Variation in postsampling treatment of avian blood affects ecophysiological interpretations


Correspondence author. E-mail:


1. The fluid component of blood is widely used in ecophysiological investigations, including measures of immune function and stable isotope ecology. After blood collection, delayed separation of blood extracellular fluids from red blood cells is known to affect the concentration of a wide range of biochemical compounds in the resulting fluid, as does prevention of clotting (producing plasma) when compared with blood allowed to clot (producing serum).

2. One challenge when investigating immune function and stable isotope ecology, therefore, is discriminating variation because of the effect of the biological factors of interest from potential methodological artefacts. This study assesses how seven widely used measures of immune function and stable isotope composition respond both to delayed separation of the cellular and fluid components and to the clotting of blood samples from two species of waterfowl.

3. Samples that remained uncentrifuged for up to 12 h did not differ from those centrifuged within 15 min of sampling from the same individuals, indicating that samples from a wide range of field conditions may remain highly comparable. However, the outcome of three of the four immunological assays and two of the three isotopic analyses was highly dependent on the type of fluid, with higher immunological activity and higher relative concentrations of heavy carbon and total nitrogen in plasma compared to serum.

4. Researchers interested in immune function and stable isotope ecology may obtain the most useful results by ensuring that they use a single fluid type in their investigations.


Blood is widely used to explore the physiology and stable isotope ecology of wild animals. Ecological immunologists utilize blood samples to investigate how immune function changes seasonally, under different environmental conditions, and between individuals, particularly in the context of sexual selection (Hasselquist 2007). Disease ecologists and veterinarians also rely on blood samples to measure specific antibodies that indicate prior infection with particular pathogens (e.g. Hoye et al. 2011). Furthermore, the concentrations of various stable isotopes in blood samples can also be used to reconstruct the composition of an individuals’ diet and determine its trophic position (Inger & Bearhop 2008), as well as assess reproductive investment strategies (Klaassen et al. 2001), patterns of habitat occupancy (Inger et al. 2008) and migration routes and timing (Hobson & Norris 2008). Each of these immunological and stable isotope applications often requires blood cells to be separated from the fluid component, plasma or serum, by centrifugation prior to analysis.

Prolonged storage of uncentrifuged human, reptilian and avian blood samples has been shown to result in significant changes to a wide range of biochemical assays owing to glucose depletion, the movement of water into cells and leakage of intracellular constituents (Boyanton & Blick 2002; Harr, Raskin, & Heard 2005). Immediate separation of plasma or serum from cells has therefore been advocated to optimize analytical results (Boyanton & Blick 2002). Yet, field conditions often result in a delay between blood collection and sample separation and storage, particularly when animals are captured in remote locations or in large numbers. Blood fluid biochemistry has also been shown to differ between samples in which clotting is prevented (producing plasma) and those in which clotting is allowed (producing serum) (Ceron et al. 2004; Mohri, Narenji Sani, & Masoodi 2008). Furthermore, collection-to-centrifugation delay may differentially affect analytical results depending on whether the blood is allowed to clot or not (Boyanton & Blick 2002). One challenge in investigating immune function and stable isotope ecology through blood sampling is, therefore, to recognize the extent to which different methodological procedures in processing wildlife blood after collection affect the biological factors of interest.

Given that wildlife studies are inconsistent in their use of plasma or serum, and in the time between collecting blood and separating the cells from the fluid fraction, this study investigates the potential for variation in blood processing to influence seven widely used measures of immune function and stable isotope composition. Specifically, this study addresses whether: (i) the length of time between taking a blood sample and separating the cellular and fluid components affects the results of immunological and stable isotope assays; (ii) the type of blood fluid (plasma or serum) affects these results; and (iii) identified methodological effects are consistent between species. Two waterfowl species – Bewick’s Swans (Cygnus columbianus bewickii, Yarrell) and Mallards (Anas platyrhynchos, Linnaeus) – are used because blood-based measurements of immune function, disease exposure, diet choice and foraging habitat have been investigated in a range of waterfowl species and because relatively large birds are required to furnish the volumes of blood needed for a repeated measures design.

