IgG4-Related Disease: Baseline clinical and laboratory features in 125 patients with biopsy-proven disease

Abstract

Purpose

IgG4-related disease (IgG4-RD) is an immune-mediated fibroinflammatory condition that can affect nearly any organ. No detailed clinical and laboratory assessments have been reported in large numbers of patients with IgG4-RD diagnoses established by strict clinicopathological correlation.

Methods

We reviewed the baseline features of 125 patients with biopsy-proven disease. The diagnosis was confirmed by pathology review according to consensus diagnostic criteria. Disease activity and damage were assessed by the IgG4-RD Responder Index (RI). Flow cytometry was used to assess levels of circulating plasmablasts.

Results

Of the 125 patients, 103 had active disease and 86 were on no treatment. Only 51% of the patients with active disease had elevated serum IgG4 concentrations. However, patients with active disease and elevated serum IgG4 concentrations were older, had a higher RI, a greater number of organs involved, lower complement levels, higher absolute eosinophil counts, and higher IgE levels compared to those with active disease but normal serum IgG4 (P<0.01 for all comparisons). The correlation between IgG4+ plasmablast level and RI (R=0.45, P=0.003) was stronger than that of total plasmablasts and RI. Seventy-six (61%) of the patients were male, but no significant differences according to gender were observed with regard to disease severity, organ involvement, or serum IgG4 concentrations. Glucocorticoids failed to produce sustained remission in the majority of patients.

Conclusion

Nearly 50% of this patient cohort with biopsy-proven, clinically-active IgG4-RD had normal serum IgG4 concentrations. Serum IgG4 elevation identify a subset with more inflammatory features. IgG4+ plasmablasts correlate well with disease activity.

Introduction

IgG4-related disease (IgG4-RD) has emerged over the last decade as a unique immune-mediated condition that links multiple fibro-inflammatory disorders previously regarded as separate entities (1). IgG4-RD, initially described in a cohort of Japanese patients with “sclerosing pancreatitis” (2,3), has now been reported across an ethnically diverse spectrum in nearly every organ (4,5). The disease has a predilection for middle-aged to elderly individuals and a tendency to affect males (1,5-8). The incidence and prevalence of IgG4-RD have not been established comprehensively because of the relative novelty of this diagnosis, but the burgeoning medical literature on this condition from all subspecialties within medicine, surgery, radiology, and pathology suggests that the disease remains under-recognized.

The gold-standard for the diagnosis of IgG4-RD, regardless of the organ(s) involved, is the identification of typical histopathology features (e.g., lymphoplasmacytic infiltrate, storiform fibrosis, obliterative phlebitis) in the context of a significant IgG4+ plasma cell infiltrate (9). Serum IgG4 concentration elevations were once considered the sine qua non of the diagnosis, but normal serum IgG4 concentrations are now well-described even in the setting of active, biopsy-proven disease (10-13). The significance of the IgG4 antibody in disease pathogenesis remains unclear (14). Recent publications have described elevations in the levels of circulating plasmablasts, suggesting that the levels of such cells serve as both a diagnostic biomarker and a reliable indicator of disease activity (12, 15).

Most literature on IgG4-RD consists of case reports and small case series focused on individual organ manifestations, particularly the pancreas, biliary tree, and salivary glands (10,16-23). We report here a detailed analysis of the clinical features of the first 125 patients with biopsy-proven IgG4-RD evaluated at our center.

Methods

Cohort overview

This study was approved by the Partners Institutional Review Board. All patients signed written informed consent. The Massachusetts General Hospital Center for IgG4-Related Disease maintains a database of all patients with IgG4-RD evaluated in the Center. Data pertaining to demographics, prior treatment, and laboratory findings at baseline evaluations were derived from the medical record.

Age at onset refers to the age at which the patient first noticed the symptoms ultimately attributed to IgG4-RD or to the timepoint at which the disease was recognized first (whichever was earlier). Organ involvement was determined by a review of the patient's history, physical examination findings, imaging results, laboratory studies, and tissue biopsies. Damage (e.g., pancreatic insufficiency, chronic kidney disease, palate destruction) was determined through an organ-by-organ assessment of function, radiographic abnormalities (e.g., organ atrophy), damage related to surgical intervention (e.g., modified pancreatectomy), or chronic pain. Disease activity was assessed by the IgG4-RD Responder Index (IgG4-RD RI)(24). An IgG4-RD RI score ≥ 1 was classified as active disease. Because of growing awareness that the serum IgG4 concentration has shortcomings as a disease biomarker (11), we report IgG4-RD-RI scores without the inclusion of serum IgG4 concentrations. The methodology of scoring the IgG4-RD RI has been described in detail and is included in the Appendix (Supplement 1).

