Sarcoidosis and increased risk of comorbidities and mortality in sweden.

Sarcoidosis is a systemic inflammatory disorder involving granuloma formation that, in more than 90% of cases, affects the lungs and intrathoracic lymph nodes. Of all patients, 30-50% show extra pulmonary disease affecting, most commonly, the skin, liver, eyes and peripheral lymph nodes. Sarcoidosis is diagnosed by biopsy of the affected organs. The etiology is unclear, but a genetic susceptibility to yet unidentified environmental factors is suggested ( 1 ). Most cases occur sporadically, but some are familial, and first-degree relatives of patients are at a higher risk of developing sarcoidosis than others ( 2 ). The annual incidence of sarcoidosis in Sweden is 11.5 per 100 000, among the highest worldwide ( 3 ). Sarcoidosis occurs in an acute or chronic form, each with different clinical presentation and prognosis ( 1 ). Most deaths due to sarcoidosis are caused by advanced pulmonary fibrosis or involvement of the central nervous system or heart ( 4 ). Studies suggest increased coexistence of several immune-mediated and chronic inflammatory diseases with sarcoidosis ( 5 , 6 ). Associations with hypertonia, chronic obstructive pulmonary disease (COPD), thyroid diseases, and diabetes mellitus have also been identified ( 5 - 10 ). Whether the risk of malignancy in sarcoidosis is increased or not has been debated. A meta-analysis including 16 original studies and more than 25,000 patients suggested a moderate, though significant, association between sarcoidosis and malignancy overall ( 11 ). Similar results were obtained in a study of hospitalized patients with sarcoidosis in Sweden ( 12 ). Previous studies on mortality in sarcoidosis are largely based on data from death certificates ( 13 - 15 ). Inclusion in these mortality studies requires correct identification of a sarcoidosis case and declaration of sarcoidosis as cause of death on the death certificate. Since sarcoidosis can be a chronic disease affecting several organs there is a risk of misclassification and it may not by declared as the cause of death in patients with a clinically concealed disease. Moreover, patients who had sarcoidosis earlier in life that improved spontaneously would be excluded. Mortality studies also have contradictory results, with some suggesting that mortality is up to twice as high among those with sarcoidosis compared to the general population while others propose that mortality is not increased in patients with sarcoidosis ( 16 - 18 ). Considering this, this study aimed to contribute with more information on the comorbidities and risk of mortality of sarcoidosis compared to matched controls and the Swedish general population.

The study population consisted of 23,484 individuals in total, 7828 cases and 15,656 controls. The sex distribution was 58.9% males, 41.1% females. Deaths occurred in a total of 451 individuals, of which 218 (2.8%) were among the cases and 233 (1.5%) were among the controls. Further demographics are presented in Table 1 . Characteristics of Cases and Controls included in the Study. a Controls are matched 2:1 on age, sex and county at the time of diagnosis The number of inpatient care visits by sarcoid patients, compared to controls, were significantly increased and were associated with diagnoses in most ICD10 chapters, as shown in Table 2 . The most apparent increase among sarcoid patients were found in diseases of the blood and immune-mechanism (chapter III), HR 66.48 (95% CI 36.48 – 121.12), which most likely is due to sarcoidosis itself. Numbers of Visits to Inpatient Care where Sarcoid Cases and Matched Controls are compared using Cox Regression. a Each individual can appear several times in different chapters, b Hazard ratio, c 95% confidence interval, d Diseases of the blood and blood-forming organs and certain disorders involving the immune-mechanism, e Symptoms, signs and abnormal clinical and laboratory findings not elsewhere classified, f Factors influencing health status and contact with health services Apart from diseases of the blood and immune-mechanism (chapter III), the HRs were highest for: diseases of the eye and adnexa, HR 2.95 (95% CI 1.88 – 4.62); respiratory system, HR 2.76 (95% CI 2.22 – 3.45); and certain infectious and parasitic diseases, HR 2.49 (95% CI 1.84 – 3.38). Further significant increases in associated comorbidities stratified into ICD10 chapters are presented in Table 2 . Further information about specific diseases