Materials and methods

Study Animals

Bewick Swans and Mallards were held in outdoor aviaries and fed a diet containing 50% grain mixture and 50% chicken mash (HAVENS Voeders, Maashees, The Netherlands) ad libitum. The experimental procedures were carried out under approval CL09·04 from the Animal Experimentation Committee (DEC) of the Royal Netherlands Academy of Arts and Sciences (KNAW), and all efforts were made to minimize animal suffering throughout the experiment.

Blood Handling Experiment

Approximately 3 mL of whole blood was collected using a 23G needle and 5 mL syringe from the tarsal vein of nine individuals of each species (five men and four women) under brief (<1 min) anaesthesia with methoxyflurane. This represents ∼1% of the circulating blood volume of a 6000 g swan and ∼5% of the circulating blood volume of a 1000 g Mallard. Immediately after drawing the blood, each sample was divided equally into eight experimental vials – five containing a commercially prepared clot activator (no. 450470, Greiner Bio-one, Kremsmünster, Austria) to produce serum and three containing a commercially prepared anti-coagulant (Lithium Heparin, no. 450477, Greiner Bio-one, Kremsmünster, Austria) to obtain plasma. One plasma sample and one serum sample were centrifuged (7000g; 10 min) within 15 min of blood collection from each bird, with the remaining samples stored at 4°C. Pairs of plasma and serum samples were then centrifuged at 2 h and 4 h postsampling, and the remaining serum samples were centrifuged at 6 h and 12 h postsampling. All plasma and serum samples were stored at −20°C until analysis.

Immunological Assays

Matson, Ricklefs, & Klasing’s (2005) haemolysis/haemagglutination assay was used to measure natural antibody-mediated agglutination and lysis of exogenous erythrocytes. Briefly, 25 μL of plasma or serum was pipetted into the first and second rows of a 96-well plate and then twofold serially diluted with sterilized PBS from the second to the eleventh rows before adding 25 μL of 1% of rabbit red blood cells in PBS to each well. Plates were incubated for 90 min in a waterbath at 37°C, then tilted at a 45° angle for 20 min before scanning for agglutination and, after a further 90 min, for lysis. All samples were run in duplicate. Lysis and agglutination were quantified as the negative log2 of the last dilution exhibiting the respective activity.

Concentration of haptoglobin (mg mL−1), a measure of the strength of the acute-phase response (Matson 2006), was measured using the ‘manual method’ of a commercial haem-binding kit (no. TP801, Tri-delta diagnostics, Morris Plains, NJ, USA), requiring 7·5 μL serum or plasma. Absorbance was measured at 630 nm on a Synergy HT microplate reader (Bio Tek, Winooski, VT, USA). All samples were run in duplicate with duplicate calibration standards on each plate.

The presence of antibodies to the nucleoprotein gene segment of influenza A viruses was tested using a commercially available blocking enzyme-linked immunosorbent assay (b-ELISA; MultiS-Screen Avian Influenza Virus Antibody Test Kit, IDEXX Laboratories, Hoofddorp, The Netherlands) following manufacturer’s instructions (Brown et al. 2009). Assays were carried out in duplicate, each using 10-L samples of plasma or serum, in combination with supplied positive and negative controls. Absorbance was measured at 620 nm using a Tecan infinite 200 plate reader (Tecan Benelux BVBA, Giessen, The Netherlands). Sample signal/noise ratios (the quotient of sample mean absorbance divided by negative control mean absorbance) >0·5 were considered negative for the presence of antibodies to avian influenza virus.