Pathology methods

All patients had biopsies from at least one organ that were reviewed and confirmed at our center. For some cases, archived pathology samples were reviewed, re-stained, and re-interpreted in order to establish the diagnosis. For certain analyses, only patients with active disease at the time of initial diagnosis were included. The pathology methods have been previously described (12,15).

Pathology definitions

The pathology diagnosis of IgG4-RD is predicated upon the presence of both specific histopathologic features and an increased number of IgG4+ plasma cells (or IgG4+/IgG+ ratio) in affected tissue (9). Careful clinicopathologic correlation is essential because no single pathology feature is diagnostic of IgG4-RD. The major histopathologic features include a dense lymphoplasmacytic infiltrate, fibrosis that has a storiform pattern in focal (or diffuse) areas, and obliterative phlebitis. At least two of these three features were present in all cases. In the immunohistochemistry studies, we considered more than 10 IgG4+ plasma cells/high power field (hpf) and an IgG4+:IgG+ ratio of at least 0.40 to be consistent with the diagnosis of IgG4-RD (9).

Serum IgG4 assays and other laboratory assessments

Serum IgG4 concentrations were measured by immunonephelometry using a Siemens BNII instrument and reagent sets obtained from either Siemens or The Binding Site. The prozone effect, known to occur with some nephelometry assays in the setting of high antigen excess (25), was avoided by diluting the samples sufficiently until consistent measurements were obtained from one serial dilution to the next. The methods for measuring circulating plasmablasts by flow cytometry have been reported (15).

Statistical analysis

All statistical tests were performed using SPSS (Version 22, Chicago, IL, USA). Statistical differences for continuous, normally-distributed data were analyzed by Student t-tests; all continuous non-normally distributed data were analyzed by non-parametric testing. Categorical variables were assessed by Chi-square or Fisher's Exact Tests, as appropriate. Correlations were determined using the Pearson correlation coefficient (normally distributed) or the Spearman correlation coefficient (nonparametric data). A P-value <0.05 was considered significant for all statistical testing.

To compare the prevalence of diabetes mellitus among patients in our cohort with the general United States population, we used the indirect standardization method (26-29) based on the 2011 prevalence of diabetes mellitus in the United States (http://www.cdc.gov/diabetes/statistics/prev/national/tprevage.htm, accessed on 10/8/2014). We calculated a standardized prevalence ratio as the ratio of observed to expected cases. We calculated 95% confidence intervals assuming a Poisson distribution; a P-value was calculated using a χ² test with a continuity correction.

Results

Patient demographics

The average age at evaluation was 55.2 years (+/- 13.9, range: 24-83)(Table 1). The average age at onset was 50.3 years (+/- 14.9, range: 12-82). At the time of initial evaluation at our center, 86 patients (69%) were on no treatment and 39 (31%) were receiving immunosuppression for IgG4-RD (usually GCs). One hundred and three patients had active disease at their initial visit and new diagnoses of IgG4-RD were established in 93 patients (74%) upon their initial visit to our clinic. Patients newly-diagnosed upon visiting our clinic had a mean duration of disease of 5.0 years +/- 7.5, compared with 5.8 years +/- 11.2 for those diagnosed before attending our clinic.

Table 1. Baseline Features of 125 Patients with Biopsy-Proven IgG4-Related Disease.

Age at Evaluation (mean +/-SD, range, years)	55.2 +/- 13.9 (24-83)
Age of Onset (mean +/-SD, range, years)	50.3 +/- 14.9 (12-82)
Disease Duration (mean +/-SD, range, years)	5.2 +/- 8.5 (0-60)
Number of Organs (mean +/-SD, range)	2.29 +/- 1.3 (1-7)
% Organ Damage	58%
% Male	60.8%
% Caucasian	76%
% Active Disease	85.6%
% on Treatment	31.2%
Responder Index (without IgG4 concentration) among active disease (mean +/- SD, range)	6.6 +/- 4.6 (1-33)
% Elevated IgG4 among active disease	51%
% Elevated IgG4 among active disease but off of treatment	55.3%†
% Prior Glucocorticoid Treatment	51%
% Responding to Glucocorticoid Treatment	83%
% Prior Biliary Stents	6.4%
% Prior Ureteral Stents	6.4%

^†There was no significant difference in the proportion of patients with an elevated serum IgG4 concentration when comparing those with active disease on treatment to those off treatment (P=0.263).

