Retroperitoneal fibrosis (RPF) is characterized by the presence of an aberrant fibro-inflammatory retroperitoneal tissue that usually develops around the infrarenal portion of the abdominal aorta and the iliac arteries.10 Because of its rarity and the lack of diagnostic and classification criteria, RPF has often been considered to embrace a spectrum of diseases of varying etiology; this has led to the inclusion of heterogeneous patient populations in most of the studies carried out so far (especially in urologic and surgical series), which is why most clinical data are unreliable or difficult to interpret. In the last 2 decades, several attempts have been made to define a correct nosology of RPF and, although large-scale systematic studies are still lacking, it has become clear that RPF is to be distinguished in idiopathic and secondary forms, where the latter have clearly identifiable causes such as infections, radiation therapy, drugs, surgery, and cancer.10 This distinction, however, is not always clear-cut, since in some patients these conditions (particularly surgery, radiation, and the use of drugs) may be considered only "predisposing factors," and a direct causal relationship cannot be demonstrated.
The rarity of RPF, the heterogeneity of the conditions included in the previously reported series, and the lack of population-based studies have also limited a correct assessment of the epidemiologic characteristics of the disease, particularly its incidence and prevalence. Additional areas of uncertainty regarding the clinical management of RPF involve the definition of an appropriate diagnostic workup, with particular regard to the role of cross-sectional imaging and retroperitoneal biopsy, and the characterization of the disease phenotype, that is, the complete spectrum of clinical manifestations clinicians must assess when evaluating a patient with RPF. Clarifying these points is a conditio sine qua non for designing future controlled therapeutic trials.
In the current issue of Medicine, the studies by van Bommel et al13 and Scheel et al,7 performed prospectively on large cohorts of patients with the idiopathic form of RPF, address most of the issues outlined above, thus extending our knowledge regarding the clinical characteristics and imaging findings of the disease. Although the 2 studies have been carried out on separate patient cohorts, their results partially overlap, which proves their appropriate patient selection and collection of the clinical data; however, each study also raises some unique points.
Idiopathic RPF has traditionally been considered a very rare condition. Based on the results of a 2004 retrospective, case-control study performed in Finland,8 its incidence was estimated to be 1 × 10−6 person-years and its prevalence 1.4 × 10−5 inhabitants. In that study, however, the cases were retrieved from hospital discharge data using diagnostic codes, a method that can be fallacious particularly in the case of idiopathic RPF, where the diagnosis requires a complex workup and often needs a prolonged follow-up for its final confirmation and for the thorough exclusion of an underlying malignancy. The study by van Bommel et al13 is instead population-based, that is, it considers all RPF cases seen at the Albert Schweitzer Hospital in Dordrecht, The Netherlands, during a study period of 10 years, and precisely reports the patient referral pattern. Using this approach, the annual incidence of idiopathic RPF is shown to be 1.3 × 10−5 inhabitants,13 which is more than 10-fold higher than that reported in the Finnish study.8 Thus, idiopathic RPF is probably much more frequent than it appeared and, given the widespread use of cross-sectional imaging and the growing awareness of this disease among clinicians, it is likely that its incidence will rise in the near future.
Idiopathic RPF usually presents with 2 (often coexisting) types of manifestations: localized and systemic. The former include pain, edema of the lower limbs, scrotal swelling with hydrocele or varicocele, and claudication, and are probably due to the mechanical or compressive effects of the retroperitoneal mass; the latter include fatigue, weight loss, anorexia, and low-grade fever, and probably reflect the systemic inflammatory nature of the disease. Pain is certainly the most common symptom; the proportions of patients presenting with pain in van Bommel's and Scheel's studies are similar and exceed 90%. Pain usually involves the abdomen, the lower back, and the flanks, and often radiates to the inguinal areas. Constipation is another common manifestation. In van Bommel's study,13 the intensity of pain significantly correlated with erythrocyte sedimentation rates, whereas the duration of symptoms was negatively associated with mass thickness.
