Short-term biological variation of serum tryptase

Introduction

Tryptase is a trypsin-like serin protease produced mainly by mast cells [1]. During normal homeostasis the serum tryptase level is low, but tryptase is released upon activation of the mast cells and the serum level increases substantially [2]. The function of tryptase is not completely elucidated, but it is believed to promote numerous proinflammatory functions [2]. Elevated serum tryptase concentrations have been observed in several pathological conditions including systemic allergic reactions, and serum tryptase has proven valuable as a biomarker in anaphylactic reactions [3] as well as in non-anaphylactic conditions [4]. The tryptase level is a minor diagnostic criterion for mastocytosis, and it is widely accepted that an increased serum baseline tryptase concentration above 20 ng/mL is suggestive of mastocytosis [5]. Additionally, tryptase serves as a potential biomarker for some hematological malignancies [6].

In the diagnostic of anaphylaxis, serum tryptase is used as a biomarker in a retrospective manner to confirm the diagnosis [3]. During an anaphylactic reaction, the serum tryptase level increases throughout the acute phase of the reaction but rapidly returns to baseline after approximately 12 h [2]. Hence, it is important to know the acute serum tryptase level as well as the baseline level to avoid misinterpreting an important increase in acute tryptase in a patient with a very low concentration of baseline serum tryptase. An international consensus has proposed that a clinically significant acute increase in tryptase is defined as 120 % x baseline+2 ng/mL [7]. This equation is widely used in the clinic, even though other equations have been proposed [3, 8, 9].

Although serum tryptase is a well-established biomarker, only sparse information is available on its biological variation (BV) [10, 11]. Knowledge of the BV is crucial when evaluating if an observed change in serum level is actually due to an anaphylactic reaction or caused by BV. Therefore, the aim of this study was to estimate the magnitude of the short-term BV of tryptase, and based on this, calculate the reference change value (RCV). This will help the interpretation of serum tryptase measurements in the clinic.

Materials and methods

Results

Biological estimates

The overall mean concentration of serum tryptase in all patients was 4.0 ng/mL (range, 1.3–8.0 ng/mL) (Table 1). For each subject, the mean concentration and range of tryptase are illustrated in Figure 1. No correlation between mean tryptase concentration for each subject and age was found (r=0.21, p=0.251). None of the subjects met the 120 % plus 2 ng/mL change between any of their measurements.

Table 1:

Components of biological variation.

	Tryptase
Number of subjects	32
Number of samplesa	175
Mean, ng/mL	4.0
Range, ng/mL	1.3–8.0
CVI, %	3.7 (3.0–4.4)
CVG, %	31.5 (23.1–39.8)
CVA, %	3.4 (2.9–4.1)
IIb	0.16
RCVc, %
Increasing	13.3
Decreasing	11.8
Samples required, nd	1

1. Values in parentheses are the 95 % confidence intervals. ^aNumber of samples analyzed after exclusion of outliers; ^bII, index of individuality; ^cRCV, reference change value at 95 % significance; ^dn, samples required to estimate homeostatic set point within 10 % with 95 % confidence. CV_I, within-subject coefficient of biological variation; CV_G, between-subject coefficient of biological variation; CV_A, analytical coefficient of variation.

Figure 1:

Serum tryptase levels in apparently healthy subjects (n=32) ordered according to age with the youngest at the bottom (22 years) and the oldest at the top (66 years) for each sex. Subjects are coloured according to the number of samples contributed by the individual: dark blue, seven or six samples; light blue, five samples; grey, three samples. The coloured vertical lines indicate the mean values and the horizontal lines shows the ranges. The coloured spots illustrate the mean value of duplicates for each measurement.

The overall CV_I was low at 3.7 % (95 % CI 3.0–4.4 %) and noticeably lower than the CV_G, which was 31.5 % (95 % CI 23.1–39.8 %) (Table 1). Accordingly, the II was very low at 0.16. CV_A was 3.4 % (95 % CI 2.9–4.1 %) and only slightly lower than the CV_I. The RCV for an increase in tryptase was 13.3 % at a 95 % level of significance. Only one sample was required to provide an estimate of the homeostatic set point within ±10 % with 95 % confidence.

Biological estimates according to sex

Biological estimates were evaluated separately for men (n=10) and for women (n=22) (Table 2). A greater variability in tryptase concentration was observed in women, although no differences were found in mean concentration or CV_I between sexes. The variability was reflected by the CV_G, which was more than 3-fold higher for women (36.7 %, 95 % CI 26.3–47.1 %) than men (12.9 %, 95 % CI 8.3–17.5 %). Consequently, the II was lower for women (0.14) than for men (0.39). The variability in mean tryptase concentrations in women was primarily found in older women (Figure 2), however, no correlation between mean concentration and age was observed for men (r=−0.028, p=0.938), women of all ages (r=0.275, p=0.215), nor women ≥50 years (r=0.628, p=0.131).

