OBJECTIVES: To provide age- and sex-specific paediatric reference intervals (RIs) for 13 haematological parameters analysed on Sysmex XN-9000 and compare different methods for estimating RIs after indirect sampling. METHODS: Via the Danish Laboratory Information System, we conducted a population-based study. We identified samples from children aged 0-18 years analysed at Aarhus University Hospital from 2019 to 2023, including samples from general practitioners only. Information about all parameters were available for all samples via linkage to the local laboratory middleware. Then, we applied two different methods. First, we excluded potential pathological samples by predefined criteria: if the child had other abnormal blood measurements at date of request, or had a blood sample of any type analysed in the period two months before to two months after. We estimated RIs stratified by age- and sex using the non-parametric percentile method. Second, we used refineR (an open source automated algorithm) to exclude pathological samples and for RI estimation. Finally, we compared our data to results from a study using the direct method. RESULTS: We identified 22,786 samples. After exclusion by predefined criteria, the population comprised 10,199 samples from 8,736 children (57 % of samples were from females and median age was 13 years). We estimated RIs for red blood cell, white blood cell and platelet indices. The two different methods showed agreement. Furthermore, our data provided results comparable to direct sampling. CONCLUSIONS: Our study provided age- and sex-specific paediatric RIs for 13 haematology parameters useful for laboratories worldwide. RIs were robust using different methods in the framework of indirect sampling. Finally, our data showed agreement with the direct method, indicating that indirect sampling could be useful for establishing RIs on haematology parameters in the future.