Evolution of surgical trends in Crohn’s disease during the biologic era: population-based cohort insights from Taiwan, a low-endemicity region
Article information
Abstract
Background/Aims
Surgery remains a crucial treatment option for Crohn’s disease (CD), even with the introduction of biological agents. This nationwide cohort study in Taiwan investigates surgery trends and the impacts of biologics in a region with a low prevalence of CD.
Methods
This retrospective population-based cohort study used Taiwan’s National Health Insurance Database from 2003 to 2018. The cohort included 725 CD patients. Patient characteristics, surgery outcomes, and impact of the biologic era on surgical risk were analyzed.
Results
During the study period, 292 CD patients (40.3%) underwent surgery, with 125 in the pre-biologic era and 167 in the post-biologic era. The incidences of intestinal surgery (IS) and perianal surgery (PS) have significantly decreased. The cumulative probabilities of IS were 20%, 35%, and 44% after 1, 5, and 10 years, respectively; the PS incidences were 3%, 5%, and 7%, respectively. The cumulative incidence of IS was significantly lower in the post-biologic era compared to the pre-biologic era (P=0.049). CD patients had high second IS incidences of 31% at 5 years after the first IS.
Conclusions
Our study demonstrates the surgical incidences have decreased in the biologic era but remained relatively high in a region with low disease prevalence. This suggests the need for further improvements in CD management.
INTRODUCTION
Crohn’s disease (CD) poses multifaceted challenges due to persistent intestinal inflammation, giving rise to complications such as intestinal obstruction, abscess, fistula, and fulminant colitis. Consequently, surgical intervention is common, with approximately 70%−80% of patients with CD undergoing surgery at some point in their lives [1,2]. Notably, up to 70% of postoperative CD patients require additional surgical intervention, indicating the relapsing nature of the disease process [3].
Biologic treatments targeting tumor necrosis factor have demonstrated efficacy in reducing the risk of surgery in CD patients, as evidenced by randomized controlled studies [4]. However, real-world studies present inconsistent findings regarding the effectiveness of biologics in reducing surgery rates. Some population-based studies indicate a declining trend in surgery rates during the biologic era [5], whereas others report no significant differences [6-8]. This variability may result from factors beyond biologic treatment, such as early and accurate diagnosis or multidisciplinary management, potentially influencing prognosis. Furthermore, perianal involvement poses a significant risk for CD-related surgeries [9,10]. However, the current status and trends of perianal surgeries are inadequately represented in real-world data.
Most existing studies on the impact of biologics on surgery rates predominantly focus on regions with high endemicity, such as Europe and North America. Although a study from Korea demonstrated a decline in surgical rates among CD patients in the post-biologic era [11], studies on CD surgical rates remain scarce in regions with a low prevalence of CD. Notably, the incidence of CD has increased in newly urbanized countries over the past decades, including Taiwan [12,13]. Hence, exploring the conditions in these regions is crucial. Therefore, this study analyzes the trends and impact of the biologic era on intestinal surgery (IS) and perianal surgery (PS) in a nationwide register-based cohort of CD patients in Taiwan.
METHODS
1. Ethical Approval
Approval for the study protocol was obtained from the National Cheng Kung University Hospital Institutional Review Board (approval number: A-EX-108-049). The informed consent was waived.
2. Data Source
This study utilized data from the National Health Insurance (NHI) Database of Taiwan, spanning from January 1, 2003, to December 30, 2018. As a comprehensive data repository, the NHI database encompasses health information for over 99.9% of Taiwan’s population [14], facilitated by the single-payer NHI program initiated on March 1, 1995. Extracted from the claims data of NHI beneficiaries, the database comprises, ambulatory care claims, inpatient claims, prescriptions, medical facilities, and board-certified specialists. To ease the financial burden on families dealing with major illnesses, the NHI recognizes 31 categories of catastrophic illness, including CD. Patients with CD can apply for a Catastrophic Illness Certificate (CIC), exempting them from copayments.
3. Study Population
This study included patients with a CIC for CD (ICD-9-CM: 555 or ICD-10-CM: K50) from 2004 to 2017 (Supplementary Table 1). The CIC application process involves a detailed review of clinical, radiologic, endoscopic, and histologic findings by experienced gastroenterologists, ensuring a high level of diagnostic accuracy. Included patients were required to have a follow-up period of at least 1 year both before and after their initial CD diagnosis. To minimize diagnostic overlap with infectious colitis in the early stages of disease, we further refined our study population by considering their medical records within the 3 years following their initial diagnosis. Specifically, we included patients who had at least 1 hospitalization or 4 outpatient visits related to CD or those who had at least 4 prescriptions for CD-related medications in any given year within this 3-year period.