Of the 125 patients, 76 (61%) were men. The majority (76%) were non-Hispanic whites but other ethnicities were also represented, including seven African-Americans, eight Hispanics, eight Asians, three South Asians, and two Arabs.

Baseline disease activity

The mean IgG4-RD RI was 5.7 (+/- 4.9; range: 0-33). The RI correlated strongly with serum IgG4 concentrations (r=0.5; P<0.001) but also correlated with the concentrations of total IgG and the individual concentrations of IgG subclasses 1, 2, and 3 (r=0.3, r=0.2, r=0.3, and r=0.2, respectively; P<0.03 for all correlations).

Organ involvement

Table 2 displays the range of organs involved. The patient cohort included disease involvement at over 23 different known anatomic sites of disease, with a mean of 2.3 +/- 1.3 organs involved (range: 1-7). The most commonly involved organs were the submandibular glands (35 cases, 28%), lymph nodes (34, 27%), orbits (28, 22%), pancreas (24, 19%), and retroperitoneum (23, 18%). There were no cases of breast or pituitary involvement. Thirty-eight percent of patients had at least three organs involved, 24% had two organs involved, and 38% of patients had only one organ involved.

Table 2. Sites of Organ Involvement in 125 Patients with Biopsy-Proven IgG4-Related Disease.

Organ Involvement
	# of cases (% of all cases)		# of Cases (% of all cases)
Submandibular	35 (28%)	Sinusitis	5 (4%)
Lymph nodes	34 (27.2%)	Prostate	4 (3.2%)
Orbit	28 (22.4%)	Meninges	3 (2.4%)
Pancreas	24 (19.2%)	Liver	3 (2.4%)
Retroperitoneal fibrosis	23 (18.4%)	Nasal Cavity	3 (2.4%)
Lung	22 (17.6%)	Sclerosing mediastinitis	2 (1.6%)
Parotid	21 (16.8%)	Sclerosing mesenteritis	2 (1.6%)
Other*	18 (14.4%)	Heart	2 (1.6%)
Kidney	15 (12%)	Gallbladder	2 (1.6%)
Aorta	14 (11.2%)	Skin	2 (1.6%)
Bile duct	12 (9.6%)	Mastoid	2 (1.6%)
Other ENT	10 (8%)	Tonsillitis	1 (0.8%)
Thyroid	7 (5.6%)
Damage
Submandibular Gland	10 (8%)	Orbit	4 (3.2%)
Pancreas	10 (8%)	Chronic Pain	4 (3.2%)
Kidney†	10 (8%)	Aorta/Valve	4 (3.2%)
Lung	10 (8%)	Sinus	3 (2.4%)
Other**	9 (7.2%)	Biliary	3 (2.4%)
Non-Aorta Vascular	6 (4.8%)	Palate	2 (1.6%)
Genitourinary	5 (4%)	Nasal Septum	2 (1.6%)
Thyroid	5 (4%)

^*Other Organ Involvement (^#1 patient, ˆ2 patients): Peripheral nerves^#, Appendix^#, Spleen^#, Cervical Fibrosis^#, Paraspinal mass^#, Palateˆ, Pulmonary Artery^#, Coronary Arteries^#, Boneˆ, Gum^#, Omentum^#, Testicle^#, Vocal Cords^#

^**Other Damage: Hearing loss^#, Jawˆ, Parotid^#, Cosmetic^#, Ear^#, Cranial Nerve^#, Small Intestine^#, Pericardium^#

^†Damage to the kidneys included both the sequelae of tubulointerstital nephritis and the development of hydronephrosis caused by retroperitoneal fibrosis.

Damage at baseline evaluation

IgG4-RD was associated with damage caused by the disease (as opposed to treatment) in a total of 73 patients (58%) at the time of the initial evaluation (Table 2). These patients accumulated 89 items of damage at a total of 22 anatomic sites. Nineteen patients (15%) had diabetes mellitus upon presentation. This figure is not significantly different from the age-standardized expected number of cases (14.8 cases or 11.9%; 95% CI 9.2%-23.7%; P=0.91).

Surgery

Fifty patients (40%) had undergone surgery during evaluations of their illness. The surgical procedures were usually performed for the purpose of diagnosis or for erroneous diagnoses of cancer. The most common surgical procedures were pancreatic surgery (Whipple procedure/partial pancreatectomy, 9 patients), followed by submandibular excision (8 patients) and ENT/sinus surgery (8 patients). Among 34 patients with lymphadenopathy, 21 had undergone a lymph node biopsy. The remaining diagnoses of lymphadenopathy were confirmed by imaging.

