Cohort profile: ESPRESSO (Epidemiology Strengthened by histoPathology Reports in Sweden)

Risk of Severe Covid-19 in Patients with Celiac Disease: A Population-Based Cohort Study

Background

Patients with celiac disease (CeD) are at increased risk of certain viral infections and of pneumococcal pneumonia, raising concerns that they may be susceptible to severe coronavirus disease 2019 (Covid-19). We aimed to quantify the association between CeD and severe outcomes related to Covid-19.

Methods

We performed a population-based cohort study, identifying individuals with CeD in Sweden, as defined by small intestinal villus atrophy diagnosed at all (n=28) Swedish pathology departments during the years spanning 1969–2017, and alive on February 1, 2020. We compared these patients to controls matched by sex, age, county, and calendar period. We performed Cox proportional hazards with follow-up through July 31, 2020, assessing risk of 1) hospital admission with a primary diagnosis of laboratory-confirmed Covid-19 (co-primary outcome); and 2) severe disease as defined by admission to intensive care unit and/or death attributed to Covid-19 (co-primary outcome).

Results

Among patients with CeD (n=40,963) and controls (n=183,892), the risk of hospital admission for Covid-19 was 2.9 and 2.2 per 1000 person-years respectively. After adjusting for comorbidities, the risk of hospitalization for Covid-19 was not significantly increased in patients with CeD (HR 1.10; 95% CI 0.80–1.50), nor was the risk of severe Covid-19 increased (HR 0.97; 95% CI 0.59–1.59). Results were similarly null when we compared CeD patients to their non-CeD siblings with regard to these outcomes. Among all patients with CeD and controls hospitalized with a diagnosis of Covid-19 (n=58 and n=202, respectively), there was no significant difference in mortality (HR for CeD compared to controls 0.96; 95% CI 0.46–2.02).

Conclusion

In this population-based study, CeD was not associated with an increased risk of hospitalization for Covid-19 or intensive care unit and/or death attributed to Covid-19.

Long-Term Incidence and Mortality of Colorectal Cancer After Endoscopic Biopsy With Normal Mucosa: A Swedish-Matched Cohort Study

INTRODUCTION

Endoscopic screening reduces colorectal cancer (CRC) incidence and mortality. Individuals with a negative result are recommended to undergo rescreening within a 10-year interval, but evidence supporting this advice is limited.

METHODS

We performed a matched cohort study using prospectively collected data from 88,798 individuals in Sweden with normal mucosa at the first colorectal biopsy (aged ≥50 years) in the nationwide gastrointestinal epidemiology strengthened by histopathology reports (ESPRESSO) (1965-2016) and 424,150 matched reference individuals from the general population. Cox proportional hazards regression estimated multivariable hazard ratios and 95% confidence intervals (CIs) of CRC incidence and mortality of incident CRCs up to 44 years of follow-up.

RESULTS

In the normal biopsy and reference groups, respectively, the 20-year incidences of CRC were 3.03% and 4.53% and the 20-year mortalities of incident CRC were 0.89% and 1.54%. The multivariable hazard ratio comparing the normal biopsy and reference groups was 0.62 for CRC incidence (95% CI = 0.58-0.66, P < 0.001) and 0.56 for mortality of incident CRC (95% CI = 0.49-0.64, P < 0.001). When assessed by time interval after biopsy, lower CRC incidence and mortality were observed throughout the follow-up. The association seemed weaker for proximal colon cancer than for rectal and distal colon cancer.

DISCUSSION

A normal colorectal biopsy was associated with lower CRC incidence and mortality for at least 20 years after the examination. Our findings confirm previous data and suggest that the screening intervals after a normal colonoscopy could be longer than the commonly recommended 10 years. It may be time to open the discussion for a revision of the international guidelines.

Validation of the diagnosis of eosinophilic esophagitis based on histopathology reports in Sweden

BACKGROUND

Eosinophilic esophagitis (EoE) is a relatively new diagnosis, where until recently a specific international classification of disease code was missing. One way to identify patients with EoE is to use histopathology codes. We validated the clinicopathological EoE diagnosis based on histopathology reports and patient charts to establish these data sources as the basis for a nationwide EoE patient cohort.

METHODS

Through the Epidemiology Strengthened by histoPathology Reports in Sweden (ESPRESSO) study, we randomly selected 165 patients from five Swedish health care regions with a histopathologic diagnosis of EoE. Patients were assigned a histopathology diagnosis of EoE if they had ≥15 eosinophils per high-power field or, in the absence of eosinophil quantification, the pathologist interpreted the biopsy as consistent with EoE. Patient charts were scrutinized to see if the other diagnostic criteria were fulfilled. Of the 131 received patient charts, 111 (85%) had sufficient information to be included in the study.

RESULTS

Of the 111 validated patients, 99 had EoE, corresponding to a positive predictive value of 89% (95% confidence interval = 82-94%). Dysphagia was the most common symptom (n = 78, 70%), followed by food impaction (n = 64, 58%) and feeding difficulties (n = 37, 33%). Twelve patients had coexisting asthma (11%) and 16 allergic rhinitis (14%). Seventeen patients underwent esophageal dilatation (15%), of which seven had more than one dilatation. Ninety-seven (87%) patients had a proton-pump inhibitor treatment ≤2 years before or after the diagnosis. Forty-two patients (38%) had been prescribed inhalation steroids and 64 (58%) had undergone esophageal radiology.