To identify new incidence cases, this study excluded patients who had any CD-related medical records in the year prior to their initial CD diagnosis. To verify the diagnosis of CD, patients who only received corticosteroid treatment in the first year following their initial diagnosis were also excluded. Moreover, patients under the age of 18 or those without registered age or sex information were also excluded.
In this study, the start date for follow-up was defined as the date of the first prescription for CD-related medication, which is also defined as the diagnosis date of CD. This study followed patients from the diagnosis date and censored them until the occurrence of surgery, death, or the end of the study period in the database (i.e., December 31, 2018), whichever came first. Additionally, the biologic era was defined as starting from July 01, 2011, coinciding with the availability of the first biologic (adalimumab) in Taiwan. The Anatomical Therapeutic Chemical codes for biologics are provided in Supplementary Table 2.
The NHI program reimburses biologics for CD patients with CIC who meet 1of the following criteria: (1) uncontrolled disease (Crohn’s Disease Activity Index ≥ 300) or severe drug-related side effects after more than 6 months of conventional treatment (5-aminosalicylic acid, steroids, and immunosuppressants); (2) unhealed perianal or abdominal wall fistulas (Crohn’s Disease Activity Index ≥ 100) despite over 6 months of conventional treatment or surgery; or (3) undergone 2 or more surgeries within a year due to CD complications despite adequate conventional treatment. The reimbursement period for biologics treatment is limited to 12 months.
4. Study Outcomes
This study focused on 2 primary outcomes concerning surgical interventions in CD patients, namely IS and PS. IS was further categorized into small bowel surgery and ileocolonic surgery. Small bowel surgery included procedures involving the small bowel and ileocolonic surgery included operations on the ileocolon, colon, or both the colon and small bowel. Restorative surgery, such as colostomy reversal, was excluded from the count of surgeries in this study. PS included surgery for anal fistula or abscess. Detailed procedure codes for these surgeries can be found in the supporting information (Supplementary Tables 3 and 4). This study assessed the annual IS and PS rates from 2007 to 2018 in CD patients and obtained the cumulative probabilities of the first and second surgeries using Kaplan-Meier curves.
5. Statistical Analysis
The statistical results are presented as counts (percentages), means (standard deviations), or medians (interquartile ranges). The surgery rate was calculated as the number of surgeries in a given year divided by the total number of CD patients in that year. To assess the trend in surgery rates over time, we conducted a linear regression analysis, treating the year as a continuous independent variable and the surgery rate as the dependent variable. The cumulative probability of surgery was evaluated using the Kaplan-Meier method, and the curves for patients in the pre- and post-biologic eras were compared using the log-rank test. The Cox proportional hazards model was used to estimate hazard ratios and 95% confidence intervals for comparing the cumulative probability of first IS and PS between the pre- and post-biologic eras. All statistical computations in this study were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA).
RESULTS
1. Baseline Clinical and Demographic Features at CD Diagnosis
A total of 725 eligible CD patients were included. The screening process of the study is shown in the Supplementary Fig. 1. The basic characteristics of the included patients are given in Table 1, and the diagnostic codes used to define baseline characteristics are provided in Supplementary Tables 5 and 6. Among CD patients, males were slightly more prevalent, and the mean age at diagnosis was 38.7 ±15.4 years. The occurrence of anemia was 24.7% in CD patients. The proportion of patients with concurrent immunosuppressive diseases, extraintestinal manifestations, or cancer was less than 10% for CD patients. During the observation period, the proportion of surgeries was 35.9% for IS and 8.7% for PS among CD patients.
Baseline Characteristics at Diagnosis of 725 Crohn’s Disease Patients
In Taiwan, adalimumab was the first biologic agent available for CD patients and became accessible in 2011. Vedolizumab and infliximab were introduced in late 2017, followed by golimumab in 2018. Consequently, the majority of patients in our study who received biologic therapy were treated with adalimumab (92.2%), while 6% received vedolizumab, and 1.8% were treated with either golimumab or infliximab. The mean duration from CD diagnosis to initiation of biologic therapy was 4.46 ± 3.52 years.
