Overall description of the registry population

Within the observational period spanning from 2000 to 2014, a cohort of 338 children and adolescents diagnosed with CD was enrolled in the registry and, consequently, deemed eligible for the current investigation. The dataset for these 338 individuals comprised a total of 3605 documentation forms from the registry, which have been incorporated into the analyses conducted for this study.

Description of the registry population at the time of inclusion in the registry

In the registry, 214 (63.3%) males and 124 (36.7%) females were enrolled. The median age of children and adolescents diagnosed with CD at time point \(\) was 11.96 years (IQR: 4.16), with males averaging 11.90 years (IQR: 4.51) and females 11.90 years (IQR: 4.23). No significant difference in age between genders was observed (t-test: p-value = 0.315). Approximately one-fourth of the children (n=88, 26.0%) belonged to age group A1a (<10 years), while approximately three-fourths (250, 74.0%) were assigned to age group A1b (10–17 years).

Table 1 Disease activity at time point \(\) by gender N=326Fig. 1

Initial disease activity with respect to: A Gender — There was no statistically significant difference in the initial disease activity between genders (t-test: p-value = 0.10), B Age group — No statistically significant difference in the initial disease activity was observed across distinct age groups (t-test: p-value = 0.39)

Initial disease activity (PCDAI) at time point \(\)

At time point \( t_0 \), an assessment of disease activity was available for 326 individuals, encompassing 95.8% of the pediatric and adolescent cohort. The average initial PCDAI score was 16.89 points (SD 12.81), with a median score of 15.0, spanning from 0 to 65 on a scale of 0 to 100. Table 1 shows the distribution of disease activity by gender at the time of inclusion in the registry (\(\)).

Table 2 Distribution of disease activity by age group

At the time of inclusion in the registry (\( t_0 \)), the mean PCDAI score for individuals in age group A1a was 15.85 points (SD 13.11), while those in age group A1b exhibited a mean score of 17.26 points (SD 12.72). Analyzing the data by gender, the mean PCDAI score at \(\) for males was 16.0 points (SD 13.22), whereas for females, it was 18.42 points (SD 17.59). No statistically significant disparity in initial disease activity was observed for both age groups and genders, as presented in Fig. 1. As shown in Table 2, approximately 30.7% of the children in age group A1a and 56.8% of the children and adolescents in age group A1b showed a PCDAI (Pediatric Crohn’s Disease Activity Index) score of < 10 points, indicating remission. At time \( t_0 \), mild disease activity was observed in 53.4% of children in age group A1a and in 12.0% of children in age group A1b. In addition, at time \( t_0 \), 11.36% of the children in age group A1a and 27.6% of the children in age group A1b had moderate to severe disease activity (Table 2).

In a subsequent analytical phase, an investigation was undertaken to elucidate the potential association between inflammation localization at time point \(\) and disease activity. To explore this relationship, a binary logistic regression analysis was executed. No significant correlations were identified between the specific localization and the PCDAI score. The results of this regression analysis are shown in Table 3.

Table 3 Results binary logistic regression analysis; dependent variable PCDAI, independent variable localizationFig. 2

Survival time analysis event= improvement in disease activity. A Stratified by age group A1a median time to improvement 2.41 years (95% CI: 2.19–2.65); A1b 1.80 years (95% CI: 1.70–1.92) log-rank test p<0.001. B Stratified by gender boys median time to improvement 1.93 years (95% CI: 1.81–2.09); girls 2.00 years (95% CI: 1.88–2.16), log-rank test: p= 0.3

Table 4 Cox proportional hazard model: independent variables: initial localization (L1–L4) age classes (A1a, A1b) and gender, Signif. codes: “\(^\)” 0.001, “\(^\)” 0.01, “\(^\)” 0.05Age, gender and localization as influencing factors on the duration until improvement of disease activity