The 2 papers then focus on the presence of risk factors for the development of idiopathic RPF. Atherosclerotic disease (that is, established coronary, cerebrovascular, or peripheral arterial insufficiency) emerges as a quite common condition among patients with idiopathic RPF, its prevalence being as high as 40% in van Bommel's13 and 20% in Scheel's7 cohort; although these frequencies are considerably high, they do not fully support the previously purported hypothesis that advanced atherosclerosis is the sole causative factor of idiopathic RPF.5 Conversely, it seems more reasonable to think that it can only be a predisposing condition in susceptible hosts.3 Case-control studies with controls matched for age, sex, and geographic area of origin are needed to establish the actual relative risk of developing RPF associated with atherosclerosis or with atherogenic conditions such as hypertension or diabetes mellitus.
In van Bommel's series,13 20% of the patients had been exposed to asbestos, and most of them had pleural plaques as well; in Scheel's cohort,7 12.5% of patients had potential occupational exposure to asbestos. Asbestos exposure has been demonstrated to increase the risk of developing RPF;8 however, the inclusion of RPF occurring in asbestos-exposed patients among the idiopathic and not the secondary forms of the disease seems appropriate, first because it is radiologically and histologically indistinguishable from RPF in non-exposed subjects, second because no clear causal relationship has hitherto been shown, and third because there are no data indicating that the response to therapy is different.
Similarly, Scheel et al7 include in their cohort patients with a history of β-blocker use or radiation therapy. As for the drugs known to be associated with the development of RPF, although a causative role for drugs such as ergot-alkaloids seems to be very likely, evidence supporting a role for commonly used compounds such as β-blockers or analgesics is limited.12 As for the inclusion of patients previously treated with radiation therapy, if the location of radiation is not abdominal and the time elapsed between radiation and RPF particularly long (as was the case for Scheel's patients), a causal relationship seems again unlikely, and therefore these forms should not be regarded as secondary.
Altogether, these data lend support to the hypothesis that many conditions act as "inciting" factors for the development of RPF, but that only some of them induce "secondary" RPF; this is the case for primary or metastatic retroperitoneal neoplasms,2 infections such as tuberculosis,1,4 multisystemic disorders such as Erdheim-Chester disease,9 and the use of drugs for which compelling evidence exists. It would be a mistake to put such cases into a study of idiopathic RPF.
The studies by van Bommel et al13 and Scheel et al7 also provide insightful data on the cross-sectional imaging (computed tomography [CT] and magnetic resonance imaging [MRI]) findings characterizing idiopathic RPF: both show that the typical presentation of idiopathic RPF is that of a periaortic, homogeneous muscle-isodense tissue, which usually surrounds the anterior and lateral sides of the infrarenal abdominal aorta and iliac arteries but does not cause aortic displacement. These data are useful for differential diagnosis with other retroperitoneal lesions (particularly malignancies), which often displace the aorta and show inhomogeneous density in both MRI and CT studies.12 According to the view of Scheel et al, the presence of periaortic/periiliac disease should be considered a prerequisite for the diagnosis of idiopathic RPF; this criterion is more stringent than that adopted in the study by van Bommel et al, which also included a few cases without periaortic involvement (for example, paracolic, prevesical RPF). Although following Scheel's view, one might exclude patients with true but atypically localized idiopathic RPF, this criterion should be considered in prospective trials in order to have homogeneous patient populations. Additionally, it is likely that-despite similar histopathologic features-the pathogenesis of RPF differs in cases with and without periaortic disease: idiopathic RPF has indeed been considered to arise as a primary aortitis that subsequently extends into the surrounding retroperitoneum, which obviously cannot be possible for RPF without perivascular involvement. The presence of concomitant inflammatory involvement of the thoracic aorta in a subset of patients is in line with this view.6
Interestingly, van Bommel et al13 also point out that localized lymphadenopathy, found in close proximity to the retroperitoneal mass, is quite frequent (25% of cases), and should not confuse the diagnosis. In addition, both papers show that renal vessel involvement may be an overlooked feature of RPF, and that in some cases it may contribute to renal insufficiency or cause reno-vascular hypertension, thus requiring specific interventional approaches.7
Finally, Scheel et al7 propose for the first time a classification of the disease into 4 different classes on the basis of the disease extent assessed using CT or MRI: class I is hallmarked by periaortic and/or periiliac involvement; class II consists of class I plus pericaval involvement, which could be associated with an increased risk of thromboembolic events; class III corresponds to class I plus ureteral involvement; and class IV is class I plus compression of the renal vessels. This scheme is interesting, and needs to be corroborated by data showing that the different classes have different outcomes or different complications and therefore need tailored treatment approaches. In addition, the scheme could be used to stratify idiopathic RPF patients on the basis of their disease extent in future controlled trials.
