Salivettes are commonly used to collect human saliva for the estimation of saliva flow rate and concentrations of steroid hormones, total protein, secretory immunoglobulin A (s-IgA) and amylase. Salivettes include an absorbent cotton roll, a plastic roll container with a perforated bottom, and a centrifuge tube. Collection of saliva is usually accomplished by placing the cotton roll under the tongue for a timed period of 1-4 min. The cotton roll is then replaced in the container and back in the centrifuge tube. Centrifugation allows collection of the saliva in the bottom of the tube, which can then be stored frozen prior to analysis. However, the use of an absorbent cotton roll may affect saliva composition and the estimation of saliva flow rate.

In the present study, with local ethics committee approval, eight healthy men (age 29 ± 3 years, body mass 74.2 ± 1.3 kg, mean ± S.E.M.) provided ~15 ml of unstimulated saliva by dribbling into a tube over a 20-30 min period following an overnight fast. The following volumes: 0.2, 0.4, 0.7, 1.0, 2.0, 3.0 and 4.0 ml (with the remaining volume of about 3 ml as a control) were placed in pre-weighed vials. Pre-weighed cotton rolls (diameter 1 cm, length 4 cm) were put into each vial (except for the control sample) and placed on a shaker at 500 r.p.m. for 2 min. After shaking, the swabs were removed and centrifuged at 1500 g for 10 min at 18 °C. Saliva volume was estimated by weighing to the nearest mg and saliva density was assumed to be 1 g ml^-1. Samples were then stored frozen at -20 °C prior to analysis of total protein, s-IgA, amylase and cortisol by methods previously described (Walsh et al. 1999). The cotton rolls were also weighed following centrifugation, so that the amount of saliva retained in the cotton material could be determined. Results were analysed using ANOVA and paired t tests applied where appropriate. Statistical significance was accepted at P < 0.05.

The cotton roll became saturated when saliva volume exceeded 2 ml. The amount of saliva retained in the cotton roll after centrifugation was not constant and ranged from 0.09 ± 0.01 to 0.28 ± 0.03 ml, for saliva volumes of 0.2 and 4.0 ml, respectively. Thus a higher percentage of the initial saliva volume was retained by the cotton roll at the lower saliva volumes (45, 22, 21, 16, 12, 9 and 7 % for initial saliva volumes of 0.2, 0.4, 0.7, 1.0, 2.0, 3.0 and 4.0 ml, respectively). The saliva levels of total protein, s-IgA, amylase and cortisol (Table 1) were all significantly affected by the presence of the cotton roll, but were not significantly influenced by the volume of saliva. Total protein, s-IgA and amylase were 24, 16 and 12 % lower in the saliva exposed to the cotton roll, respectively, whereas cortisol was 33 % higher than in the control sample.

Our findings indicate that the cotton roll collection method affects the results of total protein, s-IgA, amylase and cortisol. Estimations of saliva flow rate will also be inaccurate when the saliva volume exceeds 2 ml and/or if the volume of saliva retained in the cotton roll after centrifugation is ignored. With regard to previously reported studies in which cotton rolls were used to collect saliva, our findings suggest that the results of such studies may be compromised and need to be reconsidered.