Disease Activity Patterns of Paediatric Inflammatory Bowel Disease: A Danish Nationwide Cohort Study (1996–2018)

Abstract

Background and Aims

Inflammatory bowel diseases [IBD] are heterogeneous in the frequency and severity of their flare-ups. We aimed to describe disease activity patterns in a Danish nationwide paediatric IBD cohort.

Methods

Paediatric patients [<18 years at diagnosis] with Crohn’s disease [pCD] or ulcerative colitis [pUC] in the study period from 1996 to 2018 were identified in national registers. Disease activity [severe, moderate-to-mild, remission] was assessed at diagnosis according to medications prescribed, hospitalizations, and surgeries.

Results

In total, 1965 pCD and 1838 pUC incident patients were included in the cohort. At diagnosis, severe disease activity was found in 87%/80% of pCD/pUC and in addition 6.1% of pUC patients had undergone a colectomy during the first year after diagnosis. Five years after diagnosis, the annual proportions of pCD/pUC with no disease activity were 70%/61%, and 10 years after diagnosis the proportions were 72%/64%. Colectomy was required in 6.1, 12, and 16% of pUC patients after 1, 5 and 10 years. No improvement of disease activity was seen in the proportion of prevalent pCD [N = 2515] and pUC [N = 2428] in the study period 2000–2018 concomitant with the introduction of biological treatment. However, decreasing disease activity was the most common pattern in both pCD and pUC [43 and 47%], respectively.

Conclusions

pIBD was characterized by a high proportion of patients with severe activity at diagnosis, followed by an improvement after 5 and 10 years of follow-up. Notably, the proportion of patients with no disease activity was unchanged when biological treatment was introduced and the number of colectomies in pUC remained high.

1. Introduction

Inflammatory bowel diseases [IBD] comprise Crohn’s disease [CD] and ulcerative colitis [UC]; both are immune-mediated diseases of unknown aetiology. IBD are characterized by chronic gastrointestinal inflammation and are heterogeneous in terms of disease presentation, the frequency and severity of flare-ups, and their long-term prognoses.^1,2

Several studies have suggested that patients with paediatric-onset IBD [pIBD] have a more aggressive phenotype than those with adult-onset IBD, with more extensive intestinal inflammation, more frequent flare-ups, and a greater risk of requiring treatment with biologicals.^3–5

A single Italian paediatric-onset UC [pUC] cohort study published in 2021 reported that most pUC patients had a benign disease pattern, but one-third of its patients had chronic continuous or chronic intermittent disease activity.⁶

In a previous Danish adult inception cohort of patients diagnosed between 1962 and 1987, the proportion of the CD population in remission each year stabilized at 55%.⁷ In the UC population, 40–50% of all patients were in remission after a few years, but not exceeding 60% in the 25-year study period; the crude rate of colectomy was 20.2% after 25 years.⁸

We recently re-investigated the disease activity patterns in patients with adult-onset IBD between 1995 and 2018 in Denmark. Most of these adult patients had either an improving or quiescent disease activity pattern, meaning that disease activity was most often present only immediately after a diagnosis.⁹

When communicating ‘what to expect’ with newly diagnosed patients and their families, an illustration of the disease activity patterns would be of major importance. Therefore, we aimed to describe and illustrate the disease pattern in a Danish nationwide cohort of all IBD patients with pIBD [<18 years old at diagnosis] between 1996 and 2018, coinciding with the time when biologicals were introduced as a standard treatment. Furthermore, we compared the findings from this paediatric cohort with our similar study of adult IBD disease activity patterns.

