Materials and Methods

From January, 2005 to December, 2020, we reviewed 59 elec tronic medical records searching for biopsies with diagnosis of retroperitoneal fibrosis at Hospital Italiano de Buenos Aires (Fig. 1). After excluding secondary causes (neoplastic, post radiotherapy and surgery), 27 IRFs were eligible for IgG4-RD according to 2019 ACR/EULAR classification criteria (they met inclusion yet not exclusion criteria⁶). Eight pathology blocks were not available. Finally, 19 adults histopathologically diagnosed with IRF at Hospital Italiano de Buenos Aires were included in a dynamic retrospective cohort.

Pathology slides were reviewed by experienced patholo gists to register histopathological features, including lympho plasmacytic infiltrate, storiform fibrosis, obliterative or non-obliterative phlebitis, presence of eosinophils and germinal centers. The presence of neutrophilic infiltrates, prominent necrosis and granulomas was considered as histopathological exclusion criteria⁶. Immunohistochemistry for IgG4 (Roche, clone: MRQ-44) and IgG (Cell Marque, polyclonal) was con ducted using the Benchmark XT (Ventana) automated system and assessed for each case. The biopsy date matched the date of IRF’s diagnosis and the time elapsing from biopsy to revision was recorded. We used 2019 ACR/EULAR classifi cation criteria to identify IgG4-RPF cases. After a three-step classification criteria process, IgG4-RD was identified if the entry criteria were met (involvement of at least 1 of 11 pos sible organs in a manner consistent with IgG4-RD) and no exclusion criteria were present (32 clinical, serologic, radio logic, and pathologic items). In the third step, the total points had to be ≥ 20 after applying 8 weighted criteria domains addressing clinical and radiology findings, serologic results and pathology interpretation, including histopathology and im munohistochemistry. Only the highest-weighted item in each domain was scored⁶.

Clinical, serologic and imaging data were collected from electronic medical records. We reported the following clinical features: age, sex, allergic disease, hypertension, diabetes, collagen tissue disease, cardiovascular events, thromboembol ic events, presenting symptoms, drug therapy, invasive treat ment, hemodyalisis requirement, number of hospitalizations and death. Other IgG4-related manifestations were recognized by thoroughly analyzing specialist records, biopsies and im age descriptions. Renal failure was defined as a glomerular filtration rate < 60 ml/min: we consider it acute renal failure if it was determined at the time of IRF diagnosis and chronic renal failure if it was present for more than 3 months. We included eosinophilia (defined as > 500 cells/ul in peripheral blood at any time of evolution), serum acute-phase reactants concentrations (erythrosedimentation rate and protein reac tive C) at diagnosis and IgG4 serum levels, when available. In all cases, patients had at least one abdominal and pelvic image. Twelve contrast-enhanced computed tomography of the abdomen and pelvis, 3 computed tomography urography, 6 contrast-enhanced magnetic resonance of the abdomen and pelvis, 2 magnetic uro-resonances and 3 positron-emission tomography/computed tomography images were available. The location of soft tissue in the retroperitoneum and hydro nephrosis presence (either uni- or bilateral) were recorded.

Data are presented as median values and as interquartile ranges for continuous variables and absolute frequencies for categorical variables. Mann Whitney U test (as non-parametric test for two independent samples) and the Fisher exact test were used to compare continuous and categorical variables, respectively. We consider a statistically significant p value <0.05. All the analyses were conducted using STATA software version 13 (Stata Corp, College Station, TX, USA).

This study was approved by the Ethics Committee for Research Projects (Comité de Ética de Proyectos de Inves tigación; Project number 3353) of Hospital Italiano de Buenos Aires. Written informed consent for research purposes is always secured before conducting biopsies at the abovemen tioned institution and we pledged to contact patients whose diagnosis had changed.

