Do acute-phase reactants predict response to glucocorticoid therapy in retroperitoneal fibrosis?




To determine the value of the erythrocyte sedimentation rate (ESR) and the C-reactive protein (CRP) level at the time of diagnosis for predicting radiographic response to glucocorticoid therapy in patients with retroperitoneal fibrosis (RFP).


Data were collected retrospectively for 37 patients with an established diagnosis of RFP (the diagnosis was proven by biopsy in 31 patients), all of whom met the following inclusion criteria: 1) availability of a recorded baseline ESR and/or CRP level and results of computed tomography or magnetic resonance imaging, 2) availability of followup CRP level and/or ESR with radiographic imaging 12–24 weeks after initiation of therapy, and 3) treatment with prednisone monotherapy at a starting dosage of 40–60 mg daily. Patients were divided into 2 therapeutic response groups: group 1 showed radiographic regression, and group 2 showed no change or radiographic progression. Any progression or regression was determined by an estimated change of ≥25%.


The median baseline CRP levels were 2.2 mg/dl (interquartile range [IQR] 1.4–8.0) in group 1 and 1.2 mg/dl (IQR 0.8–4.1) in group 2 (P = 0.35). The median baseline ESR in group 1 was 57.5 mm/hour (IQR 39.2–102.5), which was not statistically different from the median ESR in group 2 (58 mm/hour [IQR 33–66]). The mean CRP level and ESR tended to be higher in patients with radiographic regression, but these differences failed to reach statistical significance. Spearman's correlation coefficient revealed no correlation between the baseline CRP level (r = −0.11, P = 0.51) or ESR (r = −0.06, P = 0.71) and the radiographic response.


The ESR and CRP level at baseline are poor predictors of a therapeutic response to glucocorticoid therapy in patients with RPF.


Retroperitoneal fibrosis (RPF) is a rare disease characterized by a fibroinflammatory retroperitoneal mass, with an anatomic predisposition for the periaortic areas (1–4). In most cases the disease is idiopathic, but it may also arise secondary to neoplasms, infections, or radiation injury, or in response to treatment with certain drugs. Idiopathic RPF can be broadly divided into 2 categories: nonaneurysmal fibrosis and perianeurysmal fibrosis. Patients with RPF may be asymptomatic if the mass does not impair critical organ function, but it often entraps the ureters and other abdominal organs, resulting in end-organ damage including renal impairment (acute renal failure) and vascular and neurologic compromise (back pain, neuritic pain, claudication, and venous thrombosis).

Although the exact pathogenesis of RPF is poorly understood, it has been thought to be a localized process in response to advanced atherosclerosis (5, 6). It is often associated with elevation of the erythrocyte sedimentation rate (ESR) and the serum C-reactive protein (CRP) level and with constitutional features such as anemia, fever, anorexia, and weight loss. Some studies have shown an association of RPF with other autoimmune diseases and autoantibodies (7–10). Histologically, the disease has 2 stages: an active cellular stage with predominant inflammation, and an inactive fibrotic stage with a paucity of inflammatory infiltrates (11–13).

The general therapeutic approach to RPF is not standardized except for the necessity for prompt ureteral decompression in case of ureteral obstruction. However, such measures do not affect the underlying pathologic process or the potential for continued anatomic spread and systemic effects. Treatment with glucocorticoids is widely used, and tamoxifen therapy has been advocated, based on observational studies that have demonstrated variable effects on clinical signs and symptoms, acute-phase reactants, and mass size (14–16). Treatment with the combination of prednisone and cyclophosphamide or azathioprine has resulted in the resolution of symptoms in some studies (17, 18). Recently, single cases and case series demonstrating the efficacy of mycophenolate mofetil in combination with prednisone have been reported (19, 20).

At present, there are no validated predictors of response to glucocorticoids. Gallium scanning has been proposed to help in predicting such a response (21), but a recent study showed that such scanning failed to discriminate between responders and nonresponders to prednisone (15). There is a general belief that an elevated ESR and an elevated CRP level at the time of diagnosis reflect active inflammation, indicating that glucocorticoid therapy might be effective. This hypothesis remains largely untested. We reviewed a large series of patients with RPF who were treated in a standardized manner with prednisone monotherapy and frequently with surgery, in an attempt to determine whether the elevated ESR or the CRP level at the time of diagnosis predicts a response to initial glucocorticoid therapy. In addition, we examined the clinical characteristics and laboratory abnormalities in this RPF cohort and assessed the frequency of comorbidities and autoantibodies.