CONCLUSION

Histopathology reports from the ESPRESSO cohort with esophageal eosinophilic inflammation are suggestive of EoE.

Mortality in Eosinophilic Esophagitis - a nationwide, population-based matched cohort study from 2005 to 2017

BACKGROUND

There is a lack of knowledge about mortality in eosinophilic esophagitis (EoE). Therefore, this study aimed to examine the mortality in EoE.

METHODS

A nationwide, population-based matched cohort study was conducted of all EoE patients in Sweden diagnosed between July 2005 and December 2017. Individuals with EoE (n = 1,625) were identified through prospectively recorded histopathology codes from all gastrointestinal pathology reports in Sweden, representing 28 pathology departments (the ESPRESSO study). Each individual with EoE was then matched with up to five reference individuals from the general population (n = 8,003) for age, sex, year of birth, and place of residence. We used the Cox proportional hazard modeling to estimate the adjusted hazard ratio (aHR) and 95% confidence interval (95% CI) while adjusting for other potential confounders. In sensitivity analyses, mortality in EoE patients was compared with mortality in their siblings.

RESULTS

Through December 2017, 34 deaths were confirmed in EoE patients (4.60 per 1,000 person-years) compared with 165 in reference individuals (4.57 per 1,000 person-years). This rate corresponds to an aHR of 0.97 (95% CI = 0.67-1.40). HRs were similar in males (aHR = 1.00 [0.66-1.51]) and females (aHR = 0.92 [0.38-2.18]). We observed no increased risk in mortality due to esophageal or other gastrointestinal cancers in patients with EoE (aHR = 1.02 [0.51-2.02]).Mortality was similar in EoE patients and their siblings (aHR = 0.91 [0.44-1.85]).

CONCLUSION

In this nationwide, population-based matched cohort study in Sweden, there was no increased risk of death in patients with EoE compared with their siblings and the general population.

Mortality in biopsy-proven alcohol-related liver disease: a population-based nationwide cohort study of 3453 patients

OBJECTIVE

Patients with alcohol-related liver disease (ALD) are at increased risk of death, but studies have rarely investigated the significance of histological severity or estimated relative risks compared with a general population. We examined mortality in a nationwide cohort of biopsy-proven ALD.

DESIGN

Population-based cohort study in Sweden comparing 3453 individuals with an International Classification of Disease (ICD) code for ALD and a liver biopsy from 1969 to 2017 with 16 535 matched general population individuals. Swedish national registers were used to ascertain overall and disease-specific mortality, starting follow-up at the latest of first ICD diagnosis or liver biopsy plus 3 months. Cox regression adjusted for relevant confounders was used to estimate HRs in ALD and histopathological subgroups.

RESULTS

Median age at diagnosis was 58 years, 65% were men and 52% had cirrhosis at baseline. Five-year cumulative mortality was 40.9% in patients with ALD compared with 5.8% in reference individuals. The risk for overall mortality was significantly increased (adjusted HR (aHR)=4.70, 95% CI 4.35 to 5.08). The risk of liver-related death was particularly high (43% of all deaths, aHR=167.6, 95% CI 101.7 to 276.3). Mortality was significantly increased also in patients with ALD without cirrhosis and was highest in the first year after baseline but persisted after ≥10 years of follow-up (aHR=2.74, 95% CI 2.37 to 3.16).

CONCLUSION

Individuals with biopsy-proven ALD have a near fivefold increased risk of death compared with the general population. Individuals with ALD without cirrhosis were also at increased risk of death, reaffirming the need to increase vigilance in the management of these individuals.

Women's Earnings are more Affected by Inflammatory Bowel Disease than Men's: A Register-Based Swedish Cohort Study

BACKGROUND AND AIMS

Patients with inflammatory bowel disease [IBD] are subject to more work disability than the general population. We aimed to estimate the monetary cost of IBD for the individual through assessment of earnings in relation to diagnosis.

METHODS

Through linkage of national registers, we identified patients aged 30-55 years at first IBD diagnosis in Sweden in 2002-2011, and same-sex IBD-free siblings. We estimated taxable earnings and disposable income from 5 years before to 5 years after diagnosis.

RESULTS

The 5961 patients [27% Crohn's disease, 68% ulcerative colitis, 4.3% IBD unclassified] had similar taxable earnings to their 7810 siblings until the year of diagnosis, when earnings decreased and remained lower than for siblings during follow-up. The adjusted difference in earnings over the entire 5-year period after diagnosis was -5% [-8212€; 95% confidence interval: -11 458 to -4967€]. The difference was greater in women than in men, and greater in Crohn's disease than in ulcerative colitis. When stratifying for sex and IBD subtype and comparing earnings during each year of follow-up, median annual earnings were lower in women with Crohn's disease and ulcerative colitis than in their sisters during all years of follow-up, whereas the men had similar annual taxable earnings to their brothers. Disposable income was similar between patients and siblings during the investigated time period.

CONCLUSION

From the year of diagnosis and at least 5 years onwards, patients with IBD had 5% lower earnings than siblings, mainly explained by differences between women with IBD and their sisters. However, there were no differences in disposable income.

