and total protein levels (which include ALB and GLOB) in Chesapeake nestlings had limited inverse relationships with ∑PFAS, ∑PFSA, and PFOS. Flo (2016) also found limited negative correlations between PFAS and plasma proteins in white-tailed eagle nestlings (
Serum-based toxicity reference values (TRV) have been developed for PFOS (Gobas et al., 2020; Newsted et al., 2005; Simcik & Bursian, 2021) based on avian reproduction studies in Japanese quail (
Among nestlings, PFOS, PFDS, PFNA, PFDA, PFUnA, PFDoA, and PFTrDA were detected in all plasma samples, and PFOS was detected at the highest concentrations. Nestling plasma PFAS concentrations and forms of PFAS differed among study locations. Highest ∑PFAS concentrations were found at the more highly populated and industrialized sites - the Anacostia and middle Potomac Rivers (Chesapeake Bay) and lower Delaware River (our North site). The lowest ∑PFAS concentrations were observed at Poplar Island (Chesapeake Bay) and the southern portion of the Delaware Bay (South). Plasma biochemistry markers were generally within published reference intervals. In general, plasma PFAS concentrations accounted for limited proportions of the variability in the monitored health indicators in the osprey nestlings, suggesting that either the indicators we measured in this study are not responsive to PFAS or plasma concentrations for most of the nestlings are below the exposure threshold that elicits a response. PFOS concentrations in nestlings from APR in the Chesapeake and Delaware North exceeded the northern bobwhite quail TRV, although these results must be interpreted with caution, as no TRVs have been developed for raptors. Osprey nestlings throughout both watersheds are exposed to a variety of additional anthropogenic pollutants and stressors that may also influence their overall health and fitness (Lazarus et al., 2015; Rattner et al., 2018) making it harder to isolate the specific role for each PFAS.
Osprey (
PFAS bioaccumulation profiles (percent contribution of all detected PFAS) for osprey (
Nonmetric multidimensional scaling plot of uscores generated from PFAS detected in osprey (
The most significant principal coordinates (PC) produced from a principal coordinate analysis (PCoA) of PFAS profiles in Chesapeake Bay osprey (
The most significant principal coordinates (PC) produced from a principal coordinate (PCoA) analysis of PFAS profiles in Delaware Bay osprey (
# C+F = number of carbons with fluorine bonds; % >LOQ = percentage of samples exceeding limit of quantitation (LOQ); mean ± standard error of the mean (SEM) (ng/mL); ACT = average contribution to total detected PFAS.
| # C+F | APR ( |
PBR ( |
Poplar ( |
||||||||||
| % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | ||
| PFHxS (C6) | 6 | 100% | 1.30 ± 0.2 | 0.73 - 2.56 | 0.3% | 100% | 0.69 ± 0.1 | 0.3 - 1.25 | 0.4% | 100% | 0.87 ± 0.2 | 0.61 - 1.27 | 0.6% |
| PFHpS (C7) | 7 | 100% | 1.08 ± 0.1 | 0.65 - 2.38 | 0.2% | 100% | 0.65 ± 0.1 | 0.32 - 1.17 | 0.3% | 100% | 0.59 ± 0.1 | 0.43 - 0.74 | 0.4% |