2. Materials and Methods

2.1. Study population

Eligible patients were identified in the Danish Civil Registration System, in which all Danes are registered with a unique ten-digit identification number given at birth. The Danish Civil Registration System provides demographic information about sex, date of birth, date of death, and migration. The source population was linked to the National Patient Registry [NPR], which contains information about all hospitalizations in Denmark since 1977, and all outpatient visits and emergency department contacts, procedures, and surgical interventions since 1995. For this study, we identified all patients registered with an IBD diagnosis [ICD-10: K50, K51] between 1996 and 2018 and then linked individuals to other registries. We then identified all individuals with IBD using an algorithm based on diagnostic codes, pathology reports, and IBD-specific surgeries and medications to ensure no misclassifications in our population. All patients with three or more IBD registrations were included. If a patient only had one or two IBD registrations, they were included if they had a record of an IBD-relevant operation and/or a pathology registration with an unambiguous SNOMED IBD code. In the remaining patients with one or two IBD registrations, we searched for redeemed prescriptions for IBD-relevant medications. First, we included patients with two or more prescriptions for mesalazine or rectal steroids, or vedolizumab, which are treatments used exclusively for IBD. Among patients not fulfilling this requirement, we nonetheless included any with a combination of at least two prescriptions or deliveries/treatments at the hospital of thiopurines, mercaptopurine, or methotrexate, or one treatment of anti-tumour necrosis factor [TNF]-α inhibitors, anti-IL 12/23 antibodies, or JAK inhibitors, but here we excluded patients with co-occurring immune-mediated dermatological or rheumatological diseases where these drugs might also be used. The sensitivity and positive predictive value [PPV] for this algorithm was estimated to be 98% and 69% respectively.¹⁰ Follow up started at the first IBD registration. For classifying CD and UC patients, cases with CD or UC diagnosis codes exclusively were categorized as CD or UC, respectively. The remaining patients were categorized as IBDU, unless: [1] 80% or more of their codes were either CD or UC; [2] they had a record of a CD-specific surgery code; and [3] if patients were recorded with CD or UC codes exclusively during the preceding 5 years of observation, they were classified as CD or UC, respectively.

Information about medications was retrieved from the Danish National Prescription Registry [DNPR]; however, records of thiopurine, methotrexate, and biological therapy were sourced from the NPR. Patients were also linked to the Danish Pathology Data Bank containing patho-anatomical information. The registry was established in 1970 and became nationwide in 1997.

We included all living patients diagnosed with IBD before the age of 18 years between January 1^st, 1977 and December 31^st, 2018. Patients from this cohort who were alive between January 1^st, 1996 and December 31^st, 2018 were defined as prevalent cases. Medications given to children were first separated from the personal identification number of the mother beginning in 1996. We therefore defined incident patients as those with a first IBD diagnosis recorded between January 1^st, 1996 and December 31^st, 2018.

Patients who had a registered diagnosis of autoimmune hepatitis, rheumatological and/or dermatological diseases and who were treated with regimens that we used to define disease activity patterns were excluded. Patients who had undergone a total colectomy more than 1 year prior to their first IBD diagnosis were also excluded [see Supplementary Materials and Table 1 for a list of diagnoses].

2.2. Definitions of disease activity

To define disease activity patterns, we used a combination of information about IBD-related hospitalizations [with a duration longer than 1 day], IBD-related surgeries [excluding perianal surgery], redeemed prescriptions for IBD treatments (oral or topical corticosteroids, oral or topical budesonide, topical 5-aminosalicylic acid [5-ASA]), and biological therapies. Remission was defined as the absence of these events. Further information about our definitions can be found in Supplementary Tables 2 and 3.

We categorized an episode of active disease as mild–moderate if it required a course of topical corticosteroids, oral budesonide, or topical 5-ASA. Episodes were categorized as severe if they involved hospitalization, surgery, oral corticosteroids, or biological therapy.

2.3. Definitions of disease activity patterns

Two different pre-defined disease activity patterns were used based on 5- and 10-year periods respectively.

2.3.1. Disease activity patterns during 5-year periods

We assessed disease activity patterns in 5-year periods and categorized each 5-year period as follows: [I] no years with active disease; [II] one or two years with at least one episode of disease activity; [III] three or four years with at least one episode of disease activity; and [IV] each of the five years had at least one episode of disease activity.

This assessment was made for all incident patients with at least 5 years of follow-up. A favourable course was defined as a reduction in the number of years with active disease in the second 5-year period compared to the first 5-year period or was categorized as having ‘No years with activity’. Vice versa, an unfavourable disease course was defined as an increase in the number of years with active disease in the second 5-year period compared to the first 5-year period or a stable number of years with active disease in both periods or colectomy in pUC patients within the two 5-year periods from diagnosis.

2.3.2. Disease activity patterns during the 10-year period

The adult disease activity patterns described by Wintjens et al.¹¹ were used to categorize the change in disease activity of incident pIBD patients who were diagnosed between 1996 and 2008 and who underwent at least 10 years of follow-up:

• [I] Decreasing activity: first 5 years included two or more years with activity, followed by 5 years with two years or fewer with activity.

• [II] Increasing activity: first 5 years included fewer than two years with activity, followed by 5 years with two or more years with activity.

• [III] Chronic continuous activity: first 10 years had at least eight years of activity.

• [IV] Quiescent activity: first 10 years had one year of activity or no activity.