Results

Ten out of 19 patients met the criteria for IgG4-RPF with a total score > 20 (median of 30, IQR25-75 28-35, p < 0.001), while 9 were considered non IgG4-RPF, as shown in Figure 1. Among the IgG4-RPF subset, 7/10 were male and the median age at the time of the biopsy was 61.5 years (IQR25-75 55 -64). It was similar in non IgG4-RPF (p = 0.2). Time elapsing from biopsy to revision was 10 years in IgG4-RPF (IQR25-751.3-11.8) and 6 years in non-IgG4- RPF (IQR25-753- 11). Clinical and radiological features of IgG4- related and non-IgG4-RPF are detailed in Table 1.

There were no significant differences between such groups regarding the pattern of comorbidities. Allergic disease did not have predominance among IgG4-RPF patients. No collagen tissue diseases were reported. Three patients featured other manifestations of IgG4-RD in the IgG4-RPF group, including lung septal thickening, lung pseudotumor, mediastinal lymph node enlargement, hypertrophic pachymeningitis and hypophysitis with central insipid diabetes (p = 0.21). Three out of 10 had inflammatory abdominal aortic aneurysms as part of the spectrum of retroperitoneal fibrosis. One patient in the IgG4-RPF group had been diagnosed with tubulointerstitial nephritis according to the biopsy conducted at the onset. Yet, this biopsy was conducted at a different hospital and the IgG4- related origin could not be determined.

Both acute and chronic renal failure and bilateral hydro nephrosis were slightly more common among IgG4-RPF patients but renal failure, hydronephrosis and requirement of renal replacement therapy did not differ among groups statistically. One patient in the IgG4-RPF group was on the national transplant waiting list at the time when data were collected.

The median erythrosedimentation rate at the onset in the IgG4-RPF group was higher than in the non-IgG4- RF-group (66 mm/h versus 31 mm/h, p = 0.14). None of the patients had eosinophilia when IRF was diagnosed. Serum IgG4 levels at the time of diagnosis were not avail able, except for two patients in the IgG4-RPF group with normal measurements.

Thirteen contrast-enhanced computed tomography (IgG4-RPF = 7 - non-IgG4-RPF = 6), 3 computed tomog raphy urography (IgG4-RPF = 2 - non-IgG4-RPF = 1), 8 contrast-enhanced magnetic resonance (IgG4-RPF = 3 - non-IgG4-RPF = 5), 2 magnetic uro-resonance (IgG4- RPF = 1-non-IgG4-RPF = 1), and 5 positron-emission tomography/computed tomography (IgG4-RPF = 3 -non- IgG4-RPF = 2) images were available. All of the IgG4-RPF had infrarenal and circumferential retroperitoneal exten sion along the aorta on different images but this was only significant in statistical terms for periaortic distribution (p = 0.01).

Among IgG4-RPF, corticosteroids were mostly used as drug therapy, followed by azathioprine and rituximab. However, most patients did not receive specific treatment. Eight out of 10 and 7/9 patients needed surgical or endo scopic intervention in both IgG4-RPF and non-IgG4-RPF groups, respectively. Double-J ureteral stenting was the most common invasive procedure, followed by ureterolysis with intraperitonization of the ureter; yet no statistical dif ferences between groups were revealed. Eight patients in each subset were admitted to hospital with a similar median number of hospitalizations (3 in IgG4-RPF versus 2, p = 0.9). No deaths were reported with a median follow-up of 4 years in IgG4-RPF (IQR25-751 - 8) and of 1.5 years in non-IgG4-RPF (IQR25-75 1-2.5).

Table 2 shows histopathological features of IgG4- related and non-IgG4-related retroperitoneal fibrosis. We included 3 needle biopsies in both groups and 7 surgical biopsy specimens in IgG4-RPF and 6 in non-IgG4-RPF subset. The mean biopsy area was greater for IgG4-RPF patients but no statistical differences were recorded (144 mm² versus 73.5 mm², p = 0.21). All IgG4-RPF patients had dense lymphocytic infiltrate (p = 0.21), 8/10 had storiform fibrosis (p = 0.02) and 6/10 had eosinophils (p = 0.17). IgG4+ cells/hpf and IgG4/IgG ratio were sig nificantly higher than in patients with non-IgG4-RPF (p = 0.002 and p = 0.01, respectively) (Fig. 2).

