Development of the EAACI% season definition a backup for a global application

CONFLICT OF INTEREST

All authors declare no conflicts of interest.

To the Editor,

Pollen seasons progress differently in their timing, course, and intensity in different countries/biogeographical regions depending on regional factors such as vegetation, elevation, urbanization, and others. The variable regional situation often provides obstacles for a standard season definition. ¹  The season definition of the European Academy of Allergy and Clinical Immunology (EAACI) was published as a pollen concentration season definition depending on the selected pollen type reaching a certain threshold after a certain period of consecutive days, ² , ³ and it was demonstrated that they can be correlated with pollen‐induced symptom loads. ⁴  The season definition was developed for studies with a medical framework, particularly to allow a prospective approach. However, pollen concentrations do not reach the same level in Europe or globally and affect persons concerned differently. ⁵  Therefore, there is a strong need to expand the scope of application of the EAACI season definition to a retrospective approach if the standard season definition criteria are not met within the site selection or during a clinical trial.

Hence, we tried to transfer the EAACI season definition criteria into a percentage definition (EAACI%) for birch and grasses in a first approach herein to include areas where the pollen concentrations do not meet the criteria of the EAACI season definition or small gaps in the data record prevent an EAACI season definition result. For this “backup” season definition, ten pollen monitoring stations across central and eastern Europe were used as the calculation basis for transforming the EAACI season definition.

A cumulative relative pollen concentration ($r_d$) was calculated for the start, end and peak days of the standard EAACI season definition applying the following formula:

$r_d=\frac{{\sum }_{i=1}^d{p}_i}{{\sum }_{i=1}^{365}{p}_i}$. Summarized, the cumulative pollen concentration ($p_i$) from the first day of the year to day $d$ is divided by the yearly total sum of the respective pollen type.

For further optimization, these values were calculated for all sites, pollen types, and years to minimize the mean absolute deviation in comparison with the standard EAACI definition.

This calculation results in four percentage values ($v$) per pollen type, which define the start and the end of the (peak) pollen season. If the relative cumulative pollen concentration exceeds one of the thresholds ($v$), the respective date marks the (peak) pollen season start/end of the EAACI% definition. For the result validation, the difference in the season's duration was compared with each of the definitions applied (Table 1).

TABLE 1.

Calculated thresholds of the EAACI% definition for birch ($v_{b,s}$) and grasses ($v_{g,s})$ for the start of the main pollen season ($v_{a,1}$), the end of the main pollen season ($v_{a,2}$), the start of the peak pollen season ($v_{a,3}$), and the end of the peak pollen season ($v_{a,4}$)

	Main season start ($v_{a,1}$)	Main season end ($v_{a,2}$)	Peak season start ($v_{a,3}$)	Peak season end ($v_{a,4}$)
Birch ($v_{b,s}$)	$v_{b,1}=$ 0.013 2 (1–2 days)	$v_{b,2}=$ 0.968 6.4 (2–10 days)	$v_{b,3}=$ 0.081 1.8 (1–3 days)	$v_{b,4}=$ 0.868 3.1 (1–5 days)
Grass ($v_{g,s})$	$v_{g,1}=$ 0.016 4.7 (2–6 days)	$v_{g,2}=$ 0.962 13.2 (5–21 days)	$v_{g,3}=$ 0.182 3.6 (1–7 days)	$v_{g,4}=$ 0.683 7.4 (2–13 days)

In addition, the mean absolute deviation for all stations and years as a comparison to the standard EAACI season definition is displayed, including the 25% and 75% quantiles in brackets.

Additional information regarding the methodology and results as well as a calculation example can be found in the Appendix S1. To evaluate the EAACI% season definition, the Austrian pollen monitoring site of Tamsweg (ATTAMS) was selected as an experimental station to compare both season definitions and to test the application to data, where the EAACI season definition fails (Figure 1).

FIGURE 1.

Course of the birch pollen season of the station Tamsweg (Austria) for the years 2015 and 2017. (A) Neither the main pollen season, nor the peak pollen season could be derived from the EAACI definition (no matching criteria), but only from EAACI% definition. (B) Both season definitions could be applied. The solid lines display the start and the end of the main pollen season and dashed lines the peak pollen season. The blue lines show the ranges of the EAACI season definition whereas the red lines display the ranges of the EAACI% definition

This first attempt demonstrates that the EAACI season definition criteria could successfully be converted into a percentage definition. We recommend using the herein described EAACI% season definition for predictive assistance in site selection or retrospective analysis in sites/countries where the EAACI season definition could not be applied during an ongoing trial. This transformed definition allows an accurate assessment of the pollen season using EAACI criteria in terms of start and end of the pollen season where the EAACI season definition would fail. A moving average season definition usually used in trend analysis, could improve the results of this definition in the future approaches. ⁶  Taking together this calculation may be helpful in trial regions of interest with low pollen concentrations, an exceptional less intense season, or other possible obstacles for pollen concentration‐based season definitions. However, further validation with clinical data from hayfever patients is needed.

Supporting information

App S1