Previous treatment

Glucocorticoids (GC) were the most common therapy administered to patients prior to their evaluation (64 patients; 51.2%). The great majority of patients in this cohort failed to achieve stable disease remissions following GC treatment. Of the 64 GC-treated patients, nine (14%) did not respond to this therapy (per patient and/or other provider report); 26 (40%) achieved remission and discontinued GC but then relapsed; 15 (23%) achieved remission but flared with attempts at tapering; 11 (17%) responded and remained on GCs at the time of evaluation; 2 (3%) were off of GCs and reported that their disease had been stabilized by GC; and 1 (1.5%) was treated with both GC and a steroid-sparing agent and achieved disease control.

Steroid-sparing medications and other treatments

Both conventional GC-sparing medications and mechanical interventions (i.e., stents) had also been employed in these patients. The GC-sparing medications included methotrexate (7 patients), mycophenolate mofetil (3 patients), tamoxifen (3 patients), azathioprine (5 patients), cyclophosphamide (2 patients), intravenous immune globulin (1 patient), and rituximab (7 patients). All patients treated with steroid-sparing agents (other than rituximab) had discontinued them prior to evaluation due to ineffective disease control. Two patients had been treated with radiation therapy for orbital disease; one was treated after surgery and did not have recurrence at the original site but the other experienced progressive disease despite the concomitant use of steroid-sparing agents. Sixteen patients (13%) had received either ureteral or biliary tract stents for complications related to IgG4-RD (8 ureteral stents, 8 biliary stents).

Laboratory findings

Among the 103 patients with active disease at the time of baseline evaluation, the mean age was 56.1 years (+/- 13.8) and the mean age at disease onset was 50.5 years (+/- 15.1). These patients had a mean of 2.3 organs (+/- 1.3) affected by their disease. None of these values were different from those of patients who had inactive disease at the time of their evaluation in our center. Twenty-nine of the patients with active disease were on treatment at the time of evaluation.

Among those with active disease, there was no significant difference in the serum IgG4 concentration between those on treatment (median 123mg/dL, IQR: 52-258; normal < 135 mg/dL) and those not on treatment (median 155mg/dL, IQR: 53-448, P=0.66). Among patients with active disease on whom erythrocyte sedimentation rate (ESR) or C-reactive protein (CRP) data were available – 98 and 92 patients, respectively – 53% had an elevated ESR (median 52mm/hr, IQR 35-78) and 40% had an elevated CRP (median 20.3mg/L, IQR 11.6-64.9). There was no difference in the distributions of ESR and CRP values (P=0.76).

Normal versus elevated serum IgG4 concentration patients

Among the 103 patients with active disease at the time of their evaluation, only 53 (51.4%) had elevated serum IgG4 concentrations. Among those with active disease who were not on treatment (76 patients), 42 (55.3%) had elevated serum IgG4 concentrations (normal <135mg/dL, P=0.26, Table 3). Patients with active disease and elevated serum IgG4 concentrations were older, had a higher RI score, a greater number of organs involved, lower complement levels, a higher absolute eosinophil count, and higher IgE levels compared to those with active disease but normal serum IgG4 levels (P<0.01 for all comparisons). Patients with pancreatic involvement were more likely to have an elevated serum IgG4 concentration than patients without pancreatic involvement (77% versus 47%, P=0.03).

Table 3. Elevated vs. Normal Serum IgG4 Concentration Among Patients with Active Disease.