| PFOS (C8) | 8 | 100% | 373 ± 24.9 | 260 - 479 | 72.5% | 100% | 118.5 ± 5.7 | 84.8 - 140 | 63.7% | 100% | 95.37 ± 7.03 | 82.7 - 107 | 61.4% |
| PFNS (C9) | 9 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| PFDS (C10) | 10 | 100% | 9.64 ± 0.5 | 7.03 - 11.70 | 1.9% | 100% | 6.04 ± 0.5 | 3.05 - 7.99 | 3.2% | 100% | 2.16 ± 0.2 | 1.74 - 2.43 | 1.4% |
| PFHxA (C6) | 6 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 100% | 0.44 ± 0.07 | 0.31 - 0.55 | 0.3% |
| PFOA (C8) | 8 | 46% | 0.28* ± 0.03 | <LOQ - 0.4 | 0.1% | 22% | 0.32* ± 0.03 | <LOQ - 0.4 | 0.2% | 0% | <LOQ | <LOQ | - |
| PFNA (C9) | 9 | 100% | 9.93 ± 1.1 | 5.4 - 15.4 | 1.9% | 100% | 5.79 ± 0.5 | 2.83 - 7.3 | 3.1% | 100% | 12.8 ± 2.9 | 7.00 - 16.9 | 8.3% |
| PFDA (C10) | 10 | 100% | 55.5 ± 7.6 | 26.5 - 90.9 | 10.8% | 100% | 14.7 ± 0.9 | 11.5 - 18.4 | 7.9% | 100% | 10.8 ± 1.7 | 7.68 - 13.6 | 7.0% |
| PFUnA (C11) | 11 | 100% | 33.5 ± 2.6 | 20.8 - 47.2 | 6.5% | 100% | 19.1 ± 0.9 | 14.1 - 21.9 | 10.3% | 100% | 11.8 ± 1.9 | 8.50 - 15.2 | 7.6% |
| PFDoA (C12) | 12 | 100% | 15.8 ± 2.2 | 10.5 - 34.0 | 3.1% | 100% | 7.99 ± 0.8 | 4.55 - 12.3 | 4.3% | 100% | 2.31 ± 0.3 | 1.77 - 2.82 | 1.5% |
| PFTrDA (C13) | 13 | 100% | 8.19 ± 1.1 | 5.1 - 17.2 | 1.6% | 100% | 7.1 ± 0.6 | 4.39 - 8.97 | 3.8% | 100% | 2.13 ± 0.2 | 1.68 - 2.44 | 1.4% |
| PFTeDA (C14) | 14 | 100% | 3.36 ± 0.9 | 1.39 - 11.3 | 0.7% | 100% | 1.94 ± 0.2 | 1.09 - 3.14 | 1.0% | 100% | 0.49 ± 0.07 | 0.39 - 0.62 | 0.3% |
| 6:2 FTS (C8) | 6 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 100% | 15.4 ± 2.1 | 12.8 - 19.8 | 9.9% |
| 8:2 FTS (C10) | 8 | 0% | <LOQ | <LOQ | - | 33% | 1.13* ± 0.01 | <LOQ - 1.21 | 0.6% | 0% | <LOQ | <LOQ | - |
| 5:3 FTCA (C8) | 5 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| 7:3 FTCA (C10) | 7 | 0% | <LOQ | <LOQ | - | 44% | 5.95* ± 0.1 | <LOQ - 6.36 | 3.2% | 0% | <LOQ | <LOQ | - |
| PFOSA (C8) | 8 | 0% | <LOQ | <LOQ | - | 44% | 0.28* ± 0.03 | <LOQ - 0.45 | 0.2% | 0% | <LOQ | <LOQ | - |
| MeFOSAA (C11) | 8 | 46% | 0.35* ± 0.1 | <LOQ - 1.14 | 0.1% | 33% | 0.28* ± 0.00 | <LOQ - 0.34 | 0.2% | 0% | <LOQ | <LOQ | - |
| EtFOSAA (C12) | 8 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| N-MeFOSE (C11) | 8 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| N-EtFOSE (C12) | 8 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| - | 514 ± 33.8 | 353 - 641 | - | - | 186 ± 8.6 | 130 - 215 | - | - | 155 ± 14.8 | 125 - 170 | - | ||
| - | 127 ± 10.8 | 73.0 - 172 | - | - | 57.0 ± 2.9 | 41.6 - 68.2 | - | - | 40.8 ± 6.9 | 27.3 - 49.8 | - | ||
| - | 385 ± 25.2 | 268 - 492 | - | - | 126 ± 6.1 | 88.9 - 150 | - | - | 99.0 ± 7.4 | 85.4 - 111 | - | ||
| - | 2.54 ± 0.5 | 0 - 6.12 | - | - | 3.52 ± 1.1 | 0 - 7.65 | - | - | 15.4 ± 2.2 | 12.8 - 19.8 | - | ||
# C+F = number of carbons with fluorine bonds; % >LOQ = percentage of samples exceeding limit of quantitation (LOQ); mean ± standard error of the mean (SEM) (ng/mL); ACT = average contribution to total detected PFAS.