Spearman’s correlation test revealed significant and accurate cor relation between ACR/EULAR score and IgG4/IgG ratio (Rho = 0.68, p = 0.001) or IgG+ cells/hpf (Rho = 0.67, p = 0.001).

We summarize the ACR/EULAR score obtained for each patient in Table 3.

Discussion

Over half of the patients included in our IRF retrospective cohort met criteria of IgG4-RPF. There are no series pub lished in Latin America in this regard, then it is not possible to compare our findings with those from other countries in the region. Compared to worldwide reports, our prevalence in IgG4-RPF gets close to the 57% reported by Khosro shahi et al in 2013 in Boston, where IgG4-RD were identi fied with a IgG4/IgG ratio cutoff >40/HPF⁷. In 2009, Zen et al described a prevalence of 58% identifying IgG4-RD in a Japanese cohort, considering an IgG4/IgG ratio cutoff of >30/HPF⁵. The Danish nationwide study was based on the 2012 consensus statement’s criteria and documented 45% definite or possible IgG4-RF (only 17% with definite diagnosis)⁸. Diagnosis relied on strict histopathologic criteria and different disease definitions, inclusion criteria or ratio cutoffs might hinder the interpretation of results. ACR/ EULAR criteria confirm IgG4-RD diagnosis in patients with lower ratios as increasing points are assigned according to histopathologic and immunohistochemistry findings. They also consider clinical, radiologic and serologic features. In our study, 3 patients meeting the recently-established ACR/EULAR criteria would have been excluded as con firmed cases if only the IgG4/IgG ratio cutoff > 40/HPF criterion was considered.

The diagnosis in retroperitoneum may not be straight forward given the extensive fibrosis which makes the detection of dense cellular infiltrates unlikely. One of our potential sources of bias is the small size of tissue excised for pathology assessment because of their ret rospective acquisition. Most samples were surgical biop sies, but the biopsy area available for histopathological analysis was greater in IgG4-RPF with a non-significant difference in comparison with non-IgG4-RPF. Misclas sification bias may occur if a larger sample could allow us to consider non-IgG4-RPF as IgG4-RD cases. Small biopsies are common and an increased IgG4/IgG ratio has been suggested as the most sensitive pathologic feature, even in the absence of other typical histologic features⁹. New criteria assess not only histopathological or immunohistochemical features, reducing this potential source of bias. Although advanced fibrosis may be pres ent in most cases, a storiform pattern is clear-cut and could be differentiated from hyaline fibrosis in non-IgG4- RPF. Khosroshahi et al and Zen et al described a high proportion of lymphoplasmacytic infiltrate and storiform fibrosis. In our cohort, all of IgG4-RPF showed these core histopathologic features (with a significant difference re garding storiform fibrosis) compared with non-IgG4-RPF. As ACR/EULAR criteria take into account many aspects other than mere histological and immunohistochemical findings, it may contribute to harmonizing the identifica tion of IgG4-RD.

IgG4- RD in our cohort has male predominance, as published in most series. The median age at the time of diagnosis matches international reports. As Khosroshahi’ s cohort, we described extra retroperitoneal manifesta tions in 30% of IgG4-RPF, especially lung involvement. Due to the lack of suspicion, some patients might not have complete evaluation. Retrospective design could have incomplete recordings but each patient had at least abdominal imaging. However, the retroperitoneal fibrosis phenotype has less organic features than other IgG4-RD presentations¹⁰. Infrequent manifestations of IgG4-RD are not included in ACR/EULAR criteria for the purpose of identifying homogeneous populations but they represent key points at the time of clinical evaluation and approach to classification as IgG4-RD cases. In our cohort, those infrequent clinical findings (hypophysitis, pachymeningitis, lymphadenopathies) agreed with the final classification as IgG4-RPF according to ACR/EULAR criteria.