	Normal Values	Elevated Serum IgG4 Concentration (>/= 135mg/dL), N=52	Normal Serum IgG4 Concentration (<135mg/dL) N=50	P-Value*
Age at Evaluation (mean, yrs)		60.3 (+/- 11.8)	51.4 (+/- 14.7)	0.001
Age of Onset (mean, yrs)		54.3 (+/- 15.1)	46.8 (+/- 15)	0.014
Gender (% Male)		58.5%	60%	1.0
Multi-Organ (%)		50.9%	28%	0.03
Single Organ (%)		28.3%	44%	0.11
Number of Organs, mean +/- SD		2.7 (+/- 1.5)	1.9 (+/- 1)	0.001
Damage		58.5%	54%	0.69
Responder Index without Serum IgG4, mean +/- SD		8.4 (+/- 5.3)	4.8 (+/- 2.9)	<0.001
Serum total IgG, median (IQR) mg/dL	767-1590	1573 (1240-2114)	1130 (868-1310)	<0.001
Serum IgG1, median (IQR) mg/dL	341-894	868 (650-1090)	596 (499-724)	<0.001
Serum IgG2, median (IQR) mg/dL	171-632	519 (381-750)	366 (271-446)	<0.001
Serum IgG3, median (IQR) mg/dL	18.4-106	76 (43-136)	71 (41-90)	0.04
Serum IgG4, median (IQR) mg/dL	2.4-135	379 (207-817)	51 (28-96)	<0.001
C3, median (IQR) mg/dL	86-184	96 (73-129)	122 (106-155)	0.001
C4, median (IQR) mg/dL	16-38	20 (11-27)	28 (22-30)	0.001
Hypocomplementemia (C3 or C4) %		42%	5.4%	<0.001
ESR, median (IQR) mm/hr	<20	34 (11-71)	16 (9-52)	0.07
CRP, median (IQR) mg/L	<8.0	5.5 (1.8-17.6)	6.5 (1.8-13.0)	0.9
Elevated ESR and/or CRP (%)		67%	46%	0.04
Absolute Eosinophil, median (IQR)/mm3	0.0-900	290 (110-595)	160 (60-250)	0.005
Serum IgE, median (IQR)/mL	0-100	128 (36-462)	46 (16-117)	0.01
Plasmablast, median (IQR)/mL		3246 (1290-5525)	3325 (1089-4712)	0.36
IgG4+ Plasmablast, median (IQR)/mL		2424.7 (781-4961)	1192 (396.9-3190)	0.17
%IgG4+/Total Plasmablast, mean +/- SD		66% +/- 16.5%	45% +/- 25.8%;	0.01

^*By Fisher's Exact Test (Categorical), Unpaired T-Test (Parametric), Mann-Whitney U Test (Non-Parametric)

Other IgG subclasses

Elevations of IgG subclasses other than IgG4 were common among all patients with active disease. Twenty-eight percent of patients with active disease at baseline had an elevated total IgG concentration, 26% had an elevated IgG1 concentration, 23% had an elevated IgG2 concentration, and 26% had an elevated IgG3 concentration. Among those with an elevated serum IgG4 concentration, the other serum IgG subclass concentrations were significantly higher compared to those with a normal serum IgG4 concentration (Table 3). In this same group, 50% had an elevated total IgG, 49% had an elevated IgG1, 36% had an elevated IgG2, and 38% had an elevated IgG3. There was no significant difference in the proportion of patients with an elevated total serum IgG, IgG1, IgG2, or IgG3 (P=0.198). Among the 48 patients with a normal serum IgG4 concentration but active disease and IgG subclass data available, IgG subclass elevations were infrequent: 2 (4%) had elevated IgG1, 5 (10%) had elevated IgG2, and 6 (13%) had elevated IgG3.

Plasmablasts among patients with active disease

Peripheral blood flow cytometry for plasmablasts was performed on 73 patients with active disease. Twenty-nine patients had flow cytometry for IgG4+ plasmablasts. Patients with normal serum IgG4 concentrations had blood plasmablast levels equivalent to those of patients with elevated serum IgG4 levels (3,325/mL [IQR 1,089-4,712] versus 3,246/mL [IQR 1,290-5,525]; P=0.36). The IgG4+ plasmablast levels among patients with normal as opposed to elevated serum IgG4 were also not statistically different (1,192/mL [IQR 397-3,190/mL] versus 2,425/mL [IQR 781-4,961]; P=0.17). However, among the patients with IgG4+ plasmablasts measured, those with elevated serum IgG4 concentrations had a higher percentage of plasmablasts that were IgG4+ than did the patients with normal serum IgG4 concentrations (mean 66% +/- 16.5%, compared with 45% +/- 25.8%; P=0.01).

Serum complement levels

Among the patients with active disease, 20% had low C3 concentrations (mean 59.4 mg/dL; normal 86-184 mg/dL) and 19% had low C4 concentrations (mean 8.0 mg/dL; normal 16-38 mg/dL). Among the twenty patients with hypocomplementemia, both C3 and C4 were depressed in eleven (55%), five of whom (45.5%) had IgG4-related kidney disease. Patients with hypocomplementemia were more likely to have elevated serum IgG1, IgG2, IgG3, and IgG4 concentrations compared to those with normal complement concentrations (P<0.02 for all comparisons).

Correlations between clinical features, serum IgG4, and flow cytometry

Increasing age at baseline evaluation showed moderate correlations with both serum IgG4 concentration (r=0.35, P=<0.001) and the number of organs involved (r=0.25, P<0.01) in univariate analyses. When controlling for age, the serum IgG4 concentration correlated moderately well with the number of organs affected (r=0.39, P<0.001). When controlling for age, there was no correlation between disease duration and serum IgG4 concentration (r= -0.07, P=0.5) or between disease duration and the number of organs affected (r=0.1, P=0.31). Thus, the number of organs affected correlated with serum IgG4 concentration but not with disease duration.