Risk of Small Bowel Adenocarcinoma, Adenomas, and Carcinoids in a Nationwide Cohort of Individuals With Celiac Disease

BACKGROUND & AIMS

The incidence of small bowel cancers is increasing. Associations have been made between celiac disease (CD) and small bowel cancers, but there have been no detailed studies of large cohorts.

METHODS

Through the nationwide Epidemiology Strengthened by Histopathology Reports in Sweden cohort study, we retrieved data from Sweden's 28 pathology departments on all individuals who received a diagnosis of CD from 1965 through 2017. Individuals with CD, defined as duodenal or jejunal villous atrophy (stage 3 Marsh score), were matched with as many as 5 randomly selected reference individuals from the general population. We used stratified Cox regression to calculate hazard ratios (HRs) for small bowel adenocarcinoma, adenomas, and carcinoids.

RESULTS

During a median follow-up of 11 years, we identified 48,119 individuals with CD (patients) and 239,249 reference individuals. Beginning at 1 year after a diagnosis of CD, 29 patients (0.06%) received a diagnosis of small bowel adenocarcinoma vs 45 reference individuals (0.02%), 7 patients received a diagnosis of carcinoids vs 31 reference individuals, and 48 patients received a diagnosis of adenomas vs 50 reference individuals. Corresponding HRs were small bowel adenocarcinoma 3.05 (95% confidence interval [CI], 1.86-4.99), carcinoids 0.59 (95% CI, 0.16-2.10), and adenomas 5.73 (95% CI, 3.70-8.88). HRs were independent of sex and age. Overall, there was 1 extra case of small bowel adenocarcinoma in every 2944 patients with CD followed for 10 years. There was an inverse association between mucosal healing risk of future small bowel adenocarcinoma (HR, 0.18; 95% CI, 0.02-1.61), although the HR failed to attain statistical significance.

CONCLUSIONS

In an analysis of a nationwide pathology database in Sweden, we found the absolute risk of small bowel adenocarcinoma is low in individuals with CD. However, risks of small bowel adenocarcinoma and adenomas (but not carcinoids) are significantly increased in people with CD compared to people without this disease.

Risk of Skin Disorders in Patients with Celiac Disease: A Population-Based Cohort Study

BACKGROUND AND AIMS

Although dermatitis herpetiformis is closely associated with celiac disease (CD), data on the relationship between CD and other dermatologic disorders have been mixed. We aimed to quantify the risk of skin disorders in patients after CD diagnosis in a population-based setting.

METHODS

Using data from all 28 pathology departments in Sweden 1969-2016, we identified patients with CD. Each patient was matched by age, sex, calendar year, and geographic region to up to 5 population controls. We calculated the risk of any skin disease and specific skin diseases using Cox proportional hazards.

RESULTS

We identified 43,300 patients with CD and 198,532 matched controls. After a median follow-up time of 11.4 years, the incidences of skin disease in CD patients and controls were 22.6 and 14.8 per 1000 person-years respectively (HR=1.55; 95%CI 1.51-1.58). Increased risks were present for eczema (HR=1.67; 95%CI 1.56-1.79), psoriasis (HR=1.55; 95%CI 1.43-1.68), urticaria (HR=1.52; 95% CI 1.42-1.64), vitiligo (HR=1.90; 95%CI 1.52-2.39), acne (HR=1.39; 95%CI 1.29-1.50), and alopecia areata (HR=1.78; 95%CI 1.43-2.20).

CONCLUSIONS

Compared to the general population, patients with CD are at increased risk of multiple common skin disorders, a risk that persists in the long-term.

A Novel Method for Quantifying Intestinal Inflammatory Burden in Inflammatory Bowel Disease Using Register Data

Background

The Swedish Quality Register for Inflammatory Bowel Disease (SWIBREG) contains clinical data for the study of inflammatory bowel disease (IBD). The Epidemiology Strengthened by histoPathology Reports in Sweden (ESPRESSO) cohort was recently established for the study of gastrointestinal histopathology. We aimed to develop and validate a histology score from ESPRESSO using clinical information from SWIBREG, and secondarily, to evaluate the association of the score on IBD-related hospitalization.

Methods

In a nationwide, population-based cohort study of patients with IBD during 1969-2017, we linked endoscopic inflammation in SWIBREG with histologic inflammation in ESPRESSO. We established a clinically interpretable model for predicting the endoscopic score from histology using scalable Bayesian rule lists to define a SNOMED-based histology score applicable to the ESPRESSO cohort. We also assessed the impact of baseline endoscopic and histology scores on time to IBD-related hospitalization.

Results

We identified 5225 individuals with IBD comprising 11,051 endoscopic assessments in SWIBREG linked to a histopathology record in ESPRESSO. We created predictive models to calculate a SNOMED-based histology score which predicted the endoscopic score. Split-sample validated areas under the ROC curves for the score predicting a non-zero endoscopic score were 0.80 (0.78-0.81) in UC, 0.70 (0.68-0.72) in CD, and 0.76 (0.73-0.78) in IBD-U. In a subset of 2741 individuals with an initial IBD diagnosis and a corresponding record in ESPRESSO with an endoscopic assessment in SWIBREG, the baseline endoscopic and histology scores were associated with time to IBD-related hospitalization (endoscopy log-rank UC p<0.001, CD p=0.020, IBD-U p<0.001; histology log-rank UC p=0.018, CD p=0.960, IBD-U p=0.034).