Table 2:

Components of biological variation according to sex.

	Women	Men
Number of subjects	22	10
Number of samplesa	115	60
Mean, ng/mL	4.0	3.8
Range, ng/mL	1.3–8.0	2.7–4.9
CVI, %	3.7 (2.9–4.4)	3.6 (2.1–5.1)
CVG, %	36.7 (26.3–47.1)	12.9 (8.3–17.5)
IIb	0.14	0.39
RCVc, %
Increasing	12.9	13.4
Decreasing	11.4	11.8
Samples required, nd	1	1

1. Values in parentheses are the 95 % confidence intervals. ^aNumber of samples analyzed after exclusion of outliers; ^bII, index of individuality; ^cRCV, reference change value at 95 % significance; ^dn, samples required to estimate homeostatic set point within 10 % with 95 % confidence. CV_I, within-subject coefficient of biological variation; CV_G, between-subject coefficient of biological variation; CV_A, analytical coefficient of variation.

Figure 2:

Relationship between age and serum tryptase mean level for each sex. Women are indicated in orange and men are indicated in blue. The lines show linear regression fits for each sex.

Day-to day and semidiurnal variance

The day-to-day variance components were calculated based on the three 9 AM samples for each subject. No significant difference in serum tryptase concentration was observed between the three days (p=0.77, Figure 3). The semidiurnal variance components were calculated based on all the samples collected on day 2 (9 AM, 12 PM, 3 PM, 6 PM, 9 PM). A significant semidiurnal variance was observed between tryptase concentrations at 9 AM and 12 PM, and 3 PM and 9 PM (p<0.0001, Figure 4).

Figure 3:

Day-to-day levels of serum tryptase. Individual results (spots) are depicted. The horizontal lines shows the mean values.

Figure 4:

Within-day levels of serum tryptase from the second day. Individual results (spots) are depicted. The horizontal lines shows the mean values. Significant differences are marked with *p<0.05 and ****p<0.0001.

The mean concentration of serum tryptase at each time-point is depicted in Table 3 together with the day-to-day and semidiurnal variance components, which were comparable with the overall biological estimates.

Table 3:

Day-to-day and semidiurnal components of biological variation.

	Tryptase
Day-to-day
Number of subjects	32
Day 1, ng/mL	4.0
Day 2, ng/mL	4.0
Day 3, ng/mL	4.0
CVI, %	3.9 (3.0–4.7)
CVG, %	32.2 (23.7–40.8)
Semidiurnal
Number of subjects	27
12 PM, ng/mL	3.8
03 PM, ng/mL	3.8
06 PM, ng/mL	4.2
09 PM, ng/mL	4.0
CVI, %	3.4 (2.1–4.7)
CVG, %	31.4 (23.2–39.6)

1. Day-to-day and semidiurnal serum tryptase levels are presented as means. Values in parentheses are the 95 % confidence intervals. CV_I, within-subject coefficient of biological variation; CV_G, between-subject coefficient of biological variation.

Discussion

Despite serum tryptase being used in the diagnosis of several clinical conditions, especially anaphylaxis and mastocytosis, only few studies have addressed the BV of the biomarker. Yet, the presence of individual BV, both long-term as well as over a short period of days or hours, could affect the diagnosis significantly. Therefore, we here presented data on the degree of short-term BV of serum tryptase in a cohort of apparently healthy individuals. We observed no significant day-to-day variation, but found minute, yet significant, variations in the concentration over the day. Furthermore, a very low intra-individual variation was identified together with a low RCV. A considerable between-subject variation was found, which was substantial larger for women than men.

The very low within-subject BV observed was in line with a previous study evaluating long-term BV of serum tryptase over a 10-week period (CV_I=5.6 %) [10]. This implicates that the fluctuations of tryptase around the homeostatic set point in steady state are very low and that tryptase is under a tight homeostatic control. This was further confirmed here by a low RCV for an increase in tryptase of only 13.3 %, which shows that for an individual with a 4.0 ng/mL mean tryptase concentration, only an increase of 0.56 ng/mL can be attributed to BV. This clearly uncovers tryptase as a very sensitive biomarker for disease monitoring. For the diagnosis of anaphylaxis, an acute increase in tryptase concentration of 120 %+2 ng/mL has been recommended [7]. Using an individual with a mean tryptase concentration of 4.0 ng/mL as an example again, this would require an increase in concentration to at least 6.8 ng/mL to confirm the diagnosis, which is far above the BV of tryptase found in the present study. Hence, our data shows that a significant increase in serum tryptase is seen considerable earlier than estimated by the recommended equation. However, it has been suggested in retrospective studies that patients with elevated serum tryptase concentrations owed to diseases such as mastocytosis, atopy, and myeloid disorder have increased basal tryptase variability [8, 21]. Therefore, studies determining the BV in patients with elevated basal serum tryptase concentration would be highly relevant.