Some limitations of the 2 studies must, however, be acknowledged; although they have been performed on large patient cohorts and the data collected in a prospective fashion, they can only give a flash of the presentation of idiopathic RPF, but cannot demonstrate what the dynamic course of the disease involves over time. It is known, for instance, that autoimmune conditions can be found in association with RPF in a number of cases. The 2 studies show that, at presentation, only a small percentage of patients have autoimmune diseases involving other organs, but since these often become overt late in the course of RPF,11 their true frequency cannot be established at RPF presentation. In addition, neither of the 2 studies systematically performed cross-sectional imaging of the chest, which is instead useful given the coexistence of thoracic periaortitis or mediastinal fibrosis in a number of idiopathic RPF patients;6 more importantly, chest CT or MRI may also allow the unexpected detection of lung or mediastinal infectious lesions such as tuberculosis,1,4 to which RPF may be secondary. These aspects need to be considered for future studies.
In conclusion, van Bommel et al13 and Scheel et al7 should be commended for the important findings provided by their papers, which extend our knowledge of RPF and pave the way for future, large-scale, controlled studies on this disease.
I would like to thank Professor Carlo Buzio for his insightful analysis of this commentary and Dr. Alessandra Palmisano for her valuable collaboration in research and clinical management of patients with RPF.
1. Greco P, Vaglio A, Corradi D, Cobelli R, Zompatori M, Buzio C. Tuberculosis as a trigger of retroperitoneal fibrosis. Clin Infect Dis
2. Koep L, Zuidema GD. The clinical significance of retroperitoneal fibrosis. Surgery
3. Martorana D, Vaglio A, Greco P, Zanetti A, Moroni G, Salvarani C, Savi M, Buzio C, Neri TM. Chronic periaortitis and HLA-DRB1*03: another clue to an autoimmune origin. Arthritis Rheum
4. Molloy CB, Filer C, Ismail A. Mycobacterium tuberculosis as a cause of chronic periaortitis. Rheumatology (Oxford)
5. Parums DV. The spectrum of chronic periaortitis. Histopathology
6. Salvarani C, Pipitone N, Versari A, Vaglio A, Serafini D, Bajocchi G, Salvo D, Buzio C, Greco P, Boiardi L. Positron emission tomography (PET): evaluation of chronic periaortitis. Arthritis Rheum
7. Scheel PJ Jr, Feeley N. Retroperitoneal fibrosis: the clinical, laboratory, and radiographic presentation. Medicine (Baltimore)
8. Uibu T, Oksa P, Auvinen A, Honkanen E, Metsarinne K, Saha H, Uitti J, Roto P. Asbestos exposure as a risk factor for retroperitoneal fibrosis. Lancet
9. Vaglio A, Corradi D, Maestri R, Callegari S, Buzio C, Salvarani C. Pericarditis heralding Erdheim-Chester disease. Circulation
10. Vaglio A, Palmisano A, Corradi D, Salvarani C, Buzio C. Retroperitoneal fibrosis: evolving concepts. Rheum Dis Clin North Am
. 2007;33:803-817, vi-vii.
11. Vaglio A, Palmisano A, Ferretti S, Alberici F, Casazza I, Salvarani C, Buzio C. Peripheral inflammatory arthritis in patients with chronic periaortitis: report of five cases and review of the literature. Rheumatology (Oxford)
12. Vaglio A, Salvarani C, Buzio C. Retroperitoneal fibrosis. Lancet
13. Van Bommel EFH, Jansen I, Hendriksz TR, Aarnoudse ALHJ. Idiopathic retroperitoneal fibrosis: prospective evaluation of incidence and clinicoradiologic presentation. Medicine (Baltimore)