This study was approved by the Institutional Review Board of the Cleveland Clinic. Sixty-three patients with an established diagnosis of RPF were treated in the Department of Rheumatology from 1990 to 2006. Their medical records were reviewed, and data were abstracted by 2 reviewers using a standardized data collection form. The inclusion criteria were as follows: 1) results of a baseline ESR or CRP level determined before the initiation of glucocorticoid therapy; 2) available results of baseline computed tomography (CT) or magnetic resonance imaging (MRI); 3) results of a followup CRP level or ESR in addition to results of radiographic imaging at 12–24 weeks after corticosteroid therapy; and 4) treatment with a stable dose of prednisone (40–60 mg) for at least 1 month. All patients with a history of malignancy were excluded.

Data collected.

Demographic information, clinical information, and laboratory data were recorded for each patient, including age, sex, current smoking, associated comorbidities such as coronary artery disease, hypertension, diabetes, and hepatitis C, and the presence of autoimmune diseases including rheumatoid arthritis, Graves' disease, systemic sclerosis, and sclerosing cholangitis. Clinical symptoms at both baseline and followup were recorded. Laboratory data included the baseline CRP level, ESR, complete blood cell count, creatinine level, and autoantibodies (including antinuclear antibodies [ANAs], rheumatoid factor [RF], and antineutrophil cytoplasmic antibodies [ANCAs]), as well as the followup ESR and CRP level.

Imaging data.

Results of the imaging studies (CT scan or MRI) were reviewed for verification of the size of the mass and serial changes at followup after 12–24 weeks. These were combined with the opinion of the treating rheumatologist regarding the radiograph documented in the chart. The primary outcome was determined by the change in the size of the retroperitoneal mass as observed radiographically. The patients were divided into 2 therapeutic response groups: group 1 had radiographic regression, and group 2 had no change or radiographic progression. An estimated change of ≥25% determined any progression or regression.

Statistical analysis.

Data summaries for categorical data are shown as frequencies and percentages, and quantitative data are expressed as the mean ± SD or the median (interquartile range [IQR]). Categorical data for patients with and those without aneurysms were compared using Fisher's exact test, and quantitative data were compared using Wilcoxon's rank sum test. Changes from baseline to followup in categorical variables were assessed using McNemar's test, and changes in continuous data were assessed using Wilcoxon's signed rank test. For quantitative data, tests of association were derived from Spearman's correlation coefficients.


Of the 63 patients, 37 met the inclusion criteria. In 31 of those 37 patients, the diagnosis was confirmed by biopsy. The demographic and clinical characteristics of the 37 patients are shown in Table 1. As reported in previous series (8, 13, 15, 22, 23), the majority of the patients were men and were smokers, with a mean age at diagnosis of 58.4 years. The most frequent comorbidity was hypertension. Associated autoimmune diseases were rare. Seven patients (19%) had aortic aneurysms with perianeurysmal fibrosis; the mean ± SD age of this group of patients was 65.3 ± 6.0 years, which was greater than that of the group without aneurysms. All 7 of these patients had hypertension.

Table 1. Demographics and associated comorbidities in patients with retroperitoneal fibrosis (RPF)*
VariableAll patients (n = 37)Nonaneurysmal RPF (n = 30)Aneurysmal RPF (n = 7)P
  • *

    Except where indicated otherwise, values are the percent of patients.

  • Differences in the aneurysmal and nonaneurysmal groups were calculated using Fisher's exact test, except differences for age, which were determined using Wilcoxon's rank sum test.

Mean ± SD age, years58.4 ± 12.456.2 ± 12.965.3 ± 6.00.07
Male sex64.958.685.70.38
Coronary artery disease21.613.842.90.12
Diabetes mellitus17.614.314.31
Hepatitis C virus5.67.101
Rheumatoid arthritis2.73.401
Systemic sclerosis2.73.401
Sclerosing cholangitis2.70140.2
Graves' disease2.73.401

Symptoms at onset were similar to those reported in previous series, with abdominal pain and weight loss being the most common (Table 2). At followup, there was a significant improvement in all symptoms except constipation and claudication (Tables 1 and 2).