Stable Isotope Analysis

A 5 μL aliquot of each sample was freeze-dried overnight in preweighed tin cups. These subsamples (approximately 200 – 500 μg) were analysed in a Euro EA 3000 elemental analyser (Eurovector, Milan, Italy) coupled through a Finnigan con-flo III interface to a Finnigan Delta V Advantage isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany). Stable isotope ratios are reported using the typical delta notation, in parts per mil (‰) such that δX = [(Rsample/Rstandard) −1]×1000, where X is 13C or 15N, R is the corresponding ratio of 13C/12C or 15N/14N, and Rstandard is the ratio of international reference standards: Vienna PeeDee limestone (PDB) for carbon and atmospheric nitrogen. Reproducibility based on replicate measurements of a casein standard (USG 40; mean δ15N = −4·57, mean δ13C = −26·28; n = 145) during the period of measurements was 0·12‰ (SD) for N and 0·36‰ (SD) for C. This analysis method also quantifies the amount (μg) of carbon and nitrogen in each sample, which were then converted to C/N molar ratios.

Statistical Analysis

The effect of sex, species, blood fluid type and time between bleeding and centrifugation on the outcome of each assay was tested using generalized estimating equations (GEEs). Raw values were used for haemolysis, δ13C, δ15N and C/N ratio, whereas transformations of measurements for haemagglutination (^1/3), haptoglobin (log10) and avian influenza b-ELISA (exp) were required to achieve a normal distribution. GEEs that assumed either a first-order autoregressive relationship within the repeated measures subjects (each individual) or an unstructured correlation matrix were tested for each dependent variable (assay). In all cases, the first-order autoregressive relationship produced the best fit on the basis of QICC values (Corrected Quasi Likelihood under Independence model Criterion; an adaptation of Akaike’s Information Criterion for repeated measures). Full factorial models were tested for each dependent variable, and interaction levels sequentially removed when not significant at the 0·05 level. Sex was not significant to the outcome of any of the seven assays and was therefore removed from the models presented. Values represent mean ± SEM. throughout the text and figures. All statistical tests were conducted in spss version 17·0.


Immunological Assays

There was no effect of sample-to-centrifuge time on any of the immunological assays (Fig. 1, Table 1). There was, however, significant variation between species in natural antibody-mediated agglutination (Table 1), with Bewick’s swans showing higher activity than Mallards in both assays (Fig. 1). Immune function also varied between plasma and serum, with significantly higher values obtained from plasma for haemagglutination activity (7·12 ± 0·16 and 6·01 ± 0·13, respectively; Fig. 1; Table 1), and haptoglobin concentration (0·13 ± 0·006 and 0·04 ± 0·005, respectively; Fig. 1; Table 1). Furthermore, each bird’s plasma showed significantly lower b-ELISA s/n ratios than its serum (Table 1). The differences between plasma and serum were consistent between the two species. Haemolysis did not differ between the two fluids in either species (Table 1).

Figure 1.

 Immunological measures, including haemolytic activity (panels a and b), haemagglutinating activity (panels c and d), haptoglobin concentration (panels e and f) and signal/noise ratio of a blocking enzyme-linked immunosorbent assay (b-ELISA) to detect antibodies to influenza A viruses (panels g and h), in blood fluid stored in vials containing a clot activator (serum, filled circles) or an anti-coagulant (plasma, unfilled circles) as a function of time between drawing the sample and centrifugation. Whole blood samples were from Bewick’s swans (panels a, c, e, g), and Mallards (panels b, d, f, h) were stored as eight equal samples per individual. Values represent means ± SEM.

Table 1.   Effect of species, fluid type and time between bleeding and centrifugation on immunological and stable isotope analysis of blood samples from captive Bewick’s swans (Cygnus columbianus bewickii; n = 9) and Mallards (Anas platyrhynchos; n = 9).
EffectHemolysisHemagglutination*Haptoglobin†AIV b-ELISA‡δ13Cδ15NC/N ratio
  1. Subsamples were stored with an anti-coagulant (plasma) or a clot activator (serum) for 0–12 h before centrifuging to separate this fluid fraction of the blood. The models presented assume a first-order autoregressive relationship within the repeated measures subject (each individual bird). Significant results are indicated in bold type.

  2. AIV, Avian influenza virus; ELISA, Enzyme-linked immunosorbent assay; QICC, Corrected Quasi Likelihood under Independence model Criterion; an adaptation of Akaike’s Information Criterion for repeated measures.