Conclusion

Histopathology data in ESPRESSO accurately predict endoscopic scores in SWIBREG. Baseline endoscopic and histologic scores were associated with time to IBD-related hospitalization, particularly in UC. The SNOMED-based histology score can be used as a measure of disease activity in future register-based IBD studies.

Mortality of Patients With Microscopic Colitis in Sweden

BACKGROUND & AIMS

Microscopic colitis is one of the most common causes of chronic diarrhea in older populations. We investigated all-cause and cause-specific mortality in patients with microscopic colitis.

METHODS

We conducted a nationwide cohort study of all cases of microscopic colitis (n = 14,333) diagnosed from 1990 through 2017 in Sweden. Cases of microscopic colitis were identified using SNOMED codes from gastrointestinal histopathology reports collected from Sweden's 28 pathology departments. Each case of microscopic colitis was matched to 5 population comparators (n = 68,700). Mortality data were ascertained from Sweden's cause of death register. We used Cox proportional hazards modeling to estimate adjusted hazard ratios (aHRs) and 95% CIs.

RESULTS

Through December of 2017, we confirmed 3014 deaths in patients with microscopic colitis (27.4/1000 person-years) and 12,534 deaths in matched population comparators (23.3/1000 person-years). This corresponded to a 10-year absolute risk difference of 3.4% (95% CI, 2.1%-4.6%) and an aHR of 1.17 (95% CI, 1.12-1.22). However, further adjustment of models for comorbidity burden reduced the relative risk of death for patients with microscopic colitis (aHR, 0.98; 95% CI, 0.94-1.02). In analyses of cause-specific death, microscopic colitis was associated with an increased risk of gastrointestinal-related death (aHR, 1.68; 95% CI, 1.38-2.05) and infection-related death (aHR, 1.42 ; 95% CI, 1.11-1.83), but not cancer-related death (aHR, 0.83; 95% CI, 0.76-0.91) or cardiovascular-related death (aHR, 1.02; 95% CI, 0.96-1.10).

CONCLUSIONS

In a nationwide cohort study in Sweden, we found that patients with microscopic colitis were at increased risk of death. However, the increase appears to be related to higher burden of comorbidities in this population.

Antibiotic use and the development of inflammatory bowel disease: a national case-control study in Sweden

BACKGROUND

Use of antibiotics in early life has been linked with childhood inflammatory bowel disease (IBD), but data for adults are mixed, and based on smaller investigations that did not compare risk among siblings with shared genetic or environmental risk factors. We aimed to investigate the association between antibiotic therapy and IBD in a large, population-based study.

METHODS

In this prospective case-control study, we identified people living in Sweden aged 16 years or older, with a diagnosis of IBD based on histology and at least one diagnosis code for IBD or its subtypes (ulcerative colitis and Crohn's disease). We identified consecutive patients with incident IBD from the ESPRESSO (Epidemiology Strengthened by histoPathology Reports in Sweden) study, cross-referenced with the Swedish Patient Register and the Prescribed Drug Register. We accrued data for cumulative antibiotic dispensations until 1 year before time of matching for patients and up to five general population controls per patient (matched on the basis of age, sex, county, and calendar year). We also included unaffected full siblings as a secondary control group. Conditional logistic regression was used to estimate multivariable-adjusted odds ratios (aORs) and 95% CIs for diagnosis of incident IBD.

FINDINGS

We identified 23 982 new patients with IBD (15 951 ulcerative colitis, 7898 Crohn's disease, 133 unclassified IBD) diagnosed between Jan 1, 2007, and Dec 31, 2016. 117 827 matched controls and 28 732 siblings were also identified. After adjusting for several risk factors, aOR in patients who had used antibiotics versus those who had never used antibiotics was 1·88 (95% CI 1·79-1·98) for diagnosis of incident IBD, 1·74 (1·64-1·85) for ulcerative colitis, and 2·27 (2·06-2·49) for Crohn's disease. aOR was higher in patients who had received one antibiotic dispensation (1·11, 1·07-1·15), two antibiotic dispensations (1·38, 1·32-1·44), and three or more antibiotic dispensations (1·55, 1·49-1·61) than patients who had none. Increased risk was noted for ulcerative colitis (aOR with three or more antibiotic dispensations 1·47, 95% CI 1·40-1·54) and Crohn's disease (1·64, 1·53-1·76) with higher estimates corresponding to broad-spectrum antibiotics. Similar but attenuated results were observed when siblings were used as the reference group, with an aOR of 1·35 (95% CI 1·28-1·43) for patients who had received three or more dispensations, compared with general population controls.

INTERPRETATION

Higher cumulative exposure to systemic antibiotic therapy, particularly treatments with greater spectrum of microbial coverage, may be associated with a greater risk of new-onset IBD and its subtypes. The association between antimicrobial treatment and IBD did not appear to differ when predisposed siblings were used as the reference controls. Our findings, if substantiated by longer-term prospective studies in humans or mechanistic preclinical investigations, suggest the need to further emphasise antibiotic stewardship to prevent the rise in dysbiosis-related chronic diseases, including IBD.

