Table 1.
Steps to improve bioanalytical precision when measuring human oxytocin concentrations.
| Issue | Concern | Proposed solution |
|---|---|---|
| Inconsistency in sample collection, processing, and storage. | Different types of samples require specific methods of collection, processing, and storage. |
Plasma: (1) Collect blood in chilled vacutainer tubes with EDTA and a protease inhibitor (since oxytocin may degrade quickly [3] especially in pregnancy when circulating levels of oxytocinase are increased [62, 146]), then place into an ice bath and refrigerate.a (2) Once samples are collected, standard protocols to separate EDTA plasma from whole blood in a refrigerated centrifuge can be used [22]. This is followed by aliquotting the cold plasma from vacutainers into chilled cryovials (i.e., securely placed on ice) and storing at −80 degrees C. As noted in the main text, serum can also be used and techniques are similar. Saliva: (1) Saliva samples should be handled in a similar manner except they can be obtained with standard saliva collection methods including Salivettes and frozen at −20 degrees C until assay [23]. Urine: (1) Current recommendations are for urinary samples to be stored at −80 degrees C, stabilized by reducing the pH with acid and extracted. Extracted samples can then be stored at −20 degrees C until assay [125]. |
| Inconsistency in use of extraction procedure. | Unable to combine and synthesize results in meta-analyses. There ends up being two different literatures on endogenous oxytocin concentrations, one where samples have been extracted and one where they have not. | The extent to which plasma proteins interfere with an immunoassay varies depending on the antibody used and may vary between individuals. Thus, the need and rationale for sample extraction to measure human plasma oxytocin is now well established [2, 35, 36]b. Therefore, until the molecular identity of “bound” oxytocin and/or high molecular weight oxytocin antibody immunoreactive molecules can be established, all plasma oxytocin measurements should use extracted samples. Data with unextracted samples can be generated in parallel, but referred to as unextracted sample oxytocin immunoreactivity to distinguish between the two methods. However, measuring changes in human salivary oxytocin concentrations using an RIA that includes an antibody that is not affected by proteins in the sample matrix does not require extraction [23]. In addition, measurement of CSF oxytocin may result in similar levels with or without extraction [38]; however, it also depends on the antibody used [55]. |
| Studies often refer to the presence of matrix effects without testing whether there are matrix effects in the samples used. | Assays run on unextracted samples may be impacted by matrix effects and therefore should be extracted. Researchers may not be extracting based on manufacturer’s information, but may not realize that recoveries, serial dilutions etc. have been done in assay buffer with purified standards, rather than in a biological matrix. | Run spiked and unspiked samples diluted in parallel and show co-linearity. Dilutions should be done in the matrix rather than assay buffer, the latter would also be diluting any interfering substances that might be present in the matrix (i.e., plasma, saliva, or urine). |
| Not all studies that concentrate and reconstitute samples report whether higher concentration of ions and/or sample pH may be impairing antibody binding and oxytocin quantitation. | When a sample is concentrated, everything else in the sample including salt ions (e.g., sodium, potassium, chloride) become concentrated. This can result in an increase in ionic strength or change in pH that can affect the way an antibody binds its target and non-specifically alter the result. Therefore, any change in ionic strength or pH must be evaluated and shown not to affect the assay. | Where samples have been concentrated and reconstituted, it is important to check that the higher concentration of ions and sample pH does not impair antibody binding. |
| Poor recovery of oxytocin following extraction. | Can impact the validity of results. | Experimenters should report the recovery of oxytocin extracted from a subset of spiked samples, and report whether values have been corrected for recovery after extraction. |
| Lack of sensitive RIA antibodies and/or sensitive EIA kits. | Large percentage of undetectable levels. | A collaborative effort is needed between labs to increase the availability and identification of sensitive and specific antibodies for oxytocin. This effort should include the development of a website that researchers can use to find updated information about which lab has what type of antibody for RIA, how much of it, and its level of sensitivity. To further increase reproducibility, labs could share data on methods and assay validation, particularly if these have not been published elsewhere. |
| Not all studies report the study-specific intra- and inter-assay coefficients of variation, or the exact values | If either the intra- or inter-assay coefficient of variation are too high, results could be invalid. Similarly, if relying on the manufacturer’s reported CVs, this may not be accurate in the given sample. | Report exact intra- and inter-assay coefficients of variation from the specific sample, not the manufacturer’s reported CVs. Intra-assay CV is best calculated as the differences between study sample duplicates. Inter-assay CV from common control samples included in successive assays. |
| Not all studies report whether samples were measured in duplicate. | Measurements from anything less than duplicate samples may be inaccurate. | Measure oxytocin in duplicate and re-run samples with poor CVs. As a general guideline, inter-assay %CV should be less than 15% while intra-assay %CV should be less than 10%. |
RIA radioimmunoassay, EIA enzyme immunoassay, CV coefficient of variation.
aIt is difficult to purchase vacutainer tubes with EDTA in the United States that also contain aprotinin, but this type of tube does exist outside of the United States. Thus, vacutainers may need to be opened after blood has been collected and aprotinin added manually. Evidence in non-human species suggests that plasma oxytocin levels measured via enzyme immunoassay can vary in unextracted samples if blood is collected in tubes with EDTA vs. heparin [92]. Further research is needed to systematically examine this issue in humans. While oxytocin is stable in plasma (other than in pregnancy) after separation from whole blood [35, 95], stability in whole blood has not been described, thus use of a protease inhibitor may prevent peptide degradation during collection and prior to processing. For aprotinin, 0.6 Trypsin Inhibitor Unit (TIU) should be used per 1 ml of blood. Use of other anticoagulants (e.g., heparin, citrate) for plasma preparation may be used but validation of oxytocin stability and recovery after down-stream processing may be needed.
bResults from a recent study suggested that the measurement of oxytocin in unextracted mouse plasma correlates well with levels in extracted samples, but it was dependent on the enzyme immunoassay kit and antibody [147]. Use of unextracted samples from other species including human plasma would require additional validation.