Elsevier

Psychoneuroendocrinology

Volume 29, Issue 10, November 2004, Pages 1229-1240
Psychoneuroendocrinology

The “trouble” with salivary testosterone

https://doi.org/10.1016/j.psyneuen.2004.02.005Get rights and content

Abstract

In a series of studies, we identify several specific issues that can limit the value of integrating salivary testosterone in biosocial research. Salivary testosterone measurements can be substantially influenced during the process of sample collection, are susceptible to interference effects caused by the leakage of blood (plasma) into saliva, and are sensitive to storage conditions when samples have been archived. There are gender differences in salivary testosterone levels and variance, the serum–saliva association, the relationship of salivary testosterone to age and pubertal development, and the stability of individual differences in salivary testosterone levels over time. The findings have important implications at several levels of analysis for research that aims to test biosocial models of testosterone–behavior relationships. Recommendations are provided to steer investigators around these “troubles” with salivary testosterone.

Introduction

The ability to measure biological variables non-invasively in saliva has created many opportunities for behavioral scientists to test biosocial models of individual differences and intra-individual change in mood, cognition, behavior and psychopathology (Booth et al., 2000, Dabbs, 1993, Kirschbaum et al., 1992, Malamud and Tabak, 1993). As the number of studies doing so has increased, special circumstances capable of compromising the value of the information generated have been documented (e.g., Magnano et al., 1989, Schwartz et al., 1998). Over the past 10 years, our studies have documented that these issues are especially salient with respect to the measurement of testosterone in saliva. In this paper, we comprehensively review these problems and pitfalls to update this information (Dabbs, 1993) and steer the next generation of studies clear of these “troubles” with salivary testosterone.

There has been renewed interest among researchers in the relationship between testosterone and a variety of behavioral (e.g., Dabbs and Morris, 1990), physiological (e.g., Walker et al., 1980), social (e.g., Booth and Dabbs, 1993) and health constructs (e.g., Boas et al., 1996, Booth et al., 1999). Access to non-invasive means of measuring testosterone (in saliva) has in part rekindled this interest. Monitoring testosterone in saliva has several obvious advantages over doing so in serum or urine. Saliva collection procedures afford researchers the ability to conduct repeated sampling over the course of minutes, hours, days or longer (Malamud and Tabak, 1993). Research assistants, caregivers and subjects themselves with minimal training can easily collect saliva samples in most circumstances. Also the reliability, precision, accuracy and analytical recovery of immunoassays designed to measure salivary testosterone are now well documented (e.g., Granger et al., 1999). The ease of collection and recent commercial availability of salivary assay protocols have contributed to the impression that salivary testosterone is “ready” to be incorporated into biosocial research on a widespread basis (Dabbs, 1990). However, there are a number of studies reporting accounts of special circumstances with potential to significantly influence the external validity of salivary testosterone results (Dabbs, 1991, Granger et al., 1999). A more fine grained examination reveals wide gaps in information available to guide researchers in the utility of this measurement approach.

In this paper, we document that salivary testosterone measurements can be substantially influenced during the process of sample collection, are susceptible to interference effects caused by the leakage of blood (plasma) into saliva and are sensitive to storage conditions when samples have been archived. We describe gender differences in salivary testosterone levels and variance, the serum–saliva association, the relationship of salivary testosterone to age and pubertal development and the stability of individual differences in salivary testosterone levels over time. Implications for testing biosocial models of testosterone–behavior relationships are discussed and guidelines and recommendations are provided.

Section snippets

Saliva sample collection

A series of experiments have extensively evaluated the effects of different sample collection techniques on the measurement of a variety of salivary biomarkers (Dabbs, 1991, Granger et al., 1999, Schwartz et al., 1998, Shirtcliff et al., 2001a). Here, we review that materials commonly used in the literature to absorb sample (i.e., cotton and polyester swabs) or stimulate saliva flow (i.e., powdered drink-mix crystals and chew gum) have the potential to substantially change salivary testosterone

Blood contamination of saliva

Ten years ago, the literature warned that when blood is present in the oral mucosa, quantitative estimates of hormones measured in saliva may be compromised (Malamud and Tabak, 1993). Blood (and its components) can leak into the oral mucosa as a result of microinjuries such as burns or abrasions. To be a meaningful index of systemic bioactivity, quantitative estimates of testosterone in saliva must be highly correlated with the levels in serum. A high serum–saliva correlation depends on

Correspondence of testosterone levels measured in saliva and serum

As noted above, to meaningfully reflect systemic bioactivity, quantitative estimates of testosterone in saliva must be highly correlated with the levels in serum. In a recent paper (Shirtcliff et al., 2002), we documented a sex difference in the association between salivary and serum testosterone measures, and then determine if this sex difference would affect the detection of testosterone–behavior associations.