FUNDING

National Institutes of Health. Crohn's and Colitis Foundation.

Costs and Use of Health Care in Patients With Celiac Disease: A Population-Based Longitudinal Study

INTRODUCTION

Celiac disease (CD) affects 1% of the population. Its effect on healthcare cost, however, is barely understood. We estimated healthcare use and cost in CD, including their temporal relationship to diagnosis.

METHODS

Through biopsy reports from Sweden's 28 pathology departments, we identified 40,951 prevalent patients with CD (villous atrophy) as of January 1, 2015, and 15,086 incident patients with CD diagnosed in 2008-2015, including 2,663 who underwent a follow-up biopsy to document mucosal healing. Each patient was compared with age- and sex-matched general population comparators (n = 187,542). Using nationwide health registers, we retrieved data on all inpatient and nonprimary outpatient care, prescribed diets, and drugs.

RESULTS

Compared with comparators, healthcare costs in 2015 were, on average, $1,075 (95% confidence interval, $864-1,278) higher in prevalent patients with CD aged <18 years, $715 ($632-803) in ages 18-64 years, and $1,010 ($799-1,230) in ages ≥65 years. Half of all costs were attributed to 5% of the prevalent patients. Annual healthcare costs were $391 higher 5 years before diagnosis and increased until 1 year after diagnosis; costs then declined but remained 75% higher than those of comparators 5 years postdiagnosis (annual difference = $1,044). Although hospitalizations, nonprimary outpatient visits, and medication use were all more common with CD, excess costs were largely unrelated to the prescription of gluten-free staples and follow-up visits for CD. Mucosal healing in CD did not reduce the healthcare costs.

DISCUSSION

The use and costs of health care are increased in CD, not only before, but for years after diagnosis. Mucosal healing does not seem to lower the healthcare costs.

Microscopic colitis and risk of cancer - a population-based cohort study

BACKGROUND AND AIMS

The association between microscopic colitis (MC) and cancer risk is unclear. Large, population-based studies are lacking.

METHODS

We conducted a nationwide cohort study of 11,758 patients with incident MC (diagnosed 1990-2016 in Sweden), 50,828 matched reference individuals and 11,614 siblings to MC patients. Data were obtained through Sweden´s pathology departments and from the Swedish Cancer Register. Adjusted hazard ratios (aHRs) were calculated using Cox proportional hazards models.

RESULTS

At the end of follow up (mean: 6.7 years), 1,239 (10.5%) of MC patients had received a cancer diagnosis, compared to 4,815 (9.5%) of reference individuals (aHR 1.08 (95%CI=1.02-1.16)). The risk of cancer was highest during the first year of follow up. The absolute excess risks for cancer at 5, 10 and 20 years after MC diagnosis were +1.0% (95%CI=0.4%-1.6%), +1.5% (0.4%-2.6%) and +3.7% (-2.3-9.6%), respectively, equivalent to one extra cancer event in every 55 individuals with MC followed for ten years.MC was associated with an increased risk of lymphoma (aHR 1.43, 1.06-1.92) and lung cancer (aHR 1.32, 1.04-1.68) but with decreased risks of colorectal (aHR 0.52, 0.40-0.66) and gastrointestinal cancers (aHR 0.72, 0.60-0.85). We found no association with breast or bladder cancer. Using siblings as reference group to minimize the impact of shared genetic and early environmental factors, patients with MC were still at an increased risk of cancer (HR=1.20; 95%CI=1.06-1.36).

CONCLUSIONS

This nationwide cohort study demonstrated an 8% increased risk of cancer in MC patients. The risk was highest during the first year of follow up.

Inflammatory bowel disease and pancreatic cancer: a Scandinavian register-based cohort study 1969-2017

BACKGROUND

Patients with inflammatory bowel disease (IBD) have an increased risk of cancer.

AIM

To assess the risk of pancreatic cancer in IBD compared to the general population.

METHODS

Patients with incident IBD 1969-2017 were identified in Danish and Swedish National Patient Registers and through biopsy data, and were matched to IBD-free reference individuals by sex, age, place of residence and year of IBD diagnosis. We linked data to Cancer and Causes of Death Registers and examined the absolute and relative risks of pancreatic cancer and pancreatic cancer death.

RESULTS

Among 161 926 patients followed for 2 000 951 person years, 442 (0.27%) were diagnosed with pancreatic cancer compared to 3386 (0.21%) of the 1 599 024 reference individuals. The 20-year cumulative incidence was 0.34% (95% confidence interval 0.30-0.38) vs 0.29% (0.28-0.30). The incidence rate was 22.1 (20.1-24.2)/100 000 person years in the patients (excluding the first year of follow-up: 20.8 [18.8-23.0]), and 16.6 (16.0-17.2) in the reference individuals. The hazard ratio (HR) for pancreatic cancer was increased overall: 1.43 (1.30-1.58), in subtypes (Crohn's disease: 1.44 [1.18-1.74]; ulcerative colitis: 1.35 [1.19-1.53]; IBD unclassified: 1.99 [1.50-2.64]) and especially in IBD patients with primary sclerosing cholangitis: 7.55 (4.94-11.5). Patients and reference individuals with pancreatic cancer did not differ in cancer stage (P = 0.17) or pancreatic cancer mortality (HR 1.07 [0.95-1.21]).