2. Characteristics of Enrolled CD Patients Who Underwent Surgery
Among the cohort, 292 CD patients underwent CD-related operations. Their characteristics are shown in Table 2. The patients were predominantly male, with a mean age of 40.4 years at the time of surgery. Prior to diagnosis, 65 of these patients (22.3%) had undergone IS and 26 (8.9%) had undergone PS. The medication profiles preceding the first operation are also shown in Table 2. The “Operation before diagnosis” variable in Table 2 refers to surgeries performed before the confirmatory CD diagnosis (22.3% in IS, 8.9% in PS).
Baseline Characteristics at Diagnosis for 292 Crohn's Disease Patients with Operation
3. Time Trends in Surgery Rates
The time trend in surgery rates for CD patients is shown in Fig. 1. A statistically significant reduction in the rate of IS was observed in CD patients (P=0.001), from 25.3% in 2007 to 4.5% in 2018. The rate of PS also significantly decreased in CD patients (P=0.028) during the study period, dropping from 5.1% in 2007 to 1.6% in 2018.
Annual intestinal and perianal surgery rates from 2007 to 2018 among Crohn’s disease patients.
4. Cumulative Risks of First Surgery
The initial surgery risks were analyzed. The results are shown in Fig. 2. In individuals with CD, the cumulative probabilities of IS were 20%, 35%, and 44% after 1, 5, and 10 years, respectively. The cumulative probabilities of PS in CD patients were 3%, 5%, and 7% after 1, 5, and 10 years, respectively. Remarkably, 8.9% of CD patients had already undergone IS before the official disease diagnosis.
Cumulative probability of first intestinal surgery (IS) and perianal surgery (PS) in patients with Crohn’s disease (CD).
5. Impact of the Biologic Era on Surgical Risk in CD
The cumulative incidence of first IS at 1, 5, and 10 years after diagnosis was 18.4%, 29.0%, and 63.6%, respectively, in the pre-biologic era. In the post-biologic era, the cumulative incidence of first IS was 12.1% at 1 year and 21.6% at 5 years after diagnosis. The cumulative incidence of first IS was significantly lower in the post-biologic era compared to the pre-biologic era (P=0.049, log-rank test) (Fig. 3). The hazard ratio for IS in the post-biologic era compared with the pre-biologic era was 0.73 (95% confidence interval, 0.53–0.99). There was no significant difference in the cumulative incidence of PS between the pre-biologic and post-biologic eras (P=0.739, log-rank test). The hazard ratio for PS in the post-biologic era compared with the pre-biologic era was 1.18 (95% confidence interval, 0.45–3.08).
Cumulative probability of first intestinal surgery in patients with Crohn’s disease (CD) in the pre-biologic era and the post-biologic era.
6. Cumulative Risks for Second Surgery
In individuals diagnosed with CD, the cumulative probabilities of having a second IS were 16%, 26%, and 31% at 1, 3, and 5 years after the first surgery, respectively. Similarly, the cumulative probabilities of second PS for CD patients were 17%, 29%, and 32% (Fig. 4).
Cumulative probability of second intestinal surgery (IS) and perianal surgery (PS) in patients with Crohn’s disease (CD).
DISCUSSION
This national population-based study revealed significant reductions in operation rates for CD patients in Taiwan from 2007 to 2018, consistent with previous Western or Asian findings [5,11]. While various factors may have contributed, advancements in biologic therapies likely played an important role in reducing surgery rates among inflammatory bowel disease patients [15,16]. Biologic therapy has been reported with a 26% reduction in the risk of surgery for CD patients compared to that of patients without such treatment [16]. Our data further show that, according to national data, the incidence of IS in this low-endemic area had decreased significantly in patients diagnosed in the post-biologic era (Fig. 3).
Beyond the use of biologics, several other factors can be associated with the surgical rate, such as penetrating behavior, ileal involvement, disease severity, and duration [17]. Additional unmeasured confounders, including physician experience, diagnostic awareness, the use of monitoring tools, and treatment escalation strategies, could also have impacts. Due to limitations of the NHI database, key clinical variables such as disease behavior, location, and severity were not available. Given these constraints, assessing only on database available parameters might lead to an incomplete understanding of their individual effects. Therefore, our study focused on evaluating the overall influence of the biologic era at a population level, rather than evaluating specific predictors.
