Baseline Serum Tryptase: Sex‐ and Age‐Specific Reference Intervals in the Pediatric and Adult Population

Abbreviations

bST

basal serum tryptase levels

CLSI

Clinical Laboratory Standards Institute

HαT

hereditary alpha‐tryptasemia

IQR

interquartile range

PARIS

Pollution and Asthma Risk: an Infant Study

RI

reference interval

SD

standard deviation

UGM₉₅

Unimodal Gaussian Model of the lowest 95th percentiles

To the Editor,

Recent advances toward a better understanding of baseline serum tryptase levels (bST) determinants, including hereditary alpha‐tryptasemia (HαT), have led to reconsidering reference intervals (RIs) for this biomarker [1, 2], with special interest in defining age‐ and sex‐specific intervals [3, 4, 5]. Direct methods to define RIs determine the n^th bST percentile in a representative group of the reference “healthy” population. This strategy requires at least n = 120 individuals by sex and age groups [6]. Conversely, indirect methods use statistical models to extrapolate RIs from large clinical databases. Skills in data analysis and concerns regarding the lack of external validation have hindered, so far, the diffusion of such strategies in practice.

In order to apply such an indirect strategy to determine age‐ and sex‐specific bST RIs, we first recruited a training cohort consisting of a nationwide database of 21,216 bST values obtained from unique participants through an ambulatory community‐based clinical database. The median age was 47.5 years, and the male‐to‐female ratio was 0.49. Baseline characteristics and raw bST concentrations in all the training cohort and per age category and sex category are presented in Table S1. The distribution of raw bST values in the overall study population is depicted in Figure 1A. Of note, 3.5% of the study population presented bST levels > 20 μg/L and 0.46% bST > 100 μg/L, potentially related to HαT, kidney disease, or underlying hematological diseases.

FIGURE 1.

Distribution of bST values in the training cohort and determined age‐ and sex‐specific reference intervals for bST values. (A) Distribution of bST values in the training cohort (n = 21,216). Of note, 0.46% (n = 97) participants of the training cohort presented with bST values > 100 μg/L, are not displayed in the histogram, but were included in the statistical analysis. The red line represents the modelization of the distribution by the UGM₉₅ model. (B–D) Reference intervals are depicted for a theoretical 1:1 male: Female population (B, gray), for females (C, orange), and males (D, purple). The plain bold midline represents the predicted median bST values in the reference population. The lower and upper dotted lines represent the predicted 2.5th and 97.5th percentiles of bST values in the reference population. BST, basal serum tryptase levels; UGM₉₅, Unimodal Gaussian Model on the lowest 95th percentiles.

Thus, statistical processing was needed to derive reliable RIs. We compared five different models to fit the lower side of log‐transformed raw bST values (see Additional Methods and Table S2). All led to similar estimates and accuracies. The more efficient model (Figure 1A) was then applied for each year of age and each sex category. Of note, the geometric SD of bST did not vary significantly across age and sex groups (Figure S1A).

Independently of age and sex, the reference median bST value was 4.6 μg/L, and reference values corresponding to the 1^st, 2.5^th, 5^th, 95^th, 97.5^th, and 99^th percentiles were 1.6, 2.2, 2.5, 8.4, 9.5, and 10.9 μg/L, respectively. Similarly to raw values (Table S1), the longitudinal bST trajectories showed lifelong higher levels for males, with a more pronounced difference among teenagers and young adults. The levels of bST increased steadily from teenagehood to old age for both sexes (Figures 1B–D and S1B).

Finally, we verified extreme bST reference percentiles in an external validation cohort, consisting of 572 HαT‐negative teenagers of 15/16 years participating in the population‐based PARIS birth cohort. The observed proportion of teenagers above their predicted 95^th, 97.5^th, and 99^th age‐ and sex‐specific percentiles was 4.9% (28/572, p = 0.98), 1.9% (11/572, p = 0.45), and 1.0% (6/572, p > 0.99), respectively. Similar results were obtained for the lowest reference percentiles and after stratifying for sex. Given the very high statistical power of these tests (always > 0.91), these results validate the reliability of our study design and bST reference percentiles to describe the general population.

This study had several strengths. First, the assessment of a real‐life cohort of over 20,000 bST values provided insights on bST values in the pediatric, adult, and geriatric populations. However, extreme ages (< 1 and > 80 years) were under‐represented in this cohort, and corresponding bST values should be cautiously interpreted. Another important limitation is the absence of clinical information or tryptase genotypes in the training cohort. Although recruitment was restricted to community‐based outpatients to limit enrichment of disorders from hospital‐based recruitment (e.g., anaphylaxis), common conditions (e.g., HαT, kidney disease, or underlying hematological malignancies) could not be entirely excluded. Another strength is the use of multiple methods to confirm the robustness of the statistical analyses. Finally, we used an external validation cohort which, although limited to population‐based teenagers, was well‐documented on the clinical and genetic levels. Extended validation at 18 months and 8/9 years in the PARIS cohort is planned for the future, and additional assessment in adults would be of great interest [7].