Although there were no significant differences between the subsets, half of IgG4-RPF patients had renal failure at diagnosis, 2 of them required renal replacement therapy and one of them was on the national transplant waiting list at the time when data were collected. We emphasize the early detection of this disease given the natural evolution to end organ damage. One of those patients had been diagnosed with tubulointerstitial nephritis according to biopsy at the onset (IgG4- related origin could not be es tablished because she was treated in another hospital), but in most patients we could not determine if there is intrinsic IgG4- renal involvement beyond obstructive renal failure.

Erythrosedimentation rate (ESR) at the onset was higher among IgG4-RPF but there were not statistical differences whatsoever. We did not have protein reac tive C (PCR) measurement in most patients. Despite the unspecific condition of ESR, inflammatory markers are elevated in IgG4-related periaortitis in contrast with other IgG4 phenotypes¹¹. Nonetheless, inflammatory markers are elevated in IRF and do not help in differentiating id iopathic and IgG4- RD¹². Although ESR and PCR levels may be high at the onset, these parameters should be interpreted carefully because they do not always express disease activity and relapses can occur when acute-phase reactants levels are normal¹³. We did not have serum IgG4 concentrations in most cases at the time of diagno sis, which may reflect lack of knowledge of this entity. In any case, increased serum level of IgG4 is less common in IgG4-RPF than in other IgG4 manifestations¹⁰. In our cohort, we highlight that we achieved a classification of IgG4-RPF even when serum IgG4 concentrations are normal or not measured, which is consistent with the robustness of the classification criteria⁶.

Most patients in our series were diagnosed as IRF be fore 2010. By this time, the awareness of IgG4-RD was low and immunohistochemistry was not usually performed to rule out this diagnosis. There was an erratic management of IRF and most of them did not receive corticosteroids. Beyond medical treatment, this manifestation required invasive treatment, either endoscopic or surgical, in almost every patient in both groups. Advanced fibrosis led to ir reversible anatomic changes requiring those interventions. Due to the biphasic nature of the IgG4-RD, early treatment before a total fibrotic phase may contribute to potential reversibility. In spite of this, subacute presentation tends to delay diagnosis anyway, and a challenge for future research is to elucidate the fibrotic phase of the disease and identify therapeutic targets.

One of the strengths of our study is that we use criteria that consider not only histopathological but also clinical, laboratory and imaging criteria. We agree with current evidence that none of these features taken in isolation are sufficient for either clinical diagnosis or accurate classifica tion of patients. We were unable to find studies related to IgG4-RPF in Latin America, and comparisons cannot be established. As IgG4-RPF is frequent among FRI in our cohort, regional studies related to IgG4-RD are needed.

Regarding the limitations of our work, the number of patients included was rather low compared to worldwide reports, because it is an infrequent manifestation of a low-prevalence disease. Due to the retrospective nature of the study, systematic clinical or serologic data were not available. IgG4-RD was not ruled out in most patients with IRF at the time of diagnosis and they underwent irregular management with no statistically significant dif ferences in terms of evolution. In the light of these results, every retroperitoneal fibrosis, especially circumferential around the infrarenal aorta or even iliac arteries, may be assessed following 2019 ACR/EULAR criteria to confirm IgG4-RD and start corticosteroid therapy. It is relevant to communicate these findings in order to get accurate and early diagnosis because IgG4-RPF may be responsive to prompt treatment.

In conclusion, over half of the patients in our IRF retrospective cohort met the criteria of IgG4-RPF. ACR/ EULAR new criteria may be helpful in order to homogenize the identification of IgG4-RD. These findings may moti vate further evaluation of this condition, increasing early recognition, accurate treatment and integral follow-up.