Table 2. Clinical symptoms at baseline and followup*
  • *

    Values are the number (percentage) of patients. At baseline, testicular pain was assessed in only 35 of the 37 patients. At followup, abdominal pain and leg edema were assessed in 36 patients; weight loss, urinary frequency, constipation, and claudication were assessed in 35 patients, and testicular pain was assessed in 34 patients.

Abdominal pain29 (75.7)7 (19.45)< 0.001
Weight loss17 (45.9)2 (5.7)< 0.001
Leg edema10 (27.0)5 (13.9)0.031
Testicular pain6 (17.1)1 (2.9)0.031
Urinary frequency5 (13.5)1 (2.9)0.06
Constipation3 (8.1)0 (0)0.50
Claudication3 (8.1)2 (5.7)1.00

Laboratory abnormalities at baseline included mild anemia (mean hemoglobin ± SD 11.54 ± 1.84 gm/dl) and elevated creatinine levels (1.59 ± 1.46 mg/dl). The median CRP level and ESR at baseline were 1.9 mg/dl (IQR 0.8–4.4) and 58 mm/hour (IQR 34–90), respectively.

The frequency of autoantibodies was low. Five (23.8%) of 21 patients tested were positive for RF, 4 (25.0%) of 16 patients were positive for ANAs, and all 4 of the patients tested for ANCAs had negative results.

Following treatment, the retroperitoneal mass regressed in 54.1% of the patients, and median CRP levels and ESR significantly decreased to 0.5 mg/dl (IQR 0.3–0.8) and 16 mm/hour (IQR 10.0–24.8), respectively (P < 0.001) (Figure 1).

Figure 1.

Change in erythrocyte sedimentation rate (ESR; mm/hour) and C-reactive protein (CRP; mg/dl) level with glucocorticoid monotherapy.

The CRP levels and ESRs at the time of diagnosis and at followup are shown in Table 3. The median baseline CRP levels were 2.2 mg/dl (IQR 1.4–8.0) in group 1 and 1.2 mg/dl (IQR 0.8–4.1) in group 2 (P = 0.35). The median baseline ESRs were 57.5 mm/hour (IQR 39.2–102.5) in group 1 and 58 mm/hour (IQR 33–66) in group 2. Both the mean CRP levels and the ESR tended to be higher in patients with radiographic regression, but these differences failed to reach statistical significance, regardless of whether the distributions were compared by Wilcoxon's rank sum test (P = 0.35 and P = 0.34, respectively) or whether the mean values of the log-transformed data were compared using a t-test (P = 0.35 and P = 0.33, respectively) (Table 3).

Table 3. CRP value and ESR at baseline and followup in the 2 groups*
VariableAll patients (n = 37)Outcome
Group 1 (n = 20)Group 2 (n = 17)PTest
  • *

    CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; IQR = interquartile range; W = Wilcoxon's rank sum test; T = t-test.

CRP, mg/dln = 34n = 18n = 16  
 Mean ± SD4.31 ± 5.925.56 ± 7.412.91 ± 3.30  
 Median (IQR)1.9 (0.8–4.4)2.2 (1.4–8.0)1.2 (0.8–4.1)0.35W
Log2(CRP)n = 34n = 18n = 16  
 Mean ± SD mg/dl1.07 ± 1.731.33 ± 1.910.77 ± 1.510.35T
CRP at followup, mg/dln = 34 n = 14  
 Mean ± SD0.91 ± 1.761.16 ± 2.270.56 ± 0.30  
 Median (IQR)0.5 (0.3–0.8)0.5 (0.3–1.0)0.5 (0.3–0.8)0.62W
Log2(CRP) at followupn = 34 n = 14  
 Mean ± SD mg/dl−0.86 ± 1.19−0.75 ± 1.43−1.02 ± 0.730.47T
ESR, mm/hour     
 Mean ± SD62.1 ± 34.669.0 ± 39.154.0 ± 27.2  
 Median (IQR)58 (34–90)57.5 (39.2–102.5)58 (33–66)0.34W
Log2(ESR), mean ± SD mm/hour5.69 ± 0.965.83 ± 0.995.52 ± 0.930.33T
ESR at followup, mm/hourn = 32n = 18n = 14  
 Mean ± SD18.72 ± 12.5318.83 ± 10.2818.57 ± 15.38  
 Median (IQR)16 (10–24.8)17.5 (11–26.2)14 (9.2–21.5)0.61W
Log2(ESR) at followupn = 32n = 18n = 14  
 Mean ± SD mm/hour3.89 ± 1.074.00 ± 0.903.76 ± 1.270.55T