Among all patients with peripheral blood plasmablast measurements (n=94), there were moderate correlations between the plasmablast levels and serum IgG4 concentration (r=0.29, P<0.01) as well as the RI (r=0.22, P<0.03). The correlation between IgG4+ plasmablast level (n=42) and RI (R=0.45, P=0.003) was stronger than that of total plasmablasts and RI.

Clinical differences according to gender

The IgG4-RD patients in our cohort were more likely to be male than female, but we detected no differences with regard to disease severity or organ involvement. A higher percentage of male patients had IgG4-related autoimmune pancreatitis (25% versus 10%), but this comparison fell short of statistical significance (P=0.06). Male (n = 76) and female (n = 49) patients were similar in terms of the age at onset (51.9 +/-14.8 years versus 47.6 +/-14.8 years; P=0.12), number of organs involved (2.3 +/- 1.4 versus 2.2 +/- 1.1, P=0.2), RI at baseline (5.9 +/-5.4 versus 5.3 +/- 3.8, P=0.5), and estimated duration of disease (5.0 +/- 9.3 vs. 5.5 +/- 7.2 years) preceding presentation (P=0.8).

Among the patients with active disease at the time of evaluation, the males (n = 61) and females (n = 42) had similar IgG4 concentrations (median 141 mg/dL, IQR: 58-493 versus 140, IQR: 52-255; P=0.41). There were also no significant differences in the proportions of male and female patients with an elevated IgG4 concentration, hypocomplementemia, multiorgan disease, or elevated inflammatory markers (data not shown). Finally, no significant differences existed between the groups with regard to peripheral blood plasmablast levels (male median 1,558/mL, IQR 118-4,705, versus female median 1,529/mL, IQR 0-3,572; P=0.43) or IgG4+ plasmablast levels (1,558/mL, IQR 118-4,705 versus 1,529/mL, IQR 0-3,572, P=0.23).

Disease patterns according to organ involvement

We detected certain patterns of laboratory abnormalities according to some types of organ involvement. Patients with renal disease, lymphadenopathy, and retroperitoneal fibrosis, for example, comprised distinctive disease subsets. Comparisons are shown in Table 4. Whereas 7 of 10 patients with active IgG4-related kidney disease were hypocomplementemic at baseline, none of the 13 patients with active, untreated RPF had hypocomplementemia (P <0.001). Among the 23 patients with RPF, only one also had kidney disease and only one had lymphadenopathy beyond the retroperitoneum. Serum IgG4 concentrations varied significantly across the subgroups, with medians of 406 mg/dL, 228 mg/dL, and 58 mg/dL, respectively (P <0.03 for all comparisons).

Table 4. Characteristics of Selected Disease Subsets: Patients with IgG4-Related Kidney Disease, IgG4-Related Lymphadenopathy, and IgG4-Related Retroperitoneal Fibrosis.

Characteristic	Kidney Disease versus No Kidney Disease (n = 15)	Lymphadenopathy (LAD) versus No LAD (n = 34)	Retroperitoneal Fibrosis (RPF) versus No RPF (n = 23)
Organs involved (mean, S.D.)	3.1 +/- 1.8 vs 2.2 +/- 1.3 P = 0.02	3.3 +/- 1.5 vs 1.9 +/- 1.1 P<0.001	1.8 +/- 1.04 vs 2.4 +/- 1.4 P=0.045
Serum IgG4 (mg/dL) (median, [IQR])	406 [137-1102] vs 132 [49-353] P=0.02	228 [74-828] vs 131 [47-287] P=0.03	58 [20-131] vs 193 [74-493] P=0.001
Elevated Serum IgG4 Concentration (%)	75% vs 48.4 % P=0.12	63.3% vs 46.6% P=0.14	19% vs 59.8% P=0.001
C3 (mg/dL) (median [IQR])	68 [57-96] vs 119 [93-150] P<0.01	No difference	No difference
C4 (mg/dL) (median [IQR])	12 [7-19] vs 25 [17-30] P<0.01	No difference	28 [23-31] vs. 23 [13-28] P=0.052
Hypo-complementemia (%)	70% vs. 18% P=0.002	No difference	0% vs. 29% P=0.03
ESR (mm/hr) (median [IQR])	67 [53-84] vs. 18 [8-53] P= 0.001	No difference	No difference
C-reactive protein (mg/L) (median [IQR])	No difference	No difference	No difference
Plasmablast (/mL) (median [IQR])	1387 [975-3786] vs. 3784 [1529-5720] P=0.03	No difference	1038 [861-2810] vs. 3953 [1579-5720] P=0.002
IgG4+ Plasmablast (/mL) (median [IQR])	No difference (N=3)	No difference (N=6)	535 [230.35-781.07] vs. 3061 [969.25-5979.6] P=0.001 (N=7)
IgG4+ Plasmablast/Total Plasmablast (%)	No difference (N=3)	No difference (N=6)	No difference (N=7)