• [V] Chronic intermittent activity: first 10 years had an activity pattern that does not meet the criteria above.

2.4. Statistical analysis

Patients were followed until December 31^st, 2018, their emigration, or death. Patients with pUC who had undergone a total colectomy were categorized to have no subsequent disease-related events and were not included in the cumulative numbers for medical treatments after their colectomy. We described the number of patients in each disease activity group by counting individual patients’ most severe event, rather than the number of events for each individual year. In the incident population, all patients were included in each yearly assessment after their diagnosis. Follow-up was analysed according to years since diagnosis, so the incident population shrank over time. In the prevalent population, every calendar year in the study period was assessed, and so the increasing yearly incidence meant that the number of patients increased over time. We investigated the distribution of patients in each category of disease activity [remission, mild–moderate, severe, and UC patients having undergone a colectomy] according to the year of their IBD diagnosis.

Descriptive data are given as numbers, percentages [%], medians, and interquartile ranges [IQRs], as appropriate. Pearson’s chi-squared test, Fisher’s exact test, Wilcoxon rank sum test, and Kruskal–Wallis rank sum test were used, as appropriate, to describe categorical and numerical variables. A significance level of p < 0.05 was considered statistically significant. All statistical analyses were performed with RStudio [2022] [RStudio Inc.].

2.5. Ethical considerations

According to Danish law, ethical approval is not required for registry studies.

3. Results

A total of 4943 prevalent pIBD patients [2515 pCD and 2428 pUC] were included in the study. Of these, 1965 pCD [78.1% of the prevalent population] and 1838 pUC [75.7% of the prevalent population] were incident cases. Figure 1 is a study population flowchart. Baseline characteristics for all incident patients are provided in Table 1. The most common causes of death were cancer [22%, N = 8/37], suicide [19%, N = 7/37], and accident/assaults [16%, N = 6/37]. Treatments are given as cumulative numbers in Supplementary Table 4. In incident patients 43% and 21% of pCD and pUC patients, respectively, received biological therapy.

3.1. Yearly status of disease activity

For incident pCD, severe disease activity occurred in 87% of patients [N = 1705/1965] in the year of their diagnosis, while only 3.0% of patients [N = 59/1965] had mild–moderate disease. The remaining 10% [N = 201/1965] of pCD patients had no disease activity registered in the year of their diagnosis. In the tenth year of disease, the proportion of pCD patients in remission during a given year increased steadily to 72% [N = 680/946] [Figure 2].

For incident pUC, severe disease activity occurred in 80% of patients [N = 1462/1838] in the year of their diagnosis, in addition to the 6.1% [N = 113/1838] who underwent a colectomy, while 6.4% of patients [N = 118/1838] had mild–moderate disease [Figure 2]. The remaining 7.9% [N = 245/1838] of pUC patients had no disease activity registered in the year of their diagnosis. After 10 years of disease, the proportion of pUC patients in remission during a given year increased to 64% [N = 602/942].

After 5, 10, and 20 years, 12% [N = 180/1441], 16% [N = 153/942], and 21% [N = 39/186] of pUC patients, respectively, had undergone a colectomy. No major differences were found between male and female patients [Supplementary Figures 1 and 2].

For the prevalent cohort, a similar assessment of disease activity was performed each year to measure the proportion of patients in each category of disease activity. Among patients with pCD and pUC, no clear changes occurred over time in the proportions in remission, with mild–moderate or severe activity. The proportion of pUC patients undergoing colectomy was also stable. The percentage of patients with pCD and pUC in remission each year was within a range of 59–85% and 50–62%, respectively. These findings are illustrated in Figure 3.

3.2. Disease activity pattern among incident patients during the 5-year periods after diagnosis

I: No disease activity was experienced in 5% of pCD and pUC patients (pCD: 5.1% [N = 70/1366] and pUC: 4.4% [N = 58/1323]).

II: The majority of both pCD and pUC patients (pCD: 51% [N = 701/1366] and pUC: 53% [N = 699/1323]) experienced one or two years with at least one episode of active disease.

III: One-third of pCD and pUC patients (pCD: 36% [N = 485/1366] and pUC: 35% [N = 461/1,323]) had episodes of active disease that occurred at least once in three or four years.

IV: Eight per cent of pCD and pUC patients (pCD: 8.1% [N = 110/1366] and pUC: 7.9% [N = 105/1323]) had episodes of active disease occurring in all five years.

In incident patients with at least 10 years of follow-up, most patients experienced a reduction in the number of years with active disease between the first and second 5-year periods, as shown in Supplementary Figure 3.