Using Spearman's correlation coefficient, there was no correlation between the baseline CRP level (r = −0.11, P = 0.51) or ESR (r = −0.06, P = 0.71) and the radiographic response of the mass, as shown in Figures 2 and 3.

Figure 2.

Correlation between baseline C-reactive protein (CRP) values and radiographic outcome. Boxplots demonstrate selected percentiles of the values: outer lines are drawn at the 10th and 90th percentiles, and the center box extends from the 25th to the 75th percentiles, with a central line drawn at the median. CT = computed tomography; MRI = magnetic resonance imaging.

Figure 3.

Correlation between baseline erythrocyte sedimentation rate (ESR) and radiographic outcome. Boxplots demonstrate selected percentiles of the values: outer lines are drawn at the 10th and 90th percentiles, and the center box extends from the 25th to the 75th percentiles, with a central line drawn at the median. CT = computed tomography; MRI = magnetic resonance imaging.


RPF continues to be a poorly understood disorder; there are gaping holes in our understanding of its pathogenesis, and the guidelines for its therapy are based largely on uncontrolled observations. Although the present study was uncontrolled, it provided 2 observations that may add to our understanding of the disease. First, the study showed a lack of correlation between the baseline levels of acute-phase reactants and the clinical response to glucocorticoid therapy. Second, there was a paucity of associated autoimmune disorders and autoantibodies.

There was no significant difference between the baseline CRP level and ESR between patients whose mass decreased in size during treatment and those in whom the mass did not decrease. There was a trend toward higher CRP and ESR values in patients with radiographic regression of the mass; such trends are inconclusive, however, because the degree of differences observed could simply be a result of chance. The results were contrary to our prior assumption that patients with early active RFP have predominant inflammation resulting in elevated ESRs and CRP levels and hence will respond better to the corticosteroid therapy.

Prednisone monotherapy (40–60 mg/day) used for 3–6 months provided symptomatic relief as well as regression of the mass in approximately half of the patients. This is comparable with results of other series that showed a similar success rate with prednisone monotherapy (14). However, we did not study the response to steroid monotherapy over a prolonged period of time or whether the effect was sustained.

It has been proposed that RPF is a systemic autoimmune disease that frequently is associated with other autoimmune diseases and autoantibody positivity (7–10). However, our study showed a low frequency of associated autoimmune diseases (3%), increased antibody positivity (ANA positivity 24%), and elevated RF levels (23%). These results are similar to those recently reported in which the frequency of ANA was 37% (15).

Our study had some limitations that are inherent to a retrospective analysis. The evaluations were not uniform, nor was measurement of the retroperitoneal masses. The patients studied were seen over a period of 15 years. For many patients, serial radiographic assessments were not performed at our institution, and therefore, the films were no longer available for adjudication. Also, other causes of an elevated ESR or CRP concentration were not explored. Nonetheless, our cohort represents one of the largest data sets for this rare disease.

In conclusion, determination of acute-phase reactants at baseline, using the ESR and the CRP level, is a poor predictor of a therapeutic response to corticosteroid therapy in patients with RPF. It would be reasonable to consider initial medical therapy with glucocorticoids for 3–6 months in all patients with RFP, regardless of their baseline levels of markers of inflammation. The beneficial effect may be explained by the inhibitory effect of glucocorticoids on fibroblastic proliferation. However, further prospective studies are needed to assess the utility of markers of inflammation as predictors of a response to steroid therapy in patients with RFP.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Magrey had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Magrey, Husni, Kushner, Calabrese.

Acquisition of data. Magrey, Husni, Kushner, Calabrese.

Analysis and interpretation of data. Magrey, Husni, Kushner, Calabrese.