In a subsequent analysis, we directed our focus toward the temporal dynamics of improvement in disease activity. The median time to achieve an improvement in disease activity in the overall population was determined to be 1.95 years (95% CI 1.87–2.07), with the median time to the subsequent follow-up visit being 1.77 years. In addition, the influence of age, represented by the different age groups A1a and A1b, as well as the influence of gender on the time to improvement in disease activity were investigated. As shown in Fig. 2A, the median time to improvement in disease activity in age group A1b was significantly shorter than in age group A1a (A1a: 2.41 years vs. A1b: 1.80 years) (Fig. 2A). The median time to the next follow-up visit was 2.18 years for age group A1a and 1.63 years for age group A1b. The median time to improvement in disease activity was 1.93 years for boys and 2.00 years for girls. No significant difference in the duration until improvement induction could be identified (Fig. 2B). A Cox regression was also performed. The result showed that the hazard ratio for achieving improvement in disease activity was significantly 39% higher in age group A1b (HR 1.43; 95% CI: 1.29–1.51, p-value< 0.0001) than in age group A1a.

Regarding follow-up time, it was observed that the median follow-up times for disease activity varied (remission: 1.98 years, mild: 1.44 years, moderate/severe: 1.51 years). The differences in follow-up times were significant in the ANOVA performed (F-test p< 0.001), with the post-hoc test indicating a significant difference only between remission and mild (Tukey-HSD p < 0.001). The follow-up times did not correlate with the timing of disease improvement.

In a further step, the original localization of the inflammation was included in the Cox model as a further possible predictor alongside age. It was found that disease localization limited to L4 is associated with a more favorable prognosis in terms of improvement in disease activity (Table 4, Fig. 3). However, it should be noted that this result is moderated by the age groups.

Fig. 3

Cox proportional hazard model reduced model: independent variable localization, age classes and gender

Therapy as an influencing factor on the duration until improvement in disease activityDrug monotherapy

At time \( t_0 \), 38.63% of children in age group A1a were administered mesalazine, while 44.8% in age group A1b received the same. Steroids were prescribed to 51.1% in age group A1a and 42.0% in A1b. Immunomodulatory therapy was initiated in 35.2% of subjects in group A1a and 41.2% in A1b at time \( t_0 \). Infliximab treatment was administered to 13.6% of children in age group A1a and 7.6% in A1b. Nutritional therapy was provided to approximately 12.5% of children and adolescents in both age groups at time \( t_0 \) (Table 5). Table 5 also presents the distribution of prescribed therapeutic combinations between the age groups.

Table 5 Distribution of treatment options between age groups A1a and A1b

In the subsequent analysis, the duration until the onset of an improvement in disease activity was examined, taking into account the specific drug therapies. First, the prescribed monotherapies were analyzed. Across the age groups A1a and A1b, the drug treatment of children and adolescents with infliximab showed the shortest median time until an improvement in disease activity was achieved (Table 6). In terms of age groups, the results show that children and adolescents in age group A1b benefit earlier from the prescribed monotherapies compared to age group A1a, as shown in Table 6 and Fig. 6 in Appendix A. Only in connection with infliximab therapy did the A1a age group show advantages over the A1b age group in the survival time analysis. At time \( t_0 \), 9% (n=31) of the children and adolescents had previously received treatment with infliximab. Among these individuals, 19 exhibited mild disease activity, while 2 demonstrated moderate to severe disease activity. Intriguingly, 10 children were in remission at this time point. However, these differences did not reach statistical significance in the log-rank test (Table 7, Fig. 6 in Appendix A).