In accordance with previously reported data [10], the CV_G was substantially higher (31.5 %) than the CV_I, indicating that the serum tryptase exhibits great individuality. Because of the low CV_I and moderate CV_G, a low II of 0.16 was observed. An II >1.4 indicates great applicability of conventional reference intervals, whereas an II <0.6 indicates limited applicability [18]. Thus, this finding shows that an individual’s normal tryptase level only span a very limited part of the reference interval and that extreme changes for an individual will not necessarily result in values outside the reference range. This supports performing repeated measurements of tryptase as recommended in the diagnosis of anaphylaxis. Furthermore, it emphasizes that using personalized reference intervals could be preferable for serum tryptase [22]. Interestingly, we observed a greater CV_G, and thereby a lower II, for women than for men. Greater variability in tryptase concentrations was primarily seen in women above 50 years. This is surprising, as no sex-dependent differences in serum tryptase levels have previously been observed [23, 24] or were observed in the current study. Potentially, our finding of low variability in men may be due to low statistical power in the male group as only 10 men were included, with eight of these being under the age of 50. However, it has previously been determined that the power of the CV_I estimate is 0.99 when 10 individuals are included who each contributes six samples, two replicates of each sample are analyzed, and the ratio between CV_A and CV_I is around 1 as for men in this study [25]. Hence, we believe that the estimates for both women and men are reliable. Notably, all individuals included in this study had serum tryptase levels within the reference interval as the highest concentration measured was 8 ng/mL [17, 23].

We observed a minute, yet significant, variation in the tryptase level during the day, but no variation from day to day. In accordance with this, Dugas-Breit et al. [11] observed a slight decrease in tryptase concentrations at 2:30 PM and 4 PM compared to baseline samples at 9:30–11 AM in nine control individuals. In the present study, we similarly observed a significant decrease at 12 PM and 3 PM. Nevertheless, the observed changes were minor (4.0 ng/mL at 9 PM, 3.8 ng/mL at 12 and 3 PM), and thus not considered to be clinically relevant, as they are well below the required change for the diagnosis of anaphylaxis.

The very low CV_I found for serum tryptase sets high demands for the analytical performance of the serum tryptase analysis. The desirable analytical performance for CV_A is below or equal to half the CV_I [18]. Even though CV_A was relatively low in our study, it was higher than desired, however, few analytical methods would fulfill a requirement of a CV_A of maximum 1.85 %. Hence, careful precautions need to be taken to minimize variation as much as possible when measuring serum tryptase. We attempted to minimize pre-analytical variation by using a standardized blood sample collection procedure performed by four technicians only. Moreover, the samples were batch-analyzed in duplicate using a standardized protocol performed by a single laboratory technician. The analysis was run on the commercial Phadia Immunoassay Analyzer using the ImmunoCap™ sandwich immunoassay. This is the current state-of-the art set up for serum tryptase analysis, and the analytical performance of the method in our laboratory was better than what has been achieved by others [10].

The strengths of our study were the application of the checklists for BV studies [13], [14], [15] and the strict protocol to reduce the impact of pre-analytical variation on the results. Moreover, we included individuals with a wide age range (range 22–66 years with 19 individuals above 34 years), which contrasts with the previous data on BV of tryptase where only younger individuals (23–35 years) were included. However, there were also some limitations to consider. Even though no long-term stability data of tryptase exists, we included blood samples that had been stored at −80 °C for four years. Yet, all samples from the same individual were analyzed simultaneously and thus had been stored for the same period of time. Another limitation of the study was that few subjects contributed to the 6 and 9 PM blood draws, resulting in these estimates being less robust. Lastly, the inclusion of subjects was based on clearly defined inclusion and exclusion criteria, which was evaluated based on self-reported information from the subjects. However, no clinical tests were performed to verify the subjects’ health status and, hence, we cannot be completely certain that all subjects were actually healthy.

In conclusion, we showed that serum tryptase is a tightly controlled biomarker with a very low within-subject BV, without any significant day-to-day variation, and with only minor semidiurnal variations. In contrast, a considerable between-subject BV was observed, signifying that reference intervals should be cautiously applied with respect to interpreting tryptase levels. Altogether, our data on the BV of tryptase provides valuable information for the clinical interpretation of tryptase levels and underlines its utility as a biomarker for anaphylaxis as well as non-anaphylactic conditions.