The observed increase in mortality rates due to ILD among patients with IBD is concerning and highlights a critical need for systematic ILD screening protocols within the IBD patient population to facilitate early detection and management.
Inflammatory bowel disease (IBD) is characterized with chronic and recurrent inflammatory involvement of gastrointestinal tract with unpredictable clinical course. It comprises 2 distinct diseases of ulcerative colitis (UC) and Crohn's disease (CD). Between 1990 and 2019, there was a notable rise in the prevalence, disability adjusted life years, and years lived with disabilities related to IBD and it is predicted to rise until 2025 (). Both CD and UC have been experiencing a global rise in incidence and prevalence through previous decades (). In North America, the incidence of UC and CD stands at 19.2 and 20.2 per 100,000 person-year, respectively (), and the prevalence of IBD surpasses 0.3% (). Patients with IBD face higher risk of death from gastrointestinal problems, infection, respiratory diseases, nonalcoholic liver diseases, cardiovascular diseases, and cancers ().
Patients with IBD are at a greater likelihood of having other chronic inflammatory conditions such as arthritis, asthma, bronchitis, psoriasis, and pericarditis compared with general population (). Respiratory involvement in patients with IBD is frequently overlooked, as it may develop insidiously with no specific or noticeable respiratory symptoms (). Interstitial lung disease (ILD), an inflammatory disease of respiratory system, is an extraintestinal manifestation in patients with IBD and it is more common than previously perceived (). Underdiagnosis of ILD in patients with IBD may result in irreversible destruction of the lung and significant morbidities ().
Recognizing the avoidable factors contributing to the mortality of patients with IBD is critical for shaping forthcoming strategies in both prevention and management and might result in improved outcome of these patients. Hence, in this study, we aimed to investigate the risk of mortality due to ILD in patients with IBD using a nationwide database.
We conducted this population-based epidemiological survey to evaluate contribution of ILD as a cause of death of patients with IBD in the United States, from 2001 to 2020. Data were obtained on September 7, 2023, from Wide-ranging Online Data for Epidemiologic Research (WONDER) database provided by the Center for Disease Control and Prevention. WONDER is an interactive, public, web-based tool. It includes data from death certificates of US citizens in all 50 states and the District of Columbia (). Underlying cause of death reported in the death certificates is coded in accordance with the International Classification of Diseases, 10th Revision (ICD-10), since 1999. In WONDER, when a category contains fewer than 10 individuals, the data are suppressed for confidentiality reasons. Rates are labeled as unreliable when the death count is less than 20.
We defined population groups based on age at death (grouped as 15-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75-84, and 85+ years old), sex (female, male), race/ethnicity (American Indian/Alaska Native, Asian-pacific islanders, Black or African American, and White), residential census region (Northeast, Midwest, South, and West), and year of death (grouped into 2000-2005, 2006-2010, 2011-2015, and 2016-2020). To access data on the cause of death, we used ICD-10 code J84 indicating ILD, ICD-10 code K50 for CD, and ICD-10 code K51 for UC. As our study spans to 2020, we use ICD-10 code U07.1 to specify the coronavirus disease 2019 (COVID-19) as a cause of death.
For evaluation of mortality, we generate number of deaths, age-adjusted mortality rates (AMR) per 1,000,000 population, with 95% confidence intervals (CIs) for groups of population based on demographic variables. The crude mortality rates were also obtained and reported but excluded from the analysis because they could potentially provide misleading information when comparing rates across different periods and age groups. The AMRs are calculated in WONDER by creating a weighted average of the age-specific death rates. For the calculation of AMR, the population of year 2000 US standard is used. The equation for calculating the AMR is , where Psi is the standard population for a specific age group i, Ps refers to the total US standard population, and Ri is the age-specific death rate.
The all-cause and ILD-related mortality rates for IBD, CD, and UC were extracted from WONDER. Crude and AMR with 95% CI were presented regarding sex, residential census region, and race/ethnicity. For age groups, the crude mortality rate with 95% CI was presented. Linear regression analysis was conducted to evaluate the trend in AMR from 2001 to 2020. The AMR future trend until the year 2030 was predicted using the linear regression model. To assure the fitness of models, a polynomial regression analysis with an order of 3 was also performed. Mortality rates through the last period and the first period were compared with determine percent change among population groups. A P value < 0.05 was considered as significant.
From 2001 through 2020, IBD was responsible for 57,967 deaths with an AMR of 10.9 (95% CI, 10.810-10.989) per 1,000,000 population. Considering 5-year intervals, the AMR of IBD increased from 10.989 (95% CI, 10.796-11.181) in 2001-2005 to 11.443 (95% CI,11.271-11.615) within 2016-2020 per 1,000,000 population. The average IBD-related AMR increased by 39.9% from 2001 to 2005 to 2016-2020 (Supplementary Digital Content, https://links.lww.com/CTG/B135).