Table 6 Survival analysis: influence of each of the five monotherapies on the duration until improvement in disease activityTable 7 Survival analysis: influence of medication (monotherapy) on the duration until improvement in disease activity by age groupFig. 4

Cox proportional hazard model: independent variable medication (monotherapy)

A Cox regression was then carried out in a further step to investigate the influence of the respective drug monotherapy on the duration until the onset of an improvement in disease activity. Globally, a significant log-rank test (p-value <0.001) was found, which indicates a heterogeneous effect of the drug monotherapies on the duration until improvement in disease activity. In the Cox proportional hazard model, infliximab monotherapy was associated with a more favorable prognosis in terms of duration to onset of improvement in disease activity (HR: 1.22, 95% CI: 1.10–1.35). Figure 4 also shows the superiority of infliximab therapy over other drug monotherapies across the two age groups. Interestingly, therapy with mesalazine (HR: 0.82, 95% CI: 0.76–0.88) and with steroids (HR: 0.92, 95% CI: 0.86–0.99) was not associated with a positive effect on the duration until the onset of improvement in disease activity. Similarly, nutritional therapy alone (HR: 0.94, 95% CI: 0.83–1.06) was not associated with a positive effect on the duration to improvement of disease activity in the registry population; however, this result was not statistically significant (Fig. 4).

Infliximab is associated with a shorter time to achieve improvement in disease activity compared to alternative monotherapies, as evidenced by the age group-specific analysis. Here, age group A1a with an HR of 1.48 (95% CI: 1.24–1.76) appears to be superior to age group A1b with an HR of 1.12 (0.99–1.27) (Fig. 7 in Appendix A). However, the advantage of infliximab in age group A1b compared to other drug monotherapies was not significant (see Fig. 7 in Appendix A).

Table 8 Survival analysis: influence of combination treatment on the duration until improvement in disease activity in all age groupsFig. 5

Cox proportional hazard model: independent variable medication (combination therapy)

Drug combination therapy

In order to also examine the impact of treatment combinations, we replicated the earlier analyses, this time incorporating the treatment combinations as independent variables. Table 8 reveals that combination treatments involving nutritional therapy and mesalazine (HR: 1.88, 95% CI: 1.50–2.62), infliximab and nutritional therapy (HR: 1.95, 95% CI: 1.20–2.62), and a combination therapy of immunosuppressants and infliximab (HR: 1.98, 95% CI: 1.75–2.27) are linked to a shorter median time for the induction of improvement in disease activity. However, the somewhat wide confidence intervals suggest that these estimates lack robustness, likely due to a limited sample size. This presumption was subsequently corroborated by Cox regression analysis, which failed to identify any statistically significant effects regarding the duration until the induction of improvement in disease activity (Fig. 5). However, Fig. 5 indicates a tendency in favor of a combination treatment involving infliximab and nutritional therapy, or a combination therapy with immunosuppressants and infliximab.

In the age group-specific analysis, combinations of treatment involving nutritional therapy and mesalazine, as well as nutritional therapy and steroids, exhibited a significantly shorter median time for inducing improvement in disease activity in age group A1b compared to age group A1a (Table 9, Fig. 8 in Appendix B). Conversely, for age group A1a, the combinations of infliximab treatment with nutritional therapy, along with the combination of immunosuppressants and infliximab, appeared to confer advantages in terms of the duration until the induction of an improvement in disease activity. However, these findings did not reach statistical significance in the log-rank test between the age groups. It is worth noting that some results are characterized by wide confidence intervals, particularly in groups with smaller case numbers (Table 9, Fig. 8 in Appendix B).

Table 9 Survival analysis: influence of combination treatments on the duration until improvement in disease activity by age group

In the subsequent Cox regression analysis (Fig. 9 in Appendix B), within age group A1a, the combination treatment involving immunosuppressants and infliximab was notably linked to a significantly higher likelihood of a shorter time to the induction of improvement in disease activity. While combinations such as nutritional therapy with mesalazine and infliximab with nutritional therapy exhibited positive trends in terms of the duration of disease improvement, these effects did not reach statistical significance (Fig. 9A in Appendix B). Notably, for age group A1b, the Cox regression did not reveal any statistically significant effects. However, the presence of very wide confidence intervals was observed, likely attributable to the limited sample size (Fig. 9B in Appendix B).