Leaving out the year of diagnosis [year 0], we assessed five periods in patients with 11 years of follow up. This was illustrated in Figure 4. About one-third of both pCD [31%] and pUC [34%] patients had no disease activity in years 1–6 and more than half of these stayed in the no year of disease activity group for years 7–11 [pCD: 52% and pUC: 62%].

Table 2 presents the therapies given according to favourable and unfavourable disease courses. In pCD patients with a favourable course there were no differences regarding biological therapy within the first 5-year period compared to patients with an unfavourable course [6.1% vs 7.9%, p = 0.3]; this was also the case for pUC patients [23% vs 19%, p = 0.2].

3.3. Disease activity patterns during the 10-year period after diagnosis

The distribution of pCD patients according to their 10-year disease activity patterns is described in Table 3.

The cumulative frequencies of IBD-related treatment within the different categories of disease activity are shown in Table 4; note that the distribution of pUC patients undergoing colectomy reflects our exclusion of disease activity in pUC patients in the years after a colectomy.

There was no difference in when thiopurines were initiated in pCD patients according to disease activity patterns. However, thiopurines were initiated earlier in pUC patients with decreasing disease activity, increasing disease activity, and quiescent disease.

The time to initiation of biological therapy differed among the pCD patterns with earlier initiation in the chronic continuous and decreasing disease activity patterns. For pUC, biological therapy was initiated earlier in patients with decreasing disease activity and quiescent disease, but when excluding patients with colectomy within the first 10 years from diagnosis, there was no significant difference among the disease activity patterns.

4. Discussion

In this Danish nationwide cohort of pIBD patients, we described the disease activity patterns of pIBD between 1996 and 2018. Overall, at diagnosis we found severe disease activity in 87% of pCD and 80% of pUC patients with decreasing disease activity over time. Despite the introduction of biological therapy during the study period, the proportions of prevalent patients had no clear improvement in disease activity throughout the study period, including pUC patients undergoing colectomy.

A multicentre, open-label, randomized, controlled trial with newly diagnosed moderate to severe pCD patients assessed the impact of first-line infliximab [five doses only] vs conventional induction therapy. With this top-down approach, no significant difference was observed in the proportion of patients in clinical remission with thiopurine maintenance after 1 year.¹²

Aloi et al. described the disease activity patterns in a population-based Italian cohort of pUC. They found decreasing disease activity in 24% of patients, increasing activity in 11%, chronic continuous activity in 19%, chronic intermittent activity in 14%, and quiescent disease in 33%.⁶ These results are also comparable to our findings. However, we observed half as many patients with chronic continuous activity, and more patients with decreasing disease activity. The Italian study used 5-year periods to describe the disease activity patterns, and included other parameters in addition to treatment escalation, hospitalization, and surgery to define disease activity, albeit in a smaller cohort. The shorter study period and additional measurements for flare-ups could have caused these differences in patient distribution. In the Italian pUC cohort, 8% had undergone a colectomy in the 5 years after their diagnosis, whereas we observed a slightly higher 5-year colectomy rate of 12%. However, both of these rates are lower than those found in other studies.^13,14

Previous studies have described differences in disease course among patients with paediatric- and adult-onset IBD. In a French cohort, the proportion of patients with disease activity was greater in pCD than in adults in a year-by-year assessment. pCD patients had more frequent flare-ups and more often required immunosuppressants; however, these studies found no difference in the rates of resections.¹⁵ In an Israeli nationwide UC cohort, immunosuppressants and biologics were more frequently used in pUC than in adult UC patients.¹⁶ As with our cohort, no changes in paediatric colectomy rates were observed despite the increased use of biological therapy. The impact of biological therapy on the disease course remains unclear, as cohort studies have found no changes in disease course or surgical rates among either paediatric- or adult-onset IBD patients.^9,17–19 However, our study did not evaluate whether the number of acute surgeries has fallen over time. pCD and pUC patients with a favourable course were not more often treated with biological therapy within the first 5-year period compared to patients with an unfavourable course. This observation is difficult to interpret as some calendar period effect probably occurred, and thus a growing proportion of patients with less severe disease course will have been exposed to biologics over time.