  3. *Transformed to the power 1/3.

  4. †Log10 transformed.

  5. ‡Exponential transformed.

Species • fluid      0·028
Species • time      0·161
Time • fluid      0·026

Stable Isotope Analysis

Time between drawing a blood sample and separating the fluid and cellular components did not have a significant effect on stable isotope or elemental composition (Fig. 2; Table 1); however, it approached statistical significance for δ13C (= 0·056) and showed a significant interaction with fluid type for C/N ratio (= 0·026). There were also significant differences between species in δ15N (Bewick’s swans: 5·96 ± 0·03 compared with Mallards: 6·57 ± 0·04; Fig. 2; Table 1) and C/N ratios (Bewick’s swans: 5·07 ± 0·02 and Mallards: 5·22 ± 0·02; Fig. 2; Table 1) despite both species being fed on an isotopically identical diet. Both δ13C and C/N ratios showed significant differences between serum and plasma within individual birds, with serum being more depleted in heavy carbon (δ13C = −21·18 ± 0·05 vs. −21·57 ± 0·04; Fig. 2; Table 1) and total nitrogen (C/N ratio = 5·09 ± 0·02 vs. 5·18 ± 0·02; Fig. 2; Table 1) compared with plasma. There were, however, no significant differences in δ15N between serum and plasma in both species (Table 1).

Figure 2.

 Stoichiometric measures, including carbon stable isotope ratio (δ13C; panels a and b), nitrogen stable isotope ratio (δ15N; panels c and d) and carbon/nitrogen molar ratio (panels e and f), in blood fluid stored in vials containing a clot activator (serum, filled circles) or an anti-coagulant (plasma, unfilled circles) as a function of time between drawing the sample and centrifugation. Whole blood samples from Bewick’s swans (panels a, c, e, g) and Mallards (panels b, d, f, h) were stored as eight equal samples per individual. Values represent means ± SEM.


Immunological Assays

Storing blood at 4°C for up to 12 h prior to separation of the cellular and fluid components did not have a significant effect on any of the immunological assays. However, there was substantial variation between the values obtained from plasma and serum in three of the four immunological assays. Given that serum is blood plasma without fibrinogen and other clotting proteins, it is not surprising that the biochemical and functional properties of serum may differ from those of plasma. Indeed, in this study, serum showed significantly lower agglutinating (Fig. 1c and d) and specific antibody-binding activity (Fig. 1g and h), as well as lower concentrations of acute-phase proteins (Fig. 1e and f) compared with plasma for both species. These results suggest that, in addition to the proteins, antibodies and hormones found in serum, the proteins involved in coagulation may also play a role in certain in-vitro immune reactions. The difference in haptoglobin concentration between plasma and serum (Fig. 1e and f) was of a similar magnitude to the differences between plasma samples from continental and insular birds, including waterfowl, reported by Matson (2006). Similarly, differences in natural antibody-mediated agglutination between plasma and serum (Fig. 1c and d) were slightly larger than those found between habitats and age groups in Red knots (Calidris canutus) (Buehler, Tieleman, & Piersma 2009) (although it is unclear whether these samples were allowed to clot prior to separation and whether the degree of coagulation may have differed between individuals). Furthermore, signal/noise ratio (s/n) of the Avian influenza virus (AIV) b-ELISA also showed lower reactivity (higher s/n) in serum compared with plasma. Such differences were of negligible consequence when s/n was relatively low (<0·3) or relatively high (>0·7). However, two swans and one mallard had average s/n close to the 0·5 threshold for assigning seropositive or negative status. In these individuals, the difference in s/n between fluid types resulted in samples stored as serum being more likely to be classified seronegative than samples stored as plasma, indicating that different thresholds may be required for different fluid types. The use of both serum and plasma in a single study may, therefore, be sufficient to obscure significant ecological relationships.