CONCLUSIONS

Patients with IBD had an excess risk of pancreatic cancer, in particular patients with primary sclerosing cholangitis. However, the cumulative incidence difference after 20 years was small: 0.05%, that is, one extra pancreatic cancer per 2000 IBD patients.

Mortality Risk in Irritable Bowel Syndrome: Results From a Nationwide Prospective Cohort Study

INTRODUCTION

Mortality concern is a frequent driver of care seeking in patients with irritable bowel syndrome (IBS). Data on mortality in IBS are scarce, and population-based studies have been limited in size. We examined mortality in IBS.

METHODS

A nationwide, matched, population-based cohort study was conducted in Sweden. We identified 45,524 patients undergoing a colorectal biopsy at any of Sweden's 28 pathology departments and with a diagnosis of IBS from 2002 to 2016 according to the National Patient Register, a nationwide registry of inpatient and outpatient specialty care. We compared the mortality risk between these individuals with IBS and age- and sex-matched reference individuals (n = 217,316) from the general population and siblings (n = 53,228). In separate analyses, we examined the role of mucosal appearance for mortality in IBS. Finally, we examined mortality in 41,427 patients with IBS not undergoing a colorectal biopsy. Cox regression estimated hazard ratios (HRs) for death.

RESULTS

During follow-up, there were 3,290 deaths in individuals with IBS (9.4/1,000 person-years) compared with 13,255 deaths in reference individuals (7.9/1,000 person-years), resulting in an HR of 1.10 (95% confidence interval [CI] = 1.05-1.14). After adjustment for confounders, IBS was not linked to mortality (HR = 0.96; 95% CI = 0.92-1.00). The risk estimates were neutral when patients with IBS were compared with their siblings. The underlying mucosal appearance on biopsy had only a marginal impact on mortality, and patients with IBS not undergoing a colorectal biopsy were at no increased risk of death (HR = 1.02; 95% CI = 0.99-1.06).

DISCUSSION

IBS does not seem to confer an increased risk of death.

Microscopic Colitis and Risk of Inflammatory Bowel Disease in a Nationwide Cohort Study

BACKGROUND & AIMS

Microscopic colitis shares pathogenetic mechanisms with inflammatory bowel disease (IBD). We studied the association between microscopic colitis and risk of incident IBD using data from a nationwide cohort study.

METHODS

We conducted a prospective cohort study of all adults who received a diagnosis of microscopic colitis from 1990 through 2017 in Sweden and risk of incident IBD. Cases of microscopic colitis (n= 13,957) were identified through Systematized Nomenclature of Medicine codes from the ESPRESSO (Epidemiology Strengthened by histoPathology Reports in Sweden) study, which included gastrointestinal pathology reports from all of Sweden's 28 centers. Individuals with microscopic colitis were matched to 5 general population controls (n = 66,820) and to unaffected siblings (n =13,943). Cox regression was used to estimate adjusted hazard ratio (aHRs) and 95% confidence intervals (CIs).

RESULTS

Through December of 2017, we identified 323 incident cases of ulcerative colitis (UC) and 108 incident cases of Crohn's disease (CD) in patients with microscopic colitis compared with 94 UC and 42 CD cases in population comparators. Mean times from diagnosis of microscopic colitis to diagnosis of CD was 3.3 ± 3.2 years and to diagnosis of UC was 3.2 ± 3.5 years. In multivariable models, microscopic colitis was associated with an aHR of 12.6 (95% CI 8.8-18.1) for CD, 17.3 (95% CI 13.7-21.8) for UC, and 16.8 (95% CI 13.9-20.3) for IBD. The 10-year absolute excess risks of CD and UC were 0.9 (95% CI 0.7-1.1) and 2.6 (95% CI 2.2-2.9) percentage points, respectively. In sensitivity analyses, comparing patients with microscopic colitis with their unaffected siblings, the aHRs of CD and UC were 5.4 (95% CI 3.2-9.2) and 9.4 (95% CI 6.4-13.8), respectively.

CONCLUSIONS

In a population-based study in Sweden, we found a significant increase in risk of incident IBD among patients with microscopic colitis. Future studies should focus on potential mechanisms underlying these observed associations.

Association Between Celiac Disease and Mortality Risk in a Swedish Population

IMPORTANCE

Celiac disease may be associated with a modest but persistent increased long-term mortality risk. It is uncertain whether this risk has changed in the era of wider diagnosis rates, less severe clinical disease, and more widespread availability of gluten-free food.

OBJECTIVE

To evaluate the association between celiac disease and mortality risk in a population-based cohort in Sweden.

DESIGN, SETTING, AND PARTICIPANTS

All individuals in Sweden with celiac disease diagnosed between 1969 and 2017 were identified through the Epidemiology Strengthened by histoPathology Reports in Sweden (ESPRESSO) cohort. Participants (n = 49 829) were observed starting on the day of the biopsy. The final date of follow-up was December 31, 2017.

EXPOSURES

Celiac disease was defined by the presence of small intestinal villus atrophy on histopathology specimens during the years 1969-2017 from Sweden's 28 pathology departments. Each individual was matched with as many as 5 control participants in the general population by age, sex, county, and calendar period.