are found in Supplementary Table 1 and 2 . In respiratory system several diagnoses has a higher rate than controls. There is an significant increase in pneumothorax (J93) HR 11.53 (95% CI 3.99 - 33.34), influenza and pneumonia (J09-J18) HR 2.98 (CI 2.41 – 3.68), chronic lower respiratory diseases (J40-J47) HR 2.53 (95% CI 1.66 – 3.85) and other respiratory diseases principally affecting the interstitium (J80-J84) HR 2.98 (95% CI 5.26 – 57.42) (see Supplement Files 2 ). Compering cases and controls for diseases of the circulatory system there is an increased HR 1.40 (95% CI 1.21 – 1.63) for the whole chapter. Subgrouping in diagnosis yield pulmonary embolism (I26) HR 4.36 (95% CI 2.26 – 7.07), cardiomyopathy (I42) HR 5.13 (95% CI 2.37 – 11.08), paroxysmal tachycardia (I47) HR 3.27 (95% CI 2.28 – 4.68) and heart failure (I50) HR 3.27 (95% CI 2.28 – 4.68) and where in high rate and significant increased. For cancer, there was an increase in malignant neoplasms of ill-defined, secondary and unspecified sites (C76−C80) HR 2.38 (95% CI 1.59 - 3.57) and malignant neoplasms of lymphoid, hematopoietic and related tissue (C81-C96) with HR 2.94 (95% CI 1.85 - 4.68). A small but significant decrease in association of comorbidities in sarcoid cases, was also seen for mental and behavioral diseases (chapter V). However, when cases and controls were stratified by sex, the results are no longer significant. When comorbidities of the sarcoid cases and control groups are independently compared to those of the general population of Sweden, it appears that the sarcoid cases contacted the inpatient healthcare clinic significantly more frequently than the general population ( Figure 2 ). The controls, however, contacted the inpatient clinic less frequently than the general population. Comparison of Comorbidities between Cases/Controls and the General Population. Every unique personal identification number was counted once a year, per chapter and county. Death occurred for 218 cases (2.8%) and for 233 (1.5%) controls. Mortality HR among sarcoid cases were 5.5 (95% CI 4.8 – 6.3) per 1000 person-years, and 2.9 (95% CI 2.6 - 3.3) per 1000 person-years in matched controls. The overall HR for mortality was 1.88 (95% CI 1.56 – 2.26) as seen in Table 3 . Mortality caused by neoplasms (chapter II) was significantly increased for sarcoid cases in both males and females. Increased mortality in males, but not in females, was seen for diseases of the circulatory system (chapter IX), respiratory system (chapter X) and external causes of mortality (chapter XX). Increased mortality HR was also seen for diseases of the blood and immune-mechanisms (chapter III) compared to controls, but this increase is probably due to the sarcoidosis-diagnosis itself. Deaths in Sarcoid Cases and Matched Controls Compared using Cox Regression. a Only chapters where at least 5 deaths are seen in either cases or the controls are reported, b Hazard ratio, c 95% confidence interval, d Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism, e Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified External causes of mortality (chapter XX) includes accidents and intentional death. Of the sarcoid cases, 17 deaths were due to accidents and 10 were intentional. Of the controls, 12 died due to accidents, and 12 deaths were intentional. Table 3 presents only the ICD10 chapters where at least five deaths among cases or controls are declared. Seven deaths among the cases and six deaths among the controls were therefore not included in the table but are included in the analysis of total mortality. Total overall mortality for sarcoid cases was greater than the general population, SMR 1.75 (95% CI 1.52-2.00). There was no significant difference in mortality between controls and the general population. Significantly more deaths were observed for sarcoid cases than for the general population due to neoplasms (chapters II), diseases of the respiratory system (chapter X) and external causes of mortality (chapter XX) as illustrated in Figure 3 . These findings differ from the results of the comparison between sarcoid cases and matched controls, which demonstrated a small increase in mortality caused by cardiovascular diseases, HR 1.64 (95% CI 1.08 – 2.48). Comparison of Mortality between Cases/Controls and the General Population. Every unique personal identification number was counted only once a year, per chapter and county. Only chapters with 5 or more deaths among cases or controls are included.

