Clues to Mortality Trends and Their Related Factors in IgG4-related Disease: a Japanese Single-center Retrospective Study

Background: Few observations on the long-term prognosis have been conducted in immunoglobulin G4-related disease (IgG4-RD) patients with various organ involvement, not limited to autoimmune pancreatitis. Especially, mortality and its related factors in patients with IgG4-RD with various organ involvement are not well known. This study aimed to clarify mortality trends and its related factors in IgG4-RD with various organ involvement. Methods: We retrospectively reviewed the medical records of patients with IgG4-RD at a single center in Japan. We calculated the crude mortality rate and the standardized mortality ratio (SMR) using national Japan mortality statistics and investigated the cause of death. We performed Cox regression analyses to assess mortality-related factors. Results: A total of 179 patients with IgG4-RD were included and the median follow-up from diagnosis was 47 months (IQR 19-96). Ten patients (5.6%) in our cohort died during the follow-up period. The crude mortality rate was 11.1 per 1,000 person-years. According to national Japan mortality statistics, 11.6 age- and sex-matched deaths would have been expected to occur within the follow-up period, resulting in an SMR of 0.86 (95% condence interval [CI] 0.41-1.59). Univariate Cox regression analyses indicated that the number of affected organs at diagnosis (hazard ratio [HR] 1.45, 95% CI 1.02-2.05), eGFR <45 mL/min/1.73m 2 at diagnosis (vs. ≥ 45, HR 8.48, 95% CI 2.42-29.79), and the presence of malignancy during the clinical course (HR 3.93, 95% CI 1.10-14.02) had a signicant impact on the time to death. Conclusions: Our ndings suggested that IgG4-RD does not signicantly affect long-term patient survival. On the other hand, multi-organ involvement and renal dysfunction as well as malignancy might be associated with higher mortality trends in IgG4-RD. Early detection and appropriate management of risk factors may improve the long-term prognosis of IgG4-RD. cytoplasmic antibody (ANCA), double-stranded DNA antibody, and anti-citrullinated protein antibody (ACPA); number of affected organs; involvement of pancreas, salivary glands, lacrimal glands, kidney, aorta/artery, retroperitoneum, and lung. We also determined presence or absence of malignancy, relapse, and death during the observation periods. Presence of malignancy during the observation periods was dened as having malignancy at or after the diagnosis of IgG4-RD. G; IgG4: Immunoglobulin G4; IgG4-RD: Immunoglobulin G4-related disease; IgG4-SC: IgG4-related sclerosing cholangitis; IQR: Interquartile range; RF: Rheumatoid factor; ROC: Receiver operating characteristic; SMR: Standardized mortality ratio; s/o: Suspect of.


Background
Immunoglobulin G4-related disease (IgG4-RD) is a multi-organ, bro-in ammatory condition with massforming lesions, organ-swelling, and/or hypertrophic lesions due to dense lymphoplasmacytic in ltrates with IgG4 positive plasma cells 1) . In this disease, an elevation of serum IgG4 levels is a characteristic serological nding, seen in most cases. Generally, patients with this disease respond well to glucocorticoid therapy, but frequently relapse.
Despite a good response to glucocorticoid, the long-term prognosis of IgG4-RD is not always favorable. As mentioned above, relapses are common, with past reports describing relapse rates of 21-46% [2][3][4] . If not treated early and appropriately, progressive brosis in the affected organs causes persistent organ dysfunction, such as chronic salivary and lacrimal gland dysfunction due to IgG4-related dacryoadenitis and sialadenitis 5) and diabetes mellitus and pancreatic exocrine insu ciency due to type 1 autoimmune pancreatitis (AIP) 6, 7) . Particularly, serious complications in IgG4-RD include biliary obstruction, cirrhosis, and portal hypertension due to IgG4-related sclerosing cholangitis (IgG4-SC) 6, 7) , renal dysfunction due to IgG4-related kidney disease 8, 9) or IgG4-related retroperitoneal brosis causing ureteral obstruction 10) , and aortic aneurysm formation and/or aortic dissection due to IgG4-related periaortitis 11) . Furthermore, some recent studies have focused on the potential risk of malignancy in IgG4-RD patients which also can worsen the long-term prognosis of patients. In fact, the mortality is reported to be signi cantly higher in AIP/IgG4-SC than that in the general population, with an estimated odds ratio of 2.07 6) . In addition, cardiovascular events are more common in the elderly and affect their prognosis. Because IgG4-RD generally occurs in the elderly 1) , cardiovascular events may also affect the prognosis of patients with this disease. However, an impact of cardiovascular events on the prognosis of patients with IgG4-RD has not been su ciently investigated.
Few observations on the long-term prognosis have been conducted in IgG4-RD patients with various organ involvement, not limited to autoimmune pancreatitis. Especially, mortality trends and its related factors in patients with IgG4-RD are not well known. Therefore, we undertook the present study to clarify mortality trends compared to the general population in Japan and to identify factors associated with mortality in IgG4-RD with various organ involvement.