Stable Isotope Analysis

Similar to the immunological assays, prolonged delays in sample processing did not have a significant effect on stable isotope composition, although delays beyond 12 h may go on to affect δ13C signatures and C/N ratios. Yet, δ13C and C/N ratios showed significant differences between serum and plasma within individual birds. Kurle (2002) found strikingly comparable results in fur seals, where plasma δ13C was 0·4‰ higher than serum δ13C from the same individuals. Together, these results indicate that the fibrinogen and other clotting proteins contained in plasma may discriminate 13C to a greater extent, resulting in a fluid that was 0·28‰ (±0·05) higher δ13C than the δ13C of the diet, compared with serum, which showed δ13C values 0·11‰ (±0·04) lower than the diet. This is particularly relevant for studies using stable isotopes to determine the timing of diet switches as small errors in the estimation of stable isotope discrimination between diet and tissue can result in large errors in foraging models (Caut, Angulo, & Courchamp 2009). Discrimination factors applied to δ13C signatures of blood components should therefore be specific to the fluid used. Unfortunately, information on the postsampling treatment of blood is rarely included in studies using stable isotope ecology. Discrimination factor values are available for plasma (e.g. Evans Ogden, Hobson, & Lank 2004; Klaassen et al. 2010) yet, without the use of anti-coagulants, even brief (<15 min) delays prior to separation will result in some coagulation and hence a fluid that no longer contains the full complement of proteins (and associated δ13C signature). A further complication for field studies may be that the coagulating proteins, and hence plasma, turn over at a slightly different rate to serum, although little work has been performed on this topic. Because δ15N showed little variation between plasma and serum, these results suggest that the type of fluid used is less likely to affect investigations of trophic level (based on δ15N) than investigations of foraging habitat, diet composition and time since diet switch (based on δ13C). The presence of coagulating proteins in plasma and not serum can also explain the higher C/N ratio (i.e. lower protein content) of serum compared with plasma in this study.

Practical Implications

Given that blood may be stored unseparated for up to 12 h at 4°C without altering subsequent measures of immune function or stable isotope composition in the fluid component, these results indicate that samples collected under a wide range of field conditions may remain highly comparable. However, there appear to be critical differences in a number of assay results between plasma and serum. Researchers interested in comparisons of immune function and stable isotope composition should therefore confine measurements to a single fluid type whenever possible. In this study, plasma and serum were achieved through addition of an anti-coagulant and a clot activator, respectively. However, given that these chemical additives may influence the chemical composition of the resulting fluid, some researchers may prefer to store blood samples in sterile vials without chemical additives. In either case, samples to be assayed as plasma critically require immediate separation or addition of anticoagulants to whole blood to maintain the full complement of proteins in the fluid. Alternatively, complete coagulation, achieved by the addition of either clotting agents or prolonged storage (12–24 h), is required to obtain serum. Given the physiological, biochemical and apparent functional differences between these two fluid types, samples allowed to partially coagulate will prove difficult to interpret. The choice of fluid type should therefore be dependent upon the practicality of obtaining a certain fluid under field conditions, as well as the performance of the fluid in the assays of interest. If such standardization is truly impractical, correction factors for differences between plasma and serum may potentially ameliorate the discord between the two fluid types. For instance, the results of this study suggest that serum was lower than plasma by 0·38‰ (±0·03) for δ13C; 0·085 mg mL−1 (±0·01) for haptoglobin; 1·28-fold (±0·16) haemagglutinating activity; and had C/N ratios that were 0·09 (±0·008) higher than plasma. These values should, however, be used with caution until a broader range of species has been tested.


I thank Bart van Lith for animal maintenance and assistance with blood sampling, Annelies Stuifzand for assistance with the immunological assays and Harry Korthals for stable isotope analysis. Bill Buttemer, Richard Inger and an anonymous reviewer provided valuable comments on an earlier version of this manuscript. This study was supported by the Research Council for Earth and Life Sciences (ALW) with financial aid from the Netherlands Organization for Scientific Research (NWO; grant 851·40·073). This is publication 5021 of the NIOO-KNAW.