In this longitudinal case-control study, the aim was to examine comorbidities and mortality in sarcoid patients in Sweden during the years 2007 – 2016. Increases in overall comorbidity and overall mortality in sarcoid patients compared to matched controls as well as the general population was observed during the observation period. Sarcoid patients showed increased overall comorbidity compared to controls. Sarcoidosis is a systemic disease that manifests in several organs. Each organ manifestation has its own code in the ICD10, and these codes are found across several different chapters. Examples of codes include: G35-Cranial nerve disorders-, H22-Disorders of iris and ciliary body- and I41-Myocarditis in diseases classified elsewhere. In this study, observations of comorbidities are presented at the ICD10 chapter level instead of analyzing individual diagnoses. Therefore, it is not possible to specify which diagnoses are causing the increase. As mentioned previously, a large meta-analysis suggested that the risk of malignancies increases slightly with sarcoidosis ( 11 ). In this study, malignancy was one of the most common reasons patients needed inpatient care, HR 1.78 (95% CI 1.50 – 2.10), which support this observation. Only cardiovascular diseases led to as many, and slightly more, inpatient visits, but for these comorbidities, the difference between cases and controls was not as large, HR 1.40 (95% CI 1.21 – 1.63). This study reveals increased mortality in sarcoid cases compared to controls, with death occurring in 2.8% of sarcoid cases and 1.5% of controls, HR 1.88 (95% CI 1.56–2.26). In previous studies, mortality rates have ranged between 9.4-14 per 1000 person-years (13, 14, 16). A recently-published Swedish study on mortality in sarcoid patients aged 18-85 years showed a mortality rate of 11 per 1000 person-years, while controls had a mortality rate of 6.7 per 1000 person-years ( 19 ). This study showed mortality rates of 5.5 (95% CI 4.8 – 6.3) per 1000 person-years among sarcoid cases and 2.9 (95% CI 2.6- 3.3) per 1000 person-years in matched controls. The lower mortality rate observed in this study might be due to the younger study population (aged 20-65 years), or to the somewhat healthier populations since neither the cases nor the controls suffered from ankylosing spondylitis, rheumatoid arthritis, Crohn’s disease or ulcerative colitis. The ICD10 chapter associated with the highest increased risk of death was diseases of the blood and immune-mechanism (chapter III), where the code for sarcoidosis diagnosis is found. Previously, sarcoidosis was reported to account for 24.7-58.8% of deaths occurring among sarcoid patients ( 14 , 17 ). In this study, sarcoidosis was considered the cause of death in 10.6% of deaths among sarcoid patients. In Scandinavia, one third of sarcoid patients is affected by the acute form, Lofgren’s syndrome, rather than the chronic form. Lofgren’s syndrome is said to have a better prognosis. The cause of death among cases evaluated in this study may be affected by low reporting of sarcoidosis as the cause of death among patients, since patients may have suffered from Lofgren’s syndrome earlier in life and were considered recovered at the time of death. Another impact to mortality, as evaluated herein, is that the downstream consequences of the patients’ sarcoidosis were coded as the cause of death rather than the sarcoidosis diagnosis itself. The ICD10 chapters associated with the highest number of deaths among sarcoid cases was neoplasms (90 deaths, 41.3%), cardiovascular diseases (40 deaths, 18.3%), external causes of mortality (27 deaths, 12.4%) and respiratory diseases (12 deaths, 5.5%). A study investing sarcoidosis-related morality in France, cardiovascular and respiratory diseases accounted for 16.8% and 7.5% of underlying causes of death while sarcoidosis was listed as a non-underlying cause of death, which is similar to the findings of this study ( 13 ). However, neoplasms accounted for 24.7% of the underlying cause of deaths which is clearly lower than the findings in this study ( 13 ). Suicide is a cause of death included in external causes of mortality (chapter XX) and, in males, this chapter was associated with a significant increase in death compared to controls, HR 2.32 (95% CI 1.25 – 4.98). It has previously been reported that the prevalence of depression in patients with sarcoidosis may be up to 60% when a standardized protocol is used to identify depression ( 20 ). However, this study show no increased number of visits to inpatient care for mental and behavioural diseases (chapter V). One explanation for this finding might be that depression was under-diagnosed among the study cases. Another, possibly more likely explanation may be that mild to moderate depression is usually treated in primary care, but this study used data from the inpatient register only. One limitation of this study was that it is not possible stratify sarcoid cases by level of severity of sarcoidosis or type of sarcoidosis (acute or chronic form). Another limitation, as this is a register study, there was no information on potential confounders such as smoking habits, however cases and controls are matched based on age, sex and geographical area and therefore one could assume the distribution of smokers and non-smokers to be similar in both groups. Furthermore, there was no information about medications, which also could be a confounder, as medications used to treat sarcoidosis can cause varying side effects. For instance Glucocorticoids, the first-line therapy, have numerous side effects that affect several different organ systems ( 21 ). The strengths of this study is study is the large study population comprising 7828 cases and 15,656 matched controls, and cases and controls were also both compared to the general population. Moreover, Sweden maintains high-quality registers that cover the entire population, together with unique personal identification numbers that can link patient data across different nationwide registers, and therefore it was possible, with few exceptions, to include anyone who was diagnosed with sarcoidosis in the years 2007-2016, and not just a selection. In conclusion, individuals with sarcoidosis are at higher risk of comorbidities and mortality compared to matched controls as well as the general population. Thus healthcare professionals who meet sarcoid patients, should strive to identify and treat comorbidities to reduce the risk of impaired quality of life and, eventually, premature death among individuals with sarcoidosis.