| # C+F | North ( |
Central ( |
South ( |
||||||||||
| % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | % >LOQ | mean ± SEM (ng/mL) | Range (ng/mL) | ACT (%) | ||
| PFHxS (C6) | 6 | 100% | 1.33 ± 0.3 | 0.65 - 3.45 | 0.2% | 89% | 1.04* ± 0.2 | <LOQ - 1.65 | 0.4% | 90% | 0.78* ± 0.1 | <LOQ - 1.03 | 0.7% |
| PFHpS (C7) | 7 | 100% | 1.11 ± 0.2 | 0.53 - 2.39 | 0.2% | 100% | 0.80 ± 0.1 | 0.33 - 1.33 | 0.3% | 90% | 0.40* ± 0.03 | <LOQ - 0.61 | 0.4% |
| PFOS (C8) | 8 | 100% | 349.8 ± 105.0 | 131.2 - 1097 | 62.0% | 100% | 119.5 ± 22.1 | 49.8 - 235.7 | 50.8% | 100% | 57.4 ± 3.4 | 41.9 - 76.2 | 53.2% |
| PFNS (C9) | 9 | 33% | 0.32* ± 0.1 | <LOQ - 0.85 | 0.1% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| PFDS (C10) | 10 | 100% | 3.93 ± 0.5 | 1.81 - 6.40 | 0.7% | 100% | 2.01 ± 0.3 | 0.96 - 3.99 | 0.9% | 100% | 1.44 ± 0.1 | 1.08 - 2.05 | 1.3% |
| PFHxA (C6) | 6 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| PFOA (C8) | 8 | 33% | 0.26* ± 0.1 | <LOQ - 0.63 | 0.0% | 78% | 0.76* ± 0.2 | <LOQ - 1.93 | 0.3% | 80% | 0.41* ± 0.1 | <LOQ - 1.12 | 0.4% |
| PFNA (C9) | 9 | 100% | 10.6 ± 2.5 | 2.32 - 24.0 | 1.9% | 100% | 20.9 ± 4.0 | 9.04 - 44.3 | 8.9% | 100% | 11.3 ± 0.9 | 8.5 - 16.9 | 10.5% |
| PFDA (C10) | 10 | 100% | 47.1 ± 8.4 | 16.1 - 88.8 | 8.4% | 100% | 28.1 ± 4.8 | 11.9 - 44.0 | 11.9% | 100% | 12.2 ± 0.6 | 9.52 - 16.5 | 11.3% |
| PFUnA (C11) | 11 | 100% | 69.4 ± 5.3 | 42.8 - 89.6 | 12.3% | 100% | 37.0 ± 6.8 | 15.5 - 66.9 | 15.7% | 100% | 16.4 ± 0.9 | 12.8 - 23.0 | 15.2% |
| PFDoA (C12) | 12 | 100% | 34.9 ± 6.0 | 14.8 - 75.2 | 6.2% | 100% | 11.9 ± 3.9 | 3.44 - 38.6 | 5.1% | 100% | 3.68 ± 0.2 | 2.93 - 5.29 | 3.4% |
| PFTrDA (C13) | 13 | 100% | 27.0 ± 3.5 | 14.0 - 41.2 | 4.8% | 100% | 9.13 ± 3.1 | 2.74 - 31.3 | 3.9% | 100% | 3.12 ± 0.2 | 2.33 - 4.17 | 2.9% |
| PFTeDA (C14) | 14 | 100% | 9.17 ± 1.8 | 3.76 - 20.7 | 1.6% | 100% | 2.55 ± 1.0 | 0.76 - 9.98 | 1.1% | 90% | 0.60* ± 0.04 | <LOQ - 0.85 | 0.6% |
| 6:2 FTS (C8) | 6 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| 8:2 FTS (C10) | 8 | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| 5:3 FTCA (C8) | 5 | 33% | 6.68* ± 0.5 | <LOQ - 10.0 | 1.2% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| 7:3 FTCA (C10) | 7 | 44% | 2.95* ± 2.2 | <LOQ - 18.8 | 0.5% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| PFOSA (C8) | 8 | 1 | 1.79 ± 0.5 | 0.38 - 3.95 | 0.3% | 89% | 0.90* ± 0.4 | <LOQ - 3.67 | 0.4% | 80% | 0.32* ± 0.01 | <LOQ - 0.39 | 0.3% |