ILD is defined as the contributor in cause of death for 692 patients with IBD from 2001 to 2020, with an AMR of 0.11 per 1,000,000 population (95% CI, 0.10-0.12). The share of ILD as a cause of mortality in patients with IBD increases from 1.02% in 2001-2005 to 1.30% in 2016-2020, with an increase in the number of deaths by almost 44% from 129 deaths in 2001-2005 to 229 deaths in 2016-2020. The average AMR per 1,000,000 population increased from 0.092 (95% CI, 0.075-0.109) in 2001-2005 to 0.143 (95% CI, 0.123-0.162) in 2016-2020 (P value, 0.01, r, 0.795). Female patients and male patients with IBD died due to ILD at an average AMR of 0.111 (95% CI, 0.099-0.124) and 0.132 (95% CI, 0.117-0.147) per 1,000,000 population through 2001 to 2020, respectively. Male patients experienced a similar upward trend as the whole population with a fluctuation through 2011-2015 (P value, 0.01), while female patients had a steady rise (P value, 0.01). The average AMR increased by 89.9% among male patients and 67.1% among female patients (Table 1, Figure 1).
The ILD-related AMR of patients with IBD per 1,000,000 citizens of Northeast rises by approximately 86.3% from 0.082 (95% CI, 0.049-0.128) in 2001-2005, 0.095 (95% CI, 0.060-0.143) in 2006-2010, 0.118 (0.074-0.178) in 2011-2015, and ultimately to 0.150 (95% CI, 0.105-0.208) in 2016-2020 (P value, 0.01). Midwest citizens experienced a decline from 0.163 (95% CI, 0.118-0.219) in 2001-2005 to 0.108 (95% CI, 0.072-0.155) in 2011-2015 and then an incline to 0.129 (95% CI, 0.094-0.172) in 2016-2020 (P value, 0.01). Most deaths (30%) happen in the South census region with an overall increase in AMR from 0.062 (95% CI, 0.040-0.091) in 2001-2005 to 0.119 (95% CI, 0.092-0.152) in 2016-2020. The highest increase was in the mortality of West citizens who experienced a 132.2% rise from 31 deaths in 2001-2005 to 72 deaths in 2016-2020 (P-value, 0.011) (Figure 2).
Among different race and ethnicities, the ILD-related AMR of patients with IBD in American Indian or Alaska Native, Asian-pacific islanders, and Black or African American was not calculated due to the low number of deaths in these groups. White population accounted for a total of 652 deaths (94%) with an AMR of 0.119 (95% CI, 0.097-0.142) in 2001-2005, 0.125 (95% CI, 0.104-0.146) in 2006-2010, 0.139 (95% CI, 0.115-0.162) in 2011-2015, and 0.167 (95% CI, 0.143-0.190) in 2016-2020. Table 1 and Figure 1 present more details on ILD-related mortality of patients with IBD. The crude numbers and graphs of polynomial regression analysis are provided in Supplementary Digital Content (see Supplementary File, https://links.lww.com/CTG/B135).
The crude death rate was not calculated for age groups of 15-24, 25-34, 34-44, and 45-54 years due to low number of deaths in these age groups, which together was 65 deaths from 2001 to 2020. For other age groups, an upward trend in the crude death rate was observed. Table 2 is presenting details of death distribution among age groups.
From 17,578 deaths in patients with IBD through 2016-2020, COVID-19 counts for 296 deaths, all recorded in 2020. None of the patients with IBD who have ILD as a cause of mortality in 2020 had concurrent COVID-19. We also perform an additional analysis by removing 2020 data (Supplementary Digital Content, https://links.lww.com/CTG/B135).
From 2001 to 2020, 35,409 US citizens died due to CD, which compromised approximately 61% of IBD deaths. The AMR of patients with CD (per 1,000,000) was 6.714, 6.476, 6.462, and 6.916, in 2001-2005, 2006-2010, 2011-2015, and 2016-2020, respectively. For UC, it was 4.296, 4.123, 4.129, and 4.544, respectively, in the same periods.
Adding the ILD as a cause of death, the crude number of deaths in 2016-2020 (134 deaths) is almost twice the number of deaths in 2001-2005 (68 deaths), with an increase in AMR from 0.051 (95% CI, 0.039-0.066) in 2001-2005 to 0.087 (95% CI, 0.071-0.104) in 2016-2020 (P value, 0.009). For patients with UC, an upward trend was observed in the ILD-related AMR of them (P value, 0.009), with a rise from 0.052 (95% CI, 0.039-0.068) in 2001-2005 to 0.061 (95% CI, 0.049-0.075) in 2016-2020 (Table 3, Figure 3).
From 2001 to 2020, the United States experienced an overall 44% increase in the ILD-related mortality of patients with IBD. Our results also revealed an upward trend in the mortality of patients with UC and patients with CD due to ILD. Both sexes had experienced an upward trend in the ILD-related mortality. We demonstrated that most of the ILD-related mortality in patients with IBD have occurred in the South US census region, with the West region experiencing the highest change.