In Sweden, all residents have equal access to publicly funded healthcare and residents have a unique personal identification number that can be used to link data across different nationwide register. The National Board of Health and Welfare (NBHW) holds several registers, including the National Patient Register (NPR). Information on hospitalizations (inpatient register) and non-primary care outpatient visits (outpatient register) can be gathered from the NPR. Another register, The Cause of Death Register, provides the basis for official statistics on both main and contributory causes of death in Sweden. These registers have a high degree of coverage due to reporting obligations. The outpatient register was used to identify cases for the study population. The Cause of Death Register provided information on date and cause of eventual death. Statistics Sweden is another Swedish authority accountable for many registers, including The Multi-Generation Register and The National Register of the Total Population, both used in this study to identify controls. All data processing conducted to match data from the different registers was performed by the NBHW and Statistics Sweden and de-identified data recorded from 2005 – 2016 was received. The study population comprised of men and women aged 20 to 65 years (the typical working age in Sweden) diagnosed with sarcoidosis according to the outpatient register from January 1, 2005 through December 31, 2016. Sarcoidosis was defined according to the ICD10 code D86. Controls were matched to cases (2:1) based on age, sex and county at the time of diagnosis ( Figure 1 ). The controls selected did not themselves have sarcoidosis, nor were they a first-degree relative of any case. The study population derived from a larger source of material that formed the basis of another study, thus neither the cases nor the controls are diagnosed with Ankylosing spondylitis-M45, Rheumatoid arthritis-M05 and -MO6, Crohn’s disease-K50 or Ulcerative colitis-K51. To include as many newly-diagnosed cases and as few revisits as possible, a wash-out was performed over the years 2005-2006 to ensure that only newly-diagnosed cases was used in the statistical calculations. Flow chart that visualizes inclusion and exclusion of the study population. To estimate comorbidities, information from the inpatient register about the main diagnoses provided on the first visit (date) after the date of the diagnosis of sarcoidosis in respective chapters of ICD10 was collected. Each individual can appear within several different chapters of the ICD10. Demographics are presented using descriptive statistics and is reported as number (n), percentage (%), and mean ± standard deviation (SD). Using Cox regression, sarcoid cases were compared with their controls, and the results are presented as hazard ratios (HR) with 95% confidence intervals (CIs). Study subjects contributed person years from the date they received their sarcoidosis diagnosis to the date of event (that is, comorbidity and/or death), emigration or end of the study. Standardized incidence ratios (SIRs) and standardized mortality ratios (SMRs) were used to compare sarcoid cases and controls against the general population. With the general population in Sweden used as a reference, data was stratified according to sex, 5-year age groups and 1-year calendar periods when calculating person-years at risk. Assuming a Poisson distribution of the observed numbers, SIRs and SMRs was calculated with 95% CIs. P-values less than 0.05 were considered statistically significant. All analyzes were made using STATA 14.0 (StataCorp LLC, TX; USA). The Swedish Ethical Review Authority (Ref 2017/252) approved the study protocol.