Participants in this analysis were 20 male and 20 female Penn State University

Sample storage and archiving

As the collection of saliva becomes more routine, samples will more commonly be archived to be assayed in the future. In this analyses, we evaluate week-to-week change in levels of testosterone in saliva samples when stored up to four weeks at 4 °C in a regular household refrigerator. Also, the long-term stability of testosterone levels in saliva was tested every 6 months when stored up to two years at −20 °C in a frost free (i.e., freeze–thaw cycling) regular household freezer, and also at −40

Modeling individual differences in salivary testosterone

Contemporary theorists speculate about reciprocal effects between hormones and behavior, and assume such effects are heavily dependent on intrinsic individual differences as well as characteristics of the social ecology (Booth et al., 2000, Brooks-Gunn et al., 1994, Gottlieb, 1992, Susman, 1997). To test these complex ideas requires a focus on identifying sources of individual differences, and employing methodologies that gather multiple waves of hormone and behavioral data over time.

General conclusion

Contemporary theorists consider testosterone–behavior relationships to be dynamic, having reciprocal, and potentially non-linear, effects that are highly dependent on intrinsic characteristics and features of the social environment (Booth et al., 2000, Brooks-Gunn et al., 1994, Campbell et al., 1997, Susman, 1997). In order to extend our understanding of these effects to new limits, investigators have begun to take advantage of saliva sampling to measure testosterone in ecologically valid

Acknowledgements

We acknowledge Robert Sapolsky’s creativity as the title of this paper is a play on the title of his book, “The Trouble with Testosterone”. Thanks are due to Ruth Merritt, Jodi Heaton, Jessica King, Jyotika Mirchandani, Andrew Kurscher, Andrew Hackett, Erica Schreffler, and Stefanie Bachman during data collection, and Laurie O’Brien, Carla Slike, Carol Boob, and Skip Nelson for technical assistance with the assays. The preliminary findings and comments shared with us by our colleagues Vangie

References (33)

  • A Booth et al.

    Biosocial perspectives on the family

    Journal of Marriage and the Family

    (2000)
  • A Booth et al.

    Testosterone and child and adolescent adjustment: the moderating role of parent–child relationships

    Developmental Psychology

    (2003)
  • J Brooks-Gunn et al.

    Studying links between hormones and negative affect: models and measures

    Journal of Research on Adolescence

    (1994)
  • A Campbell et al.

    Aggression and testosterone: testing a bio-social model

    Aggressive Behavior

    (1997)
  • J.M Dabbs

    Salivary testosterone measurements in behavioral studies

    Annals of the New York Academy of Sciences

    (1993)
  • J.M Dabbs et al.

    Testosterone and antisocial behavior in a sample of 4462 men

    Psychological Science

    (1990)
  • Cited by (327)

    • Not within spitting distance: Salivary immunoassays of estradiol have subpar validity for predicting cycle phase

      2023, Psychoneuroendocrinology
      Citation Excerpt :

      Since 2021, "hormonal assessment for confirmation of cycle phase" was made a precondition for publication at the journal Psychoneuroendocrinology (2022) and this condition has mostly been fulfilled via steroid assays in saliva, as opposed to serum or urine (see Supplementary Note 1). However, salivary immunoassays come with known issues (Granger et al., 2004; Schultheiss et al., 2018; Welker et al., 2016; Wood, 2009). Low concentrations of steroids, especially estradiol, are already very challenging to measure accurately in serum by immunoassay (Handelsman, 2017; Vesper et al., 2014a), mainly because of low specificity at lower concentrations (Garnett et al., 2020; Vesper et al., 2014b).

    View all citing articles on Scopus
    View full text