MAIN OUTCOMES AND MEASURES

The primary outcome was all-cause mortality, and the secondary outcome was cause-specific mortality. Patients with celiac disease were compared with controls using stratified Cox proportional modeling, stratifying by year of diagnosis.

RESULTS

There were 49 829 patients with celiac disease, including 24% who were diagnosed between the years 2010 and 2017. The mean (SD) age at diagnosis was 32.2 (25.2) years and 62.4% were women. During a median follow-up time of 12.5 years, 13.2% (n = 6596) died. Compared with controls (n = 246 426), overall mortality was increased in those with celiac disease (9.7 vs 8.6 deaths per 1000 person-years; absolute difference, 1.2 per 1000 person-years; hazard ratio [HR], 1.21 [95% CI, 1.17-1.25]). The relative increase in mortality risk was present in all age groups and was greatest in those diagnosed in the age range of 18 to 39 years (1.9 vs 1.1 per 1000 person-years; HR, 1.69 [95% CI, 1.47-1.94]; P values for heterogeneity comparing 18-39 years with 40-59 years and with ≥60 years were both <.001). Individuals with celiac disease were at increased risk of death from cardiovascular disease (3.5 vs 3.4 per 1000 person-years; HR, 1.08 [95% CI, 1.02-1.13]), cancer (2.7 vs 2.2 per 1000 person-years; HR, 1.29 [95% CI, 1.22-1.36]), and respiratory disease (0.6 vs 0.5 per 1000 person-years; HR, 1.21 [95% CI, 1.08-1.37]). When compared with controls, the overall mortality risk was greatest in the first year after diagnosis (15.3 vs 6.5 per 1000 person-years; HR, 2.34 [95% CI, 2.14-2.55]) but persisted beyond 10 years after diagnosis (10.5 vs 10.1 per 1000 person-years; HR, 1.15 [95% CI, 1.10-1.20]). The mortality risk was likewise present for patients diagnosed during the years 2010-2017 (7.5 vs 5.5 per 1000 person-years; HR, 1.35 [95% CI, 1.21-1.51]).

CONCLUSIONS AND RELEVANCE

In a Swedish population studied between 1969 and 2017, a diagnosis of celiac disease compared with the general population was associated with a small but statistically significant increased mortality risk.

Risk of colorectal cancer incidence and mortality after polypectomy: a Swedish record-linkage study

BACKGROUND

Long-term colorectal cancer incidence and mortality after colorectal polyp removal remains unclear. We aimed to assess colorectal cancer incidence and mortality in individuals with removal of different histological subtypes of polyps relative to the general population.

METHODS

We did a matched cohort study through prospective record linkage in Sweden in patients aged at least 18 years with a first diagnosis of colorectal polyps in the nationwide gastrointestinal ESPRESSO histopathology cohort (1993-2016). For each polyp case, we identified up to five matched reference individuals from the Total Population Register on the basis of birth year, age, sex, calendar year of biopsy, and county of residence. We excluded patients and reference individuals with a diagnosis of colorectal cancer either before or within the first 6 months after diagnosis of the index polyp. Polyps were classified by morphology codes into hyperplastic polyps, sessile serrated polyps, tubular adenomas, tubulovillous adenomas, and villous adenomas. Colorectal cancer cases were identified from the Swedish Cancer Registry, and cause-of-death data were retrieved from the Cause of Death Register. We collected information about the use of endoscopic examination before and after the index biopsy from the Swedish National Patient Registry, and counted the number of endoscopies done before and after the index biopsies. We calculated cumulative risk of colorectal cancer incidence and mortality at 3, 5, 10, and 15 years, and computed hazard ratios (HRs) and 95% CIs for colorectal cancer incidence and mortality using a stratified Cox proportional hazards model within each of the matched pairs.

FINDINGS

178 377 patients with colorectal polyps and 864 831 matched reference individuals from the general population were included in our study. The mean age of patients at polyp diagnosis was 58·6 (SD 13·9) years for hyperplastic polyps, 59·7 (14·2) years for sessile serrated polyps, 63·9 (12·9) years for tubular adenomas, 67·1 (12·1) years for tubulovillous adenomas, and 68·9 (11·8) years for villous adenomas. During a median of 6·6 years (IQR 3·0-11·6) of follow-up, we documented 4278 incident colorectal cancers and 1269 colorectal cancer-related deaths in patients with a polyp, and 14 350 incident colorectal cancers and 5242 colorectal cancer deaths in general reference individuals. The 10-year cumulative incidence of colorectal cancer was 1·6% (95% CI 1·5-1·7) for hyperplastic polyps, 2·5% (1·9-3·3) for sessile serrated polyps, 2·7% (2·5-2·9) for tubular adenomas, 5·1% (4·8-5·4) for tubulovillous adenomas, and 8·6% (7·4-10·1) for villous adenomas compared with 2·1% (2·0-2·1) in reference individuals. Compared with reference individuals, patients with any polyps had an increased risk of colorectal cancer, with multivariable HR of 1·11 (95% CI 1·02-1·22) for hyperplastic polyps, 1·77 (1·34-2·34) for sessile serrated polyps, 1·41 (1·30-1·52) for tubular adenomas, 2·56 (2·36-2·78) for tubulovillous adenomas, and 3·82 (3·07-4·76) for villous adenomas (p<0·05 for all polyp subtypes). There was a higher proportion of incident proximal colon cancer in patients with serrated (hyperplastic and sessile) polyps (52-57%) than in those with conventional (tubular, tubulovillous, and villous) adenomas (30-46%). For colorectal cancer mortality, a positive association was found for sessile serrated polyps (HR 1·74, 95% CI 1·08-2·79), tubulovillous adenomas (1·95, 1·69-2·24), and villous adenomas (3·45, 2·40-4·95), but not for hyperplastic polyps (0·90, 0·76-1·06) or tubular adenomas (0·97, 0·84-1·12).