Patients and Materials
Between January 1, 2014, and December 31, 2020, we included 179 consecutive patents with IgG4-RD. The diagnosis of IgG4-RD was made by experts based on ful llment of the Japanese comprehensive diagnostic criteria 12) and/or the 2019 ACR/EULAR classi cation criteria for IgG4-RD 13) .
Retrospectively reviewing the medical records, we determined the following clinical factors at the time of diagnosis: age; gender; presence or absence of allergy, dyslipidemia, hypertension, diabetes mellitus, smoking history, and cardiovascular diseases; serum levels of IgG, IgG4, IgE, CH50, C3, C4, C-reactive protein (CRP), and creatinine; estimated glomerular in ltration rate (eGFR) calculated on the basis of a 3-variable Japanese equation 14) ; peripheral blood eosinophil counts; presence or absence of rheumatoid factor (RF), anti-nuclear antibody (ANA), and various disease-speci c autoantibodies including SSA/Ro antibody, anti-neutrophil cytoplasmic antibody (ANCA), double-stranded DNA antibody, and anti-citrullinated protein antibody (ACPA); number of affected organs; involvement of pancreas, salivary glands, lacrimal glands, kidney, aorta/artery, retroperitoneum, and lung. We also determined presence or absence of malignancy, relapse, and death during the observation periods. Presence of malignancy during the observation periods was de ned as having malignancy at or after the diagnosis of IgG4-RD.
Using the collected data, we calculated the all-cause and malignancy-related mortality rate and the standardized mortality ratio (SMR) using national Japan mortality statistics and investigated the cause of death. Incidence of malignancy during the clinical course and standardized incidence ratio (SIR) were also calculated using Cancer Statistics provided by Cancer Information Service, National Cancer Center, Japan (Vital Statistics of Japan, Ministry of Health, Labour and Welfare).
The signi cance of differences between groups was determined using Mann-Whitney U test, while the signi cance of differences in frequencies was analyzed with Fisher's exact probability test. Factors related to mortality were assessed by Cox regression analysis. For the time-to-event analysis, we considered that the follow-up of the patients started at the time of diagnosis and was terminated at the time of the end of the observation period (December 31, 2020), loss of follow-up, or death. In these Cox regression analyses, for continuous variables, units for increments to calculate hazard ratios were set at 1 year for age; 100 mg/dL for serum IgG4 levels; 1 mg/dL for serum creatinine and CRP levels; 10 mL/min/1.73m 2 for eGFR. Receiver operating characteristic (ROC) curve analyses were performed to test the usefulness of certain parameters for the prediction of mortality and to determine the appropriate cut-off values. According to the presence or absence of the detected risk factors, survival rates were calculated using the Kaplan-Meier method. Log-rank test was used to compare survival times of patients with and without malignancies. Signi cant differences were recognized at P < 0.05.
Ten patients (5.6%) in our cohort died during the follow-up period ( Table 2). Five died of malignancy, one of respiratory failure, two of infectious pneumonia, one of sudden cardiac event, and one of suspected aortic aneurysmal rupture. Median period from diagnosis to death was 44 months (IQR 28-79). The crude all-cause mortality rate was 11.1 per 1,000 person-years. According to national Japan mortality statistics, 11.6 age-and sex-matched deaths would have been expected to occur within the follow-up period, resulting in an SMR of 0.86 (95% con dence interval [CI] 0.41-1.59). The cancer-related mortality was not signi cantly increased compared with the general population (4.42 expected deaths versus 5 observed deaths; SMR 1.13, 95%CI 0.36-2.64). On the other hand, according to the Cancer Statistics in Japan, the expected number of cases of ageand sex-matched malignancies to occur within the follow-up period would be 12.7. The actual number of cases with malignancies was 23 in this cohort. Therefore, there was an increased SIR of 1.81 (95%CI 1.14-2.71) for all malignancies compared with the general population.
To clarify the characteristics of the non-surviving group, we compared various clinical features of the surviving and non-surviving groups using Mann-Whitney U test and Fisher's exact probability test ( Table 1). The serum levels of IgE, CRP, and creatinine as well as the frequency of malignancy in the non-surviving group were signi cantly higher than those in the surviving group, whereas eGFR was signi cantly lower in the nonsurviving group.
Factors related to mortality trends were assessed by univariate Cox regression analysis ( To help determine the extent to which the number of affected organs and eGFR can contribute to the reliable prediction of mortality during the clinical course, and to identify their appropriate cut-offs, we used ROC curves ( Figure 1). Concerning the number of affected organs, the area under the ROC curve was 0.640 (95% CI: 0.433-0.846, P=0.137), suggesting that this parameter does not accurately predict mortality. However, we found that multiple organ involvement in more than 4 organs was the most appropriate cut-off and yielded a sensitivity of 60.0% and speci city of 72.7% in predicting mortality ( Figure 1A). Concerning eGFR, the area under the ROC curve was 0.717 (95% CI: 0.544-0.889, P=0.021). We found that eGFR of less than 73.6 mL/min/1.73m 2 was a sensitive cut-off that yielded a sensitivity of 90.0% and speci city of 47.6% and eGFR of less than 38.9 mL/min/1.73m 2 was a speci c cut-off that yielded a sensitivity of 50.0% and speci city of 91.7% in predicting mortality ( Figure 1B).
According to the presence or absence of malignancy, ≥4 affected organs or ≤3 at diagnosis, and eGFR of <45 or ≥45 mL/min/1.73m 2 at diagnosis, survival rates were calculated using the Kaplan-Meier method. Kaplan-Meier analyses also showed signi cant differences in the cumulative rates of death between patients with and without malignancy during the clinical course (P=0.028) (Figure 2A), ≥4 affected organs at diagnosis (P=0.024) ( Figure 2B), and eGFR of <45 mL/min/1.73m 2 at diagnosis (P<0.01) ( Figure 2C).