| MeFOSAA (C11) | 8 | 67% | 0.33* ± 0.03 | <LOQ - 0.59 | 0.1% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| EtFOSAA (C12) | 8 | 56% | 0.18* ± 0.03 | <LOQ - 0.37 | 0.0% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| N-MeFOSE (C11) | 8 | 44% | 0.05* ± 0.0 | <LOQ - 3.75 | 0.0% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| N-EtFOSE (C12) | 8 | 44% | 0.05* ± 0.01 | <LOQ - 3.75 | 0.0% | 0% | <LOQ | <LOQ | - | 0% | <LOQ | <LOQ | - |
| - | 563.6 ± 115.5 | 272.4 - 1332 | - | - | 235.4 ± 42.2 | 100.8 - 446.0 | - | - | 107.8 ± 5.2 | 88.9 - 138.7 | - | ||
| - | 198.6 ± 18.7 | 127.9 - 308.1 | - | - | 110.5 ± 19.6 | 48.8 - 200.0 | - | - | 47.6 ± 2.8 | 36.5 - 67.0 | - | ||
| - | 356.2 ± 106.0 | 134.4 - 1109 | - | - | 123.3 ± 22.7 | 51.8 - 242.3 | - | - | 59.9 ± 3.4 | 44.4 - 78.3 | - | ||
| - | 8.67 ± 3.7 | 0.86 - 29.9 | - | - | 1.59 ± 0.9 | 0 - 8.36 | - | - | 0.27 ± 0.1 | 0 - 0.4 | - | ||
| Delaware | Chesapeake | |||||||||
| Overall | North-Central | South - Central | South - North | Overall | PBR -APR | Poplar - APR | Poplar - PBR | |||
| 0.2600 | 2.692 | NA | NA | NA | 10.33 | 0.1747 | 0.1747 | |||
| 13.37 | 0.2323 | 14.95 | 0.0673 | 1.0000 | ||||||
| 21.96 | 22.76 | 0.0579 | ||||||||
| 21.64 | 0.2141 | 19.17 | ||||||||
| 10.42 | 0.0730 | <LOQ | <LOQ | <LOQ | <LOQ | |||||
| 8.129 | 0.5680 | 11.84 | 0.1832 | |||||||
| 19.11 | 0.0987 | 22.62 | 0.1045 | |||||||
| 27.79 | 19.57 | |||||||||
| 27.58 | 19.95 | |||||||||
| 29.37 | 5.716 | 0.9700 | ||||||||
| 29.99 | 7.636 | 0.1530 | ||||||||
| 15.5 | 0.0794 | 0.2530 | <LOQ | <LOQ | <LOQ | <LOQ | ||||
| 26.17 | 22.62 | 0.1071 | ||||||||
| 25.28 | 22.62 | 0.1071 | ||||||||
| 21.79 | 22.91 | |||||||||
| 22.61 | 0.2308 | 30.02 | 0.8776 | |||||||
ALB = albumin, AST = aspartate aminotransferase, BA = bile acids, Ca++ = calcium, CK =creatine kinase, GLOB = globulin, GLU = glucose, K+ = potassium in nmol/L, Na+ = sodium, PHOS = phosphorus, TP = total protein, and UA = uric acid. Hapt = haptoglobin-like activity, IgY = immunoglobulin gamma, DNA dam = DNA damage, T3 = triiodothyronine, T4 = thyroxine.
| Analyte | Units | Chesapeake | Delaware | ||||||||||||||
| Overall | APR | PBR | Poplar | Overall | North | Central | South | ||||||||||