Despite these limitations, we sought to address concerns regarding potential confounding factors. A prospectively registered study from Taiwan showed similar rates of penetrating disease (18% in 2006–2015 vs. 16.67% in 2016–2023) and ileal involvement (33.33% in 2006–2015 vs. 30.77% in 2016–2023) [18]. Therefore, we speculated that the phenotypic distribution of CD was likely stable during the period of our study. Although azathioprine may represent a potential confounder, previous meta-analyses and 1 clinical study have not demonstrated a significant association between its use and a reduced risk of surgery [19,20]. Moreover, our recently published study also based on NHI database but in different definition showed the most characteristics of patients diagnosed in pre- and post-biological era were similar [21]. Therefore, based on current literature, surgery-related confounders appear to have a limited impact on our study conclusions, although the possibility of residual confounding cannot be entirely excluded.
However, our study revealed a reduction only in IS rates, not PS rates, among CD patients in the post-biologic era. This result aligns with a recent study showing that the probability of the first PS for CD has remained stable across different eras [22]. This stability may be attributed to the refractory nature of perianal CD and its predisposition to relapse. Additionally, biological therapy has not demonstrated a dramatic effect on treating perianal CD. Two studies found that the incidence and recurrence rates of perianal CD remain unchanged in the biologic era [10,22]. One study reported that only 24% and 32% of CD patients with perianal fistulas achieved radiological healing after 1 and 3 years of combination therapy with thiopurines and anti-tumor necrosis factor, respectively [23]. The study found that 57% of CD patients who achieved healing experienced recurrent fistula after discontinuing biologics [23]. Therefore, many patients with perianal CD still require surgical intervention despite biologic therapy [9]. Another possible explanation is the restrictions imposed by Taiwan’s NHI on biologic reimbursement policy due to budgetary considerations. According to the policy, CD patients must exhibit an inadequate response to conventional therapy for over 6 months before biologics are initiated, with treatment discontinuation mandated after 52 weeks [24]. This limitation may impede the efficacy of biologics in preventing surgery, especially for perianal disease, a common complication requiring early and long-term continuous biologic therapy [25,26].
The cumulative risk of IS for CD patients in our study is consistent with a prior meta-analysis reported in 2013 [27]. However, a more recent meta-analysis reported that contemporary risks in the 21st century for IS rate in CD patients were 12.3%, 18.0%, and 26.2% at 1, 5, and 10 years post-diagnosis, which are lower than those found in our study [28]. The higher surgery rate observed in our CD patient cohort may be attributed to several factors. First, low awareness of IBD due to the low CD prevalence led to delayed diagnosis and treatment, resulting in higher complications and operation rates [29-31]. Second, endemic gastrointestinal infections in Taiwan pose diagnostic and therapeutic challenges due to conflicting therapies for infection and CD, particularly tuberculosis [32]. Finally, the first biologic was not available until 2011, and several restrictions on the application of biologics in Taiwan’s NHI program, as mentioned earlier, could have contributed to the relatively inadequate treatment compared to that in other countries.
In contrast, cumulative risk of first PS in our study is substantially lower than in previous studies [33]. This finding may be attributed to the low prevalence of perianal CD in Taiwan, which was reported to be 14.8% in a recent study [34]. Consequently, the proportion of perianal CD-related surgeries among all patients is relatively low. Moreover, perianal CD in East Asia may present with less severe disease and a better prognosis compared to Western countries [35]. The low surgical rate may also be associated with multidisciplinary management and medical and radiologic treatment advancements [36]. However, the cumulative risk of second PS was high, which could be attributed to relapsing and refractory characteristics of perianal CD, as mentioned previously [10,22,37].
Our study identified a higher incidence of second IS than those in recent studies, which reported second IS rates of 11%−14% at 5 years and 16% at 10 years after diagnosis in the 21st century [38-40]. This disparity may be attributed to factors similar to those contributing to the higher first IS rate observed in our study, especially the policy of forced discontinuation of biologics after 52 weeks. In Taiwan, after the forced discontinuation of biologics, 3-quarters of CD patients experienced disease relapse [41], and more than half of CD patients encountered suboptimal outcomes within 1-2 years [24]. It is reasonable that the high cumulative probability of a second IS for CD under this policy results from the short duration of the biologic treatment (52 weeks), which may not achieve endoscopic remission in most patients [42]. Given the chronic and relapsing nature of CD, consistent and uninterrupted disease control in the therapeutic window following the first operation is crucial to preventing a second IS.