INTERPRETATION

In a largely screening-naive population, compared with individuals from the general population, patients with any polyps had a higher colorectal cancer incidence, and those with sessile serrated polyps, tubulovillous adenomas, and villous adenomas had a higher colorectal cancer mortality.

FUNDING

US National Institutes of Health, American Cancer Society, American Gastroenterological Association, Union for International Cancer Control.

Validation of serrated polyps (SPs) in Swedish pathology registers

Background

Little is known about the natural history of serrated polyps (SPs), partly due to the lack of large-scale epidemiologic data. In this study, we examined the validity of SP identification according to SNOMED (Systematised Nomenclature of Medicine) codes and free text from colorectal histopathology reports.

Methods

Through the ESPRESSO (Epidemiology Strengthened by histoPathology Reports in Sweden) study, we retrieved data on SPs from all pathology departments in Sweden in 2015–2017 by using SNOMED codes and free-text search in colorectal histopathology reports. Randomly selected individuals with a histopathology report of SPs were validated against patient charts using a structured, retrospective review.

Results

SPs were confirmed in 101/106 individuals with a histopathology report of SPs, yielding a positive predictive value (PPV) of 95% (95%CI = 89–98%). By year of diagnosis, the PPV was 89% (95%CI = 69–97%), 96% (95%CI = 81–99%) and 97% (95%CI = 89–99%) for individuals diagnosed before 2001 (n = 19), between 2001 and 2010 (n = 26) and after 2010 (n = 61), respectively. According to search method, the PPV for individuals identified by SNOMED codes was 100% (95%CI = 93–100%), and 93% (95%CI = 86–97%) using free-text search. Recorded location (colon vs. rectum) was correct in 94% of all SP histopathology reports (95%CI = 84–98%) identified by SNOMED codes. Individuals with SPs were classified into hyperplastic polyps (n = 34; 32%), traditional serrated adenomas (n = 3; 3%), sessile serrated adenomas/polyps (SSA/Ps) (n = 70; 66%), unspecified SPs (n = 3, 3%), and false positive SPs (n = 5, 5%). For individuals identified by SNOMED codes, SSA/Ps were confirmed in 49/52 individuals, resulting in a PPV of 94% (95%CI: 84–98%). In total, 57% had ≥2 polyps (1: n = 44, 2–3: n = 33 and ≥ 4: n = 27). Some 46% of SPs (n = 71) originated from the proximal colon and 24% were ≥ 10 mm in size (n = 37). Heredity for colorectal cancer, intestinal polyposis syndromes, or both was reported in seven individuals (7%). Common comorbidities included diverticulosis (n = 45, 42%), colorectal cancer (n = 19, 18%), and inflammatory bowel disease (n = 10, 9%).

Conclusion

Colorectal histopathology reports are a reliable data source to identify individuals with SPs.

A nationwide cohort study of the incidence of microscopic colitis in Sweden

BACKGROUND

Epidemiological studies of microscopic colitis have shown varying but increasing incidence rates.

AIM

To assess the incidence of microscopic colitis in Sweden.

METHODS

Nationwide cohort study performed in 1995-2015 based on biopsy reports. Age-specific and age-standardised incidence rates were calculated.

RESULTS

We identified 13 844 patients with an incident diagnosis of microscopic colitis. Lymphocytic colitis (n = 9238) constituted 67% and collagenous colitis (n = 4606) 33% of microscopic colitis. The mean age at time of diagnosis of microscopic colitis was 60.2 years (58.6 for lymphocytic colitis, 63.3 for collagenous colitis). The lifetime risk of developing microscopic colitis was 0.87% in women (95% confidence interval, CI: 0.85-0.88) and 0.35% in men (95% CI: 0.34-0.36). From 2006, the overall incidence of microscopic colitis was approximately 10.5 cases per 100 000 person-years (95% CI: 9.8-11.3) with higher rates in women (72% of cases, incidence rate ratio = 2.4 (95% CI: 2.3-2.5) and the elderly with increasing rates up to 75-79 years. From 2006-2015, there was a significant increase of 1% per year (P = 0.02) in the overall microscopic colitis incidence rate in women; the estimated annual percent change was similar, although not statistically significant, in men (P = 0.15).

CONCLUSIONS

In Sweden, the incidence of microscopic colitis is still increasing in women, although the rate appears to be stabilising. The incidence is particularly high in women and the elderly up to age 75-79 years. Finally, across a lifetime, 1 in 115 females and 1 in 286 males are expected to be diagnosed with microscopic colitis and thus posing a considerable disease burden.