^*By Fisher's Exact Test (Categorical), Unpaired T-Test (Parametric), Mann-Whitney U Test (Non-Parametric)

Discussion

We describe the baseline clinical and laboratory features in 125 patients with biopsy-proven IgG4-RD. Our study is unique in comparison to other IgG4-RD publications to date, most of which have been multi-center efforts focusing on autoimmune pancreatitis or small case series. Only three studies have described cohorts of IgG4-RD patients with a variety of organ manifestations. The largest of these studies to date described 114 Japanese patients identified by a review of 20 years of pathology records (7). The focus of that study was the pathology of IgG4-RD in different organs rather than the clinical characteristics or laboratory features of disease. Chen et al. described 28 Chinese patients at a single center (30) and Ebbo et al. reported 25 patients from a multi-center registry comprised of reports by French physicians (6). Both of those series required an elevated serum IgG4 concentration as a criterion for entry.

A striking observation from our study is that only 51% of patients with active, untreated IgG4-RD had elevated serum IgG4 concentrations at baseline. This figure contrasts sharply with previous reports of patient series collected according to different criteria (6,10-11,19,30). Because serum elevations of IgG4 were the basis upon which IgG4-RD was first identified (2,3), early studies and proposed diagnostic criteria generally attached crucial importance to serum IgG4 concentrations (31). However, the use of serum IgG4 concentration elevations as a principal means of identifying patients inevitably skews study populations toward those with high serum levels. In our study, in contrast, the diagnosis was based on clinicopathologic correlation regardless of the serum IgG4 concentration. Besides differences in criteria used for the diagnosis of IgG4-RD, a variety of other explanations, including differences in patient demographics (e.g., ethnic background) and organ involvement may also contribute to the lower rate in our cohort. We note that our study included 23 patients (18% of the overall cohort) with RPF, 85% of whom had normal serum IgG4 concentrations despite having biopsy-proven IgG4-RD. In contrast, nearly 80% of those patients with pancreatic involvement (n=24) had an elevated serum IgG4 concentration. The enrichment of our study population with RPF patients probably partly explains the lower frequency of IgG4 serum elevations in our cohort. When the RPF patients were excluded, 33.3% of the remaining 57 patients on no treatment at baseline still had normal serum IgG4 concentrations.

Although our data highlight the shortcomings of serum IgG4 concentrations in the diagnosis and serial assessment of IgG4-RD, they also suggest that elevated serum IgG4 concentrations identify a subtype of IgG4-RD that is more “inflammatory” compared to patients with normal serum IgG4. That is, serum elevations in IgG4 concentration denote a large subset of IgG4-RD that is characterized by a greater likelihood of multiple organ involvement, high inflammatory markers, low complement levels, and (possibly) greater refractoriness to treatment compared to patients with normal IgG4 levels (10,12,32-33).

Findings from this study confirm our previous observations that IgG4-RD patients have elevated levels of circulating plasmablasts that are independent of serum IgG4 concentrations (12,15). Moreover, the levels of circulating plasmablasts – especially IgG4+ plasmablasts – correlate well with disease activity as assessed by the IgG4-RD RI. Longitudinal studies of the utility of peripheral blood flow cytometry for gauging the need for re-treatment are an important agenda for future study.

Our study provides important additional information about the utility of GC in IgG4-RD. GC were commonly employed as a treatment modality in our patients and 83% of the patients treated with GC responded at least initially to this therapy, a finding consistent with prior reports (6,34,35). However, 77% of the patients treated with GC failed to achieve stable, steroid-free remissions following the discontinuation of GC, a somewhat higher prevalence than prior reports (36). This figure may be biased somewhat by the nature of our referral practice, which is perhaps more likely to see patients with severe, treatment-refractory disease. Many patients either failed to respond to GC or experienced disease flares while still on tapering GC doses, confirming an observation from other studies (6,34,37). Even though GC are still regarded as the initial treatment of choice, our data highlight a growing concern that control of IgG4-RD presently necessitates either ongoing maintenance treatment with GC or repeated treatment courses for disease flares (19,34-36,38-39). Such approaches are problematic for a disease that targets middle-aged to elderly individuals and frequently affects pancreatic function.