To the best of our knowledge, the share of ILD in the outcome of patients with IBD has not been explored previously. Studies focusing on ILD in patients with IBD are mostly case reports which may be due to lack of defined protocol to screen for ILD in patients with IBD, presentation of patients with unspecific symptoms, and mostly asymptomatic involvement (). While keeping in mind that we do not have incidence and prevalence rate of ILD in patients with IBD through WONDER database, 2 hypotheses warrant investigation; first, we may be observing a higher rate of ILD diagnosis in patients with IBD due to improved coding of causes of death. Second, alongside the rising AMR of ILD in both the general population and among patients with IBD, there might also be an increasing incidence rate of ILD in patients with IBD. Involvement of the respiratory system in patients with IBD has been reported since the 1970s. However, there was a debate whether ILD in patients with IBD is an extraintestinal complication of the disease, a side effect of medications, or a coincidence of 2 unrelated diseases (). Given the shared embryological origin of the gastrointestinal and respiratory systems, the involvement of lung in IBD is not surprising. Individuals who have a genetic tendency toward IBD, have a higher likelihood of developing ILD (). Moreover, a common pathophysiologic pathway, proinflammatory fibroblasts, is involved in the pathogenic activation of both diseases (). Smoking is another factor which may be involved in the pathogenesis of pulmonary involvement in patients with IBD. Smoking exhibits contrary effect on CD and UC, elevating the risk of CD while providing a protective effect against UC (). In nonsmoking female patients with UC, a higher prevalence of lower airway involvement is documented which may precede the manifestation of IBD in younger patients (). In this study, we lack data on the prevalence of smoking in our study population; hence, we cannot define the contribution of smoking in the trend of ILD-related AMR.
Pulmonary manifestations are more prevalent in adults () and include airway inflammation and obstruction, bronchiectasis, bronchiolitis obliterans organizing pneumonia, vasculitis, pleurisy, and ILD (). Symptoms of respiratory involvement are unspecific including productive cough, chest pain, and dyspnea (). Parenchymal disease frequently develops regardless of the activity level of IBD (). Latent involvement is more prevalent than apparent respiratory symptoms (). Zhao et al () conducted pulmonary function test (PFT) for 74 patients with IBD; abnormal PFT and decreased residual volume were observed in 75% and 51% of patients, respectively. They reported that only as few as 2.7% of patients with IBD had respiratory symptoms. Except for decrease in diffusing capacity of the lungs for carbon monoxide (DLCO), there is no specific PFT pattern in patients with IBD. The impairment includes obstructive and/or restrictive pattern or isolated reduction in forced expiratory volume in second one and forced vital capacity ().
Patients with IBD have an overall 1.4 times higher risk of death compared with the general population. Respiratory involvement accounts for the second cause of death in patients with IBD and is responsible for 13.2% and 14.2% of the mortality of patients with CD and patients with UC, respectively (). Studies have demonstrated an increasing trend in the risk of death from respiratory diseases among patients with CD and patients with UC (). In this study, we found that the share of ILD as a cause of death of patients with IBD has been doubled through 2001 to 2020. However, as we used the WONDER database, which is based on death certificate, we cannot make any statement about the incidence of ILD in the population of patients with IBD. Indeed, we hope our study could raise the question about the prevalence and incidence of ILD in this population of patients with IBD as it seems to contribute to their death.
In this study, we explore the contribution of ILD to the mortality of patients with IBD, in a population-based epidemiological study through a time span of 20 years. This wide range data could reveal the rising proportion of ILD in the mortality of patients with IBD and added to the significance of our results. However, this study encounters some limitations. The database we used in this study relies on death certificates, which may be biased toward underreporting or overreporting of causes of death by clinicians. This also may have happened for recording the COVID-19 as a cause of death in 2020. Moreover, as many cases of pulmonary involvement in patients with IBD remains undiagnosed, it is probable that the actual rate of ILD-related mortality in patients with IBD is higher than what is reported in the death certificates. This study evaluates AMR among all patients with IBD; however, it fails to evaluate it in ethnic/race groups other than White patients. While it seems to be related to the lower number of mortality in these ethnic/race groups, we cannot rule out the possibility of underreporting. Finally, considering smoking as a potential confounding factor in this study, we used ICD-10 code F17.2 to define tobacco consumption as a cause of death, which identifies 113 patients with IBD between 2001 and 2020. This underscores the likelihood of underreporting regarding the impact of smoking on the study's outcomes.
Contribution of ILD in the outcome of IBD as a cause of mortality reveals the probability of underdiagnosis and high prevalence of ILD in patients with IBD. Moreover, the rising trend of ILD-related mortality of patients with IBD underscores the importance of considering and managing lung involvement, as a preventable cause of death in these patients to reduce mortality rate. This insight could lead to the development of screening programs, ultimately enhancing outcomes of patients with IBD.
Guarantor of the article: Mehdi Mirsaeidi, MD, MPH.
Specific author contributions: A. Vaezi: conceptualization, formal analysis, drafting the manuscript, she has approved the final draft submitted; T. Ashby: interpreting data, validation, she has approved the final draft submitted; M. Schweitzer: interpreting data, validation, he has approved the final draft submitted; P. Ghali: interpreting data, validation, he has approved the final draft submitted; M. Mirsaeidi: conceptualization, methodology, validation, supervision, drafting the manuscript, he has approved the final draft submitted.
Financial support: None to report.