Discussion
The present study analyzing 179 patients with IgG4-RD showed that the all-cause mortality trends of patients with IgG4-RD did not signi cantly differ from those of the age-and sex-matched Japanese population. The malignancy-related mortality was not increased compared with the matched Japanese population, but the incidence of malignancy was increased. Multi-organ involvement and renal dysfunction at diagnosis as well as malignancy developing during the clinical course were associated with higher mortality trends. Especially, the cumulative survival rate was signi cantly lower in the patients with malignancy compared with those without. These new ndings may provide clues to better long-term management and screening of IgG4-RD.
Although several studies have analyzed the long-term prognosis of IgG4-RD, it has not been fully elucidated because of its rarity. The nationwide epidemiological survey of 1,370 Japanese patients with AIP in 2016 15) reported that 37 (2.7%) patients died within approximately two years of the diagnosis. Huggett et al 6) studied 115 patients with AIP/IgG4-SC and found that 11 patients (10%) died during a mean follow-up of 33 months.
Using National UK mortality statistics, the authors suggested that there was an increased all-cause mortality ratio of 2.07 (95%CI 1.07-3.55). In our study of 179 patients with IgG4-RD, 10 patients (5.5%) died during the follow-up. There was no signi cant increase in the mortality trends compared to the age-and sex-matched population in Japan. This result was contrary to the previous report by Huggett et al 6) , but this might be due to the fact that our cases, unlike the cases of Huggett et al 6) , had not only AIP/IgG4-SC but also various affected organs including mainly salivary and lacrimal gland lesions. On the other hand, these results suggest that, with proper management, the prognosis of patients with IgG4-RD may be similar to that of the general population.
Some recent studies have focused on the potential risk of malignancy in patients with IgG4-RD. Con icting studies have noted both an increased occurrence of malignancies in this group 6, [15][16][17][18][19][20][21][22] and no such increase in others 23,24) , making the association between IgG4-RD and malignancy controversial even now. In our study, the risk of any malignancy at diagnosis or during follow-up was increased (SIR 1.81, 95%CI 1.14-2.71) when compared with the matched national population. In addition, Cox regression analysis indicated that malignancy during the clinical course had a signi cant impact on the time to death. The Kaplan-Meier survival curve also showed that all-cause mortality increased during the clinical course in the patients with malignancy as compared to those without. These results suggest that the periodical screening of systemic organs in patients with IgG4-RD may lead to a further improvement of their prognosis through the early detection and treatment of malignancy.
Furthermore, we found that the number of affected organs and renal dysfunction at diagnosis were associated with mortality trends in IgG4-RD. The number of affected organs is recognized as one of the factors re ecting disease activity in general, and it may also be related to long-term prognosis. Accordingly, the periodical screening of systemic organs is also important for the early detection of subclinical new or recurrent organ involvement. On the other hand, acute kidney injury 25,26) and chronic kidney disease 27) are known to be associated with patient mortality, and this study also suggested that renal dysfunction was an important prognostic factor in IgG4-RD. Our ROC analysis demonstrated eGFR <73.6 mL/min/1.73m 2 to be a sensitive cut-off and eGFR <38.9 mL/min/1.73m 2 a speci c cut-off for prediction of mortality. Of note, IgG4-RKD was not diagnosed in 2 of the 5 non-surviving patients with <38.9 mL/min/1.73m 2 in this study ( Table 2), meaning that renal dysfunction due to not only IgG4-RKD but also any other causes may be associated with a poor prognosis. In patients with IgG4-RD, therefore, it is necessary to also pay attention to the presence of these factors. On the other hand, further studies are necessary to clarify which therapeutic interventions improve the prognosis of patients with multiple organ involvement and/or renal dysfunction.
In this study, cardiovascular events and their classical risk factors, an important cause and risk factors of death in the general population, respectively, were not clearly associated with the mortality trends in patients with IgG4-RD despite glucocorticoid maintenance therapy. Cardiovascular events and their classical risk factors are more common in the elderly irrespective of having IgG4-RD or not and affect their prognosis.
Although IgG4-RD is more prevalent in the elderly 1) , an impact of cardiovascular events and their risk factors on the prognosis of patients with IgG4-RD was not evident in the present cohort. However, in IgG4-related periarteritis, aneurysm or rupture of the affected aorta/artery and/or ischemic events in the coronary arterial lesions are recognized as fatal complications 11) , suggesting that cardiovascular events may be an important cause of death depending on the affected organ in IgG4-RD. It is important to further investigate and clarify the impact of cardiovascular events on the long-term prognosis of this disease in multicenter studies with a larger number of patients.
There are some limitations in this study. First, it was designed as a single-center cohort study. We should con rm the results in populations with different characteristics such as race, affected organ, and treatment. Second, in assessing the incidence of malignancy, this study had a possible risk of ascertainment bias because the incidence of malignancy may be elevated by earlier detection in patients receiving various radiological examinations for follow-up of IgG4-RD. Finally, we should treat conclusions drawn from a comparison with national statistics with caution, because the analysis was not adjusted for other risk factors such as smoking, alcohol intake, and obesity. The cohort size was relatively large for this rare disease, but the number of the patients who died was not su cient to perform multivariate analysis. Further studies involving more patients are needed.
In conclusion, our ndings suggested that IgG4-RD does not signi cantly affect long-term patient survival. On the other hand, multi-organ involvement and renal dysfunction at diagnosis as well as malignancy during the clinical course might be associated with higher mortality trends in IgG4-RD. Early detection of risk factors and appropriate management and intervention are considered to be necessary to improve the long-term prognosis of IgG4-RD.