IgG4-RD is widely acknowledged to affect men more commonly than women, as observed in our cohort (6,18-19,40). The gender disparity in this cohort, however, was less prominent than often reported in the literature (male:female 1.6:1). Previous reports have described higher serum IgG4 concentrations among healthy males compared to healthy females (41), but in this study we observed no difference in the serum IgG4 concentrations between male and female patients. We also found no differences in organ involvement or disease severity as measured by the extent of organ involvement through the IgG4-RD RI.

The concept of permanent “damage” resulting from IgG4-RD has not been addressed in a large cohort of patients with diverse organ involvement. More than half (58%) of the patients in our cohort had sustained organ damage from IgG4-RD at the time of their evaluation in our center. The organs particularly susceptible to damage in our cohort were the pancreas, lungs, kidneys, and major salivary glands. Particularly concerning is the large number of patients (40%) who had undergone surgical procedures (excluding fine-needle aspiration) before the diagnosis of IgG4-RD could be established. The most frequent surgical procedures were pancreatic resections and submandibular gland excision; in most cases, the diagnosis of IgG4-RD was based on findings related to the procedure. Broader education and greater awareness of IgG4-RD as a mimicker of malignancy may obviate the need for many surgical procedures.

Our data may provide insights into the mechanism of the hypocomplementemia observed in many patients with IgG4-RD. The IgG4 molecule is believed to bind complement poorly under nearly all circumstances. Activation of the classical complement pathway by IgG4 is therefore an improbable explanation for the marked hypocomplementemia sometimes observed (18,40,42). In our cohort, patients with hypocomplementemia and elevated serum IgG4 concentrations also tended to have higher concentrations of other serum IgG subclasses, suggesting that one of the other IgG subtypes (e.g., IgG1 or 2) is responsible for low complement levels (43). Future studies might evaluate for the presence of anti-hinge antibody (targeted against the IgG4 hinge) among IgG4-RD patients with an elevated serum IgG4 concentration and hypocomplementemia. Such antibodies were identified recently in rheumatoid arthritis patients with complement activation (42).

Substantial variations in disease expression exist in IgG4. Major differences can be observed, for example, in comparisons across organ systems. As an example, patients with IgG4-related RPF and IgG4-related kidney disease differ substantially in terms of the number of organs involved, complement levels, and serum IgG4 concentrations (32). The variability of disease expression according to organ involvement is one likely explanation for the failure to recognize IgG4-RD as a unified, multi-organ disease until this century.

Our study has a number of important strengths. First, it is the largest study of biopsy-proven IgG4-RD affecting a broad array of organs reported to date, as well as the first pathology-based study to emphasize the clinical manifestations of disease. Second, we had the opportunity to evaluate many patients at a time when they were on no treatment. A particular strength of this work is that 103 patients had active disease and most were on no treatment at the time of their evaluation. We were therefore able to comment in detail on the baseline clinical appearance and laboratory characteristics of IgG4-RD. Third, we were able to further evaluate the significance of circulating plasmablasts in IgG4-RD.

Our study also has a number of limitations. Our center is located within a Rheumatology Unit and therefore our patient population is skewed toward patients with multi-organ disease or at least the types of organ manifestations most likely to be evaluated by rheumatologists. Pancreatic and other gastrointestinal manifestations of IgG4-RD may be under-represented in our study because patients with disease isolated to those sites are often managed primarily by gastroenterologists. Nevertheless, patients with type 1 (IgG4-related) autoimmune pancreatitis and IgG4-related sclerosing cholangitis contributed significantly to our study population. It is also possible that the clinical and laboratory features of our cohort of Western patients differs in some important ways from those of patients from other parts of the world. Some investigators have proposed that IgG4-RD patients from Japan and those in Western countries differ in some important respects. This hypothesis needs to be addressed through comparison studies of patient cohorts collected at different sites across the world. Finally, because of the retrospective nature of some aspects of our data collection, some elements of the data set (e.g., age at onset and organ involvement) may be affected by recall bias.

In conclusion, we have reported a large cohort of patients with IgG4-RD and diagnoses proven by histopathology. These data substantially augment our understanding of IgG4-RD, now more than a decade after its initial description. Serum IgG4 concentrations appear to be less important than originally believed for the purposes of diagnosis. Our findings have important implications for the diagnosis and management of IgG4-RD and may help guide the design of future studies.