We previously described the disease activity patterns in a Danish adult IBD cohort during the same period.⁹ We found that more patients in the paediatric cohort had severe disease in the year of diagnosis compared with adults. Eighty perc ent of the adult CD cohort experienced severe disease in the first year after a diagnosis compared to 87% in our pCD cohort. In the adult UC cohort, 3.4% of patients underwent a colectomy in the year of diagnosis, yet we observed twice [6.4%] as many colectomies in the pUC cohort; this difference persisted as the adult 10- and 20-year colectomy rates were 9.5 and 14% [pUC: 16 and 21%], respectively. The 10-year disease activity patterns describing a more aggressive disease course [chronic continuous, chronic intermittent, and increasing activity] were more frequent in the pCD cohort, at 13, 20, and 6%, respectively, than in the adult cohort, at 6, 16, and 4%, respectively. The distribution of 10-year patterns among UC patients was almost identical in the adult and paediatric cohorts, despite the fact that thiopurines and biological therapy were more frequently given for each of the disease patterns in both pCD and pUC than in the adult cohort. The use of systemic corticosteroids in CD patients was similar in the adult and paediatric patients in terms of disease activity pattern; however, pUC patients were more likely to be treated with systemic corticosteroids for each disease activity pattern than their adult UC counterparts.

Current monitoring and therapy reduce the disease activity on an individual level and most patients have a benign disease activity pattern. This is reassuring for the newly diagnosed patients and their families. However, ~10% of both pCD and pUC patients experience chronic continuous disease activity. It is well known that the chronic activity for a paediatric patient often results in disease-related complications and difficult educational attainment, all possibly resulting in poor adult quality of life.

The major strengths of this study are the nationwide design with minimal loss to follow-up in a Danish setting of free and universal healthcare services, which reduced selection bias and ensured a reporting of the full spectrum of disease severity. The limitations of this study include missing data on nutritional therapy and not capturing cases with oral 5-ASA monotherapy to treat flares. Furthermore, patients might also have had topical 5-ASA and/or steroids stored at home and for use during flare-ups; if so, these episodes of activity were not recorded and might have caused us to underestimate episodes of mild–moderate disease activity. Information about clinical activity scores, endoscopic activity, disease location and behaviour, and smoking habits was not available. We did not assess perianal disease in pCD patients. The minor decrease in the number of patients with active disease from 2010 onwards was a product of our methodology, whereby there were fewer newly diagnosed patients in later years, as potential patients did not yet meet the criteria that we used to identify IBD patients; the complex definition of IBD will automatically result in a decreasing number and proportion of incident pIBD patients for the last available years, since many incident patients will not have time to fulfil the multi-hit algorithm. Furthermore, due to our methodology, pUC patients undergoing colectomy were categorized according to the timing of the operation in the 10-year disease activity patterns. Thus, patients with early colectomy were categorized as having decreasing disease activity or quiescent disease, as this study did not consider complications in the years after colectomy. Similarly, when reporting 5-year period patterns, if a pUC patient had a colectomy 2 years after diagnosis, the following 3 years would be considered as no disease activity. However, we took this into account when reporting if the course was favourable or unfavourable.

In conclusion, an increasing proportion of pIBD patients will probably enter remission year-on-year after their diagnosis. Despite the introduction and over time slightly increased use of biological therapy during the study period [1996–2018], the proportion of prevalent pIBD patients in remission during each calendar year of follow-up remained unchanged. We observed a more severe 10-year disease activity pattern in pCD, a more frequent use of corticosteroids within each activity pattern in pUC, and higher rates of colectomy in pUC than in their adult counterparts. These data suggest a more severe disease phenotype in pIBD, and we await the long-term results on the disease activity pattern of the nowadays more early, intensive, and serologically monitored biological therapy.

Funding

No funding was received for this study.

Conflict of Interest

MDW: grants from Novo Nordisk Foundation and Bristol Meyers Squibb, unrelated to the work submitted.SJ, MM, VW: no disclosures relating to the work submitted. JB: personal fees from AbbVie, Janssen-Cilag, Celgene, and Pfizer. Grants and personal fees from MSD. Grants and personal fees from Takeda for participating in advisory boards. Grants and personal fees from Tillotts Pharma. Personal fees from Samsung Bioepis. Grants from Bristol Myers Squibb and Novo Nordisk. All unrelated to the work submitted.

Author Contributions

MDW: wrote the original draft, carried out the analyses, and helped conceptualize the study. SJ, MM: responsible for reviewing and editing the manuscript, interpreting the data, and revising the manuscript for important intellectual content. VW, JB: study conceptualization, supervision of MDW, reviewing and editing the manuscript, interpreting the data, and revising the manuscript for important intellectual content.

Data Availability

The data referred to in this article will be shared upon reasonable request made to the corresponding author.

References

Author notes