| 19.0 ± 0.5 | 22 | 18.4 ± 0.7 | 10 | 19.7 ± 0.8 | 9 | 19.0 ± 2.0 | 3 | 30.2 ± 1.5 | 28 | 30.6 ± 3.1 | 9 | 29.9 ± 2.9 | 9 | 30.1 ± 2.0 | 10 | ||
| 1125 ± 72.8 | 22 | 1032 ± 115.6 | 10 | 1217 ± 110.3 | 9 | 1157 ± 233.9 | 3 | 2397 ± 108.3 | 28 | 2545± 209.8 | 9 | 2417 ± 244.2 | 9 | 2247 ± 104.2 | 10 | ||
| 11.8 ± 1.2 | 22 | 13.4 ± 1.8 | 10 | 10.90 ± 2.0 | 9 | 9.03 ± 1.1 | 3 | 13.6 ± 1.0 | 28 | 13.2 ± 1.3 | 9 | 15.9 ± 2.1 | 9 | 13.0 ± 1.9 | 10 | ||
| 228.3 ± 8.9 | 22 | 214.9 ± 18.9 | 10 | 236.0 ± 7.3 | 9 | 249.6 ± 9.8 | 3 | 258.0 ± 2.6 | 28 | 256.6 ± 4.3 | 9 | 260.3 ± 5.0 | 9 | 257.2 ± 4.5 | 10 | ||
| 9.20 ± 0.2 | 22 | 8.88 ± 0.3 | 10 | 9.53 ± 0.3 | 9 | 9.23 ± 0.4 | 3 | 10.5 ± 0.1 | 28 | 10.5 ± 0.2 | 9 | 10.5 ± 0.1 | 9 | 10.6 ± 0.1 | 10 | ||
| 9.94 ± 0.2 | 22 | 10.2 ± 0.4 | 10 | 9.72 ± 0.2 | 9 | 9.77 ± 0.4 | 3 | 6.76 ± 0.1 | 28 | 6.5 ± 0.2 | 9 | 6.63 ± 0.3 | 9 | 7.10 ± 0.2 | 10 | ||
| 2.70 ± 0.04 | 22 | 2.61 ± 0.1 | 10 | 2.8 ± 0.1 | 9 | 2.73 ± 0.1 | 3 | 2.87 ± 0.04 | 28 | 2.86 ± 0.1 | 9 | 2.86 ± 0.1 | 9 | 2.90 ± 0.1 | 10 | ||
| 1.67 ± 0.03 | 22 | 1.62 ± 0.1 | 10 | 1.73 ± 0.1 | 9 | 1.63 ± 0.1 | 3 | 1.76 ± 0.03 | 28 | 1.74 ± 0.1 | 9 | 1.74 ± 0.1 | 9 | 1.79 ± 0.04 | 10 | ||
| 1.04 ± 0.03 | 22 | 0.99 ± 0.04 | 10 | 1.07 ± 0.1 | 9 | 1.10 ± 0.01 | 3 | 1.10 ± 0.02 | 28 | 1.08 ± 0.03 | 9 | 1.1 ± 0.1 | 9 | 1.13 ± 0.04 | 10 | ||
| 5.28 ± 0.1 | 22 | 5.26 ± 0.1 | 10 | 5.14 ± 0.2 | 9 | 5.77 ± 0.23 | 3 | 4.79 ± 0.1 | 28 | 4.74 ± 0.1 | 9 | 4.7 ± 0.12 | 9 | 4.89 ± 0.1 | 10 | ||
| 145.0 ± 0.6 | 22 | 145.4 ± 0.42 | 10 | 145.7 ± 0.8 | 9 | 141.7 ± 3.0 | 3 | 145.1 ± 0.3 | 28 | 144.9 ± 0.6 | 9 | 145.1 ± 0.5 | 9 | 145.4 ± 0.4 | 10 | ||
| 0.24 ± 0.01 | 23 | 0.25 ± 0.03 | 11 | 0.23 ± 0.01 | 9 | 0.24 ± 0.01 | 3 | 0.22 ± 0.01 | 28 | 0.23 ± 0.01 | 9 | 0.21 ± 0.01 | 9 | 0.20 ± 0.01 | 10 | ||
| 1171 ± 85.6 | 23 | 1163 ± 126.1 | 11 | 1174 ± 115.2 | 9 | 1269 ± 426.5 | 3 | 1435 ± 90.8 | 28 | 1346 ± 128.3 | 9 | 1539 ± 181.5 | 9 | 1421 ± 166.6 | 10 | ||
| 34.6 ± 1.6 | 23 | 34.7 ± 2.01 | 11 | 37.4 ± 2.9 | 9 | 26.1 ± 2.1 | 3 | 60.7 ± 3.2 | 28 | 61.3 ± 7.3 | 9 | 56.3 ± 3.8 | 9 | 64.4 ± 5.3 | 10 | ||
| 1.13 ± 0.1 | 22 | 1.22 ± 0.1 | 11 | 1.08 ± 0.1 | 9 | 0.66 ± 0.01 | 2 | 1.12 ± 0.08 | 28 | 0.88 ± 0.05 | 9 | 1.03 ± 0.1 | 9 | 1.42 ± 0.2 | 10 | ||
| 11.9 ± 0.8 | 22 | 11.7 ± 1.03 | 11 | 11.1 ± 1.1 | 9 | 12.3 ± 0.7 | 2 | 15.4 ± 1.1 | 28 | 13.9 ± 1.5 | 9 | 16.3 ± 2.1 | 9 | 15.7 ± 2.3 | 10 | ||