Saliva Sampling in Children and Young People: Acceptability and Reliability Data From Three Exploratory Studies

Abstract

In recent years, improvements in laboratory techniques have aroused new interest in the use of saliva as an alternative medium in which to measure hormones. A number of sampling methods are available, including passive drool, suction, spitting and absorption [1]. The sampling method may be informed by the hormone being analysed, age and preference. Salivary cortisol and cortisone concentrations correlate strongly with those made in serum in children [2] and adults [3], and are stable across a range of temperatures [4]. Samples collected at home can be returned to the laboratory by post. Saliva sampling is minimally invasive, pain-free,enables multiple samples to be collected across the day at home, and is less costly than the collection of blood samples by a health care professional [3]. Late-night salivary cortisol is recognised as a useful tool in the investigation of suspected Cushings syndrome in children [5] and adults [6], and salivary androgen profiles show promise in the monitoring of children with congenital adrenal hyperplasia [7]. Saliva has been used as a medium in a large number of paediatric research studies. An international study, collecting a morning sample at home, and a sample before and after a blood test, concluded saliva sampling was feasible [8], and a study examining the utility of testing for COVID-19 in saliva from children, also concluded this was feasible and acceptable [9]. A recent review concluded that saliva collection was well tolerated by children and adolescents, although it should be noted that most studies included in this review collected only one sample per participant, and data reporting how many samples were of an adequate volume for analysis were not reported [10].

In order to learn more about the acceptability and adequacy of saliva sampling in a paediatric population, we reviewed data collected from three exploratory studies: (A) Turner Syndrome [11], (B) exaggerated adrenarche, and (C) healthy volunteers [2]. Participants collected the first saliva sample using Salivette devices during a study visit in studies A and C. Further samples were collected at home every 2 h during waking hours and analysed for cortisol and cortisone. Participants in study B collected a single saliva sample by passive drool at home, 30 min after waking, which was analysed for androgens. Participants and parents/carers in studies A and B also completed questionnaires to assess the acceptability of this sampling method. Study C was performed before study A or B. Our decision to collect structured acceptability data in studies A and B was informed by anecdotal feedback from participants and parent/carers in study C. Study C, in which most samples were collected at home, was included to increase the sample size for the analysis of the reliability of saliva collection.

Participants did not eat, drink or brush their teeth within the hour prior to sample collection. Responses to questionnaires were measured using 10-point Likert-type scales (10–most favourable). A score exceeding 5 out of 10, based on established thresholds from similar studies, indicated acceptability [12].

Ethics committee approvals were obtained for all studies (21/SC/0292, 23/PR/1398, 15/LO/1840 and 20/SW/0162).

In total, 114 participants (mean age 10.8 ± 2.5 years; 51.8% female) collected 689 samples in three studies. Participant characteristics are given in Table 1. Across the studies, 1.5% of returned samples were inadequate for biochemical analysis.

Questionnaire responses were collected from 27 parents/carers and 26 participants. Parents/carers rated the ease of saliva collection methods a mean score of 9.0 ± 1.9 in Study A and 6.1 ± 2.4 in Study B. Nine (90.0%) parent/carers in Study A and 16 (94.1%) in Study B found the sample collection kits acceptable to use. In Study A, nine (90.0%) parents/carers rated the frequency of sample collection during waking hours as acceptable, with a mean ease score of 8.4 ± 1.7. In Study B, parents/carers rated their comfort in collecting a saliva sample from their child with a mean score of 9.0 ± 1.6, and 16 (94.1%) parents/carers found this study activity acceptable. Participants rated their levels of comfort during sample collection as a mean score of 6.4 ± 2.0.

In Study B, when theoretically comparing saliva to blood and urine sampling, 15 (88.2%) parents/carers preferred saliva over blood and 13 (76.5%) preferred saliva over urine sampling. Saliva was preferred to blood or urine sampling by 15 (94%) participants.

We concluded that saliva sampling for cortisol, cortisone and androgen profiles in children and young people is acceptable and reliable. The majority of returned samples were adequate for biochemical analysis, supporting data from previous studies, indicating that saliva can be reliably collected at home [3]. Additionally, most parent/carers found it acceptable to collect samples from their child every 2 waking hours, allowing for the creation of a hormone day curve profile. However, fewer were able to collect a sample between 23:00 and midnight, to identify the cortisol/cortisone nadir. It may be that children being investigated in the clinical setting for signs of cortisol excess would be more likely to collect this last night sample than our study participants.

Overall, participants perceived saliva sampling as an acceptable clinical investigation, with most feeling comfortable during sample collection. This contrasts with reports of pain and anxiety associated with venepuncture, making it predictable that most participants preferred saliva over traditional methods of sampling [13]. From our data, there is a suggestion that sample collection by Salivette may be easier than passive drool; however, the differences in study cohorts should be considered when interpreting these findings. Children collecting saliva samples by passive drool were younger, and it may be that they found the oral coordination necessary to collect a sample by this method difficult. This may account in part, for the higher number of inadequate samples in this cohort.

There are limitations to our findings. Although data from a large number of participants and samples are reported, these results may not be applicable to all patients, for example, those with poor oral motor coordination or special educational needs. More data are required to confirm whether passive drool is less acceptable and reliable than collection by Salivette. Additionally, neither protocol directly compared saliva to blood and urine sampling.

It is also important to consider the barriers to implementing saliva sampling in clinical practice. These include establishing reference ranges, managing logistical challenges of home sample collection and transport, and addressing the longer processing times compared to traditional methods. Further efforts are required to overcome these challenges before saliva sampling can be widely adopted as a routine investigation in clinical practice.

In conclusion, these data demonstrate high acceptability and reliability of saliva sampling as an endocrinology investigation, providing promising evidence for its utility as a non-invasive, convenient, and patient-centred method, which may enhance adherence to diagnostic and monitoring protocols, particularly within paediatrics.

The authors declare no conflicts of interest.