Declarations
Ethics approval and consent to participate This study was approved by the institutional ethics board of the Medical Ethics Committee of Kanazawa University, and informed consent for the use of all data was obtained from each patient. The research was conducted in compliance with the Declaration of Helsinki.

Consent for publication
Not applicable.

Availability of data and materials
The data used and/or analyzed during the present study will be available on reasonable request to the corresponding author.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by Health and Labour Sciences Research Grants for the Study of Intractable Diseases from the Ministry of Health, Labor and Welfare, Japan.
Authors' contributions HK and IM designed the research, acquired data, contributed to analysis and interpretation of data, and drafted and revised the manuscript. ST, SS, TY, RH, RN, TZ, SH, and KI acquired data and provided intellectual content of critical importance to the work described. MK designed the research, acquired data, contributed to analysis and interpretation of data, provided intellectual content of critical importance to the work described, and revised the manuscript. All authors read and approved the nal manuscript.    Figure 1 Receiver operating characteristic curve analysis. (A) Receiver operating characteristic (ROC) curves to identify the appropriate cut-offs of the number of involved organs for the prediction of mortality show that involvement of more than 3.5 organs yields a sensitivity of 60.0% and a speci city of 72.8% (circle). (B) ROC curves to identify the appropriate eGFR cut-offs for the prediction of mortality show that eGFR of less than 38.9 mL/min/1.73m2 yields a sensitivity of 50.0% and a speci city of 91.7% (square) and eGFR of less than 73.6 mL/min/1.73m2 yields a sensitivity of 90.0% and a speci city of 47.6% (circle).