Beyond their pathophysiological implications, the clinical relevance of sex‐ and age‐dependent bST variations might be questioned. Up to 20% (541/2842) of bST values > 8.4 μg/L turned out to fall within participants' respective personalized RIs and should have been considered normal. We also reaffirm that bST assessment in healthy asymptomatic individuals, if an elevated value is found, will inevitably lead to diagnostic dilemmas. Figure 2 proposes an algorithm to clarify the use of personalized bST RIs to modulate the extent of investigations with or without clinical suspicion of clonal mast cell disorder.

FIGURE 2.

Proposed algorithm for the diagnostic work‐up and follow‐up of patients according to their age‐ and sex‐specific bST percentiles and the presence or absence of clinical suspicion of clonal mast cell disorder. ¹Relevant clinical events might include: Cutaneous lesions evocative of mastocytosis (e.g., Urticaria Pigmentosa/maculopapular cutaneous mastocytosis); mast cell activation syndrome; severe hymenoptera venom anaphylaxis; severe idiopathic anaphylaxis; anaphylaxis with REMA score ≥ 2; early, severe, and unexplained osteopenia/osteoporosis. ²Bone marrow investigations might comprise the following: search for mast cell cluster on bone marrow biopsy; search for mast cell abnormal morphology on bone marrow biopsy or bone marrow smear; search for mast cell aberrant expression of CD2, CD25, or CD30 on bone marrow biopsy or bone marrow aspirate; search for KIT D816V or other KIT activating mutations on bone marrow aspirate. Of note, positive PB KIT D816V should trigger bone marrow investigations unless already performed. ³For instance, individuals with α‐duplications usually display bST < 30 μg/L, while some rare individuals with remarkably high α‐tryptase copy number can present with bST > 100 μg/L. ⁴Other explanations for mildly elevated bST might include: Overweight/obesity, chronic tobacco smoke exposure, other cardiovascular risk factors/metabolic syndrome, chronic urticaria, chronic kidney disease. Other explanations for highly elevated bST might include: Active chronic urticaria, advanced chronic kidney disease, myeloid hematological malignancy apart from clonal mast cell disorder. ⁵We propose that an elevation of follow‐up bST levels above +30% of baseline bST (i.e., at diagnosis) should be regarded as a significant bST increase over time. BM, bone marrow; bST, basal serum tryptase levels; cMCD, clonal mast cell disorder; PB KIT D816V, search for the KIT D816V mutation in the peripheral blood by a recommended method; REMA, Spanish Network on Mastocytosis; SM, systemic mastocytosis.

In conclusion, this study determined and validated sex‐ and age‐specific bST RIs in the pediatric and adult population. Personalized bST percentiles and RIs are freely available through an online tool at: https://sfa.lesallergies.fr/tryptase_calculator.

Author Contributions

Study conception and design: Y.C., J.V., I.M., C.K. Data acquisition: Y.C., C.B., S.C., C.R., F.R., I.M., C.K. Data analysis: Y.C., H.N., A.V.‐D. Manuscript drafting: Y.C., A.V.‐D., C.K. Manuscript revision: Y.C., A.V.‐D., C.B., P.A.A., E.F., J.G., P.M., M.M., A.N., E.S., C.R., F.R., J.V., I.M., C.K.

Conflicts of Interest

Y. Chantran serves on the scientific advisory board for Thermo Fisher Scientific, received honoraria from Thermo Fisher Scientific, and research grants from Blueprint Medicines and Thermo Fisher Scientific. M. Michel received honoraria from Thermo Fisher Scientific. J. Vitte reports speaker and consultancy fees in the past 5 years from Astra Zeneca, HpVac, L'Oréal, Novartis, Sanofi, Thermo Fisher Scientific, Zambon, and travel support from Stallergènes‐Greer, outside the submitted work. C. Klingebiel received honoraria from Thermo Fisher Scientific. J.G. reports speaker and travel support in the past 5 years from ALK, Stallergenes‐Greer, Thermo Fisher Scientific, and Menarini, outside the submitted work. Éric Fromentin—ALK‐Abelló, Menarini, Stallergenes Greer, Thermo Fisher, and Viatris. The other authors declare that they have no relevant conflicts of interest.

Supporting information

Figure S1: Geometric standard deviations of bST levels in each age and sex group and the modelization of age‐ and sex‐specific bST reference intervals.

Table S1: Baseline characteristics and tryptase values in the training cohort.

Table S2: Comparison of methods for indirect determination of the bST reference values.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.