Similar to the case of Taiwan’s NHI program, limitations on biologic treatment coverage in public insurance have been reported in Canada and Mexico, leading to increased hospitalizations and emergency room visits [43,44]. Belgium and U.S. studies also showed that adherence to biologic treatment has been associated with reduced healthcare costs [45], decreased hospitalization expenses [46], and improved overall survival and quality of life for patients with inflammatory bowel disease [47]. Moreover, reimbursing biologics was cost-effective and led to a decrease in the lifelong surgery rate for Chinese CD patients [47]. This evidence highlights the need to reevaluate biologic insurance reimbursement policies in Taiwan and other countries with similar budget constraints, advocating for more flexible approaches to enhance care for CD patients.
The strengths of our study include the utilization of the NHI database, which covers over 99.9% of Taiwan’s population, and the inclusion of CD patients with catastrophic illness registration, ensuring comprehensive and accurate data. Moreover, focusing on current practice patterns and outcomes in Taiwan, a low-endemicity region, provides valuable insights. Nevertheless, our study has several limitations. As an observational study, establishing direct causality for the observed reduction in surgery rates is challenging. The absence of detailed clinical data, including disease characteristics, severity, and specific surgery types, limits a comprehensive understanding [5,48]. Second, the majority of patients receiving biologic therapy in our study were treated with adalimumab (92%). The long-term impact of newer advanced therapy on surgical trends needs further investigation. Third, some CD patients may not have been included due to the absence of CIC registration. However, because CIC registration is required for accessing biologic therapy, the most clinically relevant cases were likely captured. Although the lack of non-CIC patients may affect estimates of first-surgery rates, the impact on second-surgery rates and comparisons between the pre- and post-biologic eras is expected to be minimal. Fourth, the limited sample size in our study precluded a meaningful analysis of the impact of the biologic era on the risk of a second surgery.
In conclusion, our study reveals a significant reduction in operation rates for CD patients in Taiwan, a region with a low disease incidence. The impact on reducing surgical incidence in biologic era is demonstrated, particularly for IS. Although establishing a direct cause-and-effect relationship remains challenging, the observed trends suggest the importance of continuous application of biologic therapy. Our findings also suggest that the cumulative risks of first and second surgeries are higher compared to those reported in contemporary studies, indicating the need for ongoing improvement in CD management.
Notes
Funding Source
This research was supported by research grants from the Ministry of Health and Welfare (MOHW113-TDU-B-211-114008), National Cheng Kung University Hospital (NCKUH-11302055), and National Science and Technology Council (NSTC 113-2314-B-006-057-MY3).
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Data Availability Statement
The data used for this research were obtained from the National Health Insurance (NHI) Database of Taiwan. Data will be made available on request.
Author Contributions
Conceptualization: Yang EH, Chen NY, Cheng CL, Chuang CH. Data curation: Yang EH, Chen NY, Chang YC. Formal analysis: Yang EH, Chen NY, Cheng CL, Chang YC, Chuang CH. Funding acquisition: Chuang CH. Investigation: Chen NY, Cheng CL, Chen PC, Wu LH, Kang JW, Chiang HC, Chen PJ, Lin BW, Kuo HY, Chuang CH. Methodology: Yang EH, Chen NY, Cheng CL, Chuang CH. Project administration: Chuang CH. Resources: Cheng CL, Chuang CH. Software: Chen NY, Chang YC. Supervision: Cheng CL, Chuang CH. Validation: Cheng CL, Chen PC, Wu LH, Chuang CH. Visualization: Yang EH, Chen NY, Chang YC. Writing - original draft: Yang EH. Writing - review & editing: all authors. Approval of final manuscript: all authors.
Supplementary Material
Supplementary materials are available at the Intestinal Research website (https://www.irjournal.org).
Supplementary Table 1.
Operational Definitions for Inflammatory Bowel Disease Diagnosis
Supplementary Table 2.
Operational Definitions for Biologics
Supplementary Table 3.
Intestinal Surgery Procedure Code and Definition
Supplementary Table 4.
Perianal Surgery Procedure Code and Definition
Supplementary Table 5.
Operational Definitions for Inflammatory Bowel Disease-Related Surgery
Supplementary Table 6.
Operational Definitions for Baseline Disease
Supplementary Fig. 1.
Flow diagram of study population selection. IBD, inflammatory bowel disease; CIC, Catastrophic Illness Certificate; NHI, National Health Insurance; CD, Crohn's disease.