- Email: [email protected]
Long-term TNF-α Blockade in Patients with Amyloid A Amyloidosis Complicating Rheumatic Diseases
CLINICAL RESEARCH STUDY
Long-term TNF-␣ Blockade in Patients with Amyloid A Amyloidosis Complicating Rheumatic Diseases Antonio Fernández-Nebro, MD, PhD,a Alejandro Olivé, MD, PhD,b María Carmen Castro, MD, PhD,c Angela Herranz Varela, MD,d Elena Riera, MD,b Maria V. Irigoyen, MD,a María Jesús García de Yébenes, MD, PhD,d Rosario García-Vicuña, MD, PhDe a Servicio de Reumatología y Departamento de Medicina de la Facultad de Medicina, Hospital Regional Universitario Carlos Haya, Málaga, Spain; bRheumatology Section, Hospital Universitario Germans Trias i Pujol, Badalona, Barcelona, Spain; c Servicio de Reumatología, Hospital Universitario Reina Sofía, Córdoba, Spain; dUnidad de Investigación, Fundación Española de Reumatología, Madrid, Spain; eRheumatology Service, Hospital Universitario de la Princesa, Universidad Autónoma de Madrid, Madrid, Spain.
ABSTRACT OBJECTIVE: To evaluate the effectiveness and safety of anti-tumor necrosis factor therapy in patients with amyloid A amyloidosis. METHODS: Multicenter, controlled, dynamic prospective cohort study of 36 patients with amyloid A amyloidosis (94% kidney involvement) treated with anti-tumor necrosis factor agents (drug exposure of 102.97 patient-years). As an external control group, 35 propensity score-matched non-amyloid patients were chosen from the Base de Datos de Productos Biológicos de la Sociedad Española de Reumatología registry. The end points were kidney response and progression, anti-tumor necrosis factor continuation rate, patient survival, and adverse events. RESULTS: At the end of follow-up, a kidney response was observed in 12 of 22 patients (54.5%) and a kidney progression was observed in 6 of 36 patients (17%). The kidney amyloidosis remained stable in 16 of 36 patients (44%). The level of acute phase reactants diminished but did not reach the normal level. The continuation rates of anti-tumor necrosis factor drugs among patients with amyloid A amyloidosis after 1, 2, 3, and 4 or more years were 80%, 80%, 61%, and 52%, respectively, comparable to controls. The 5-year cumulative survival of amyloid A amyloidosis cases was 90.6%, and the 10-year survival was 78.5%. In a multivariate Cox regression analysis, the duration of amyloidosis and the level of proteinuria at the onset of anti-tumor necrosis factor treatment were independent predictors of treatment failure, whereas the level of proteinuria was the only factor that predicts mortality. Most adverse events were similar in both groups, although the number of infections was 3 times higher in amyloid A amyloidosis cases. CONCLUSION: Anti-tumor necrosis factor drugs are effective in treating amyloid A amyloidosis, although they might increase the risk of infection. © 2010 Elsevier Inc. All rights reserved. • The American Journal of Medicine (2010) 123, 454-461 KEYWORDS: Amyloid A amyloidosis; Anti-tumor necrosis factor-alpha blockade; Effectiveness; Immune-mediated inflammatory diseases; Infections
Funding: None. The drugs administered to our patients were supplied by the sanitary service as part of the medical care that they normally received. Conflict of Interest: See last page of article. Authorship: See last page of article.
0002-9343/$ -see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2009.11.010
Reprint requests should be addressed to Antonio Fernández-Nebro, MD, PhD, Servicio de Reumatología, Hospital Regional Universitario Carlos Haya, Pabellón C: “Hospital Civil,” Plaza del Hospital Civil s/n, 29009, Málaga, Spain. E-mail address: [email protected]
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
455
The discovery that clinically unrelated conditions, such as biological therapy was not officially accepted for the underrheumatoid arthritis and Crohn’s disease, share similar allying disease. terations in the immune system has led to a shift in the management of immune-mediated inflammatory diseases External Controls from organ-based symptom relief to mechanism-based Controls were extracted from the Base de Datos de Protreatment. Currently, anti-tumor ductos Biológicos de la Sociedad necrosis factor-␣ drugs offer a Española de Reumatología regissuitable treatment for many imtry (BIOBADASER), a detailed CLINICAL SIGNIFICANCE mune-mediated inflammatory disdescription of which has been pubeases refractory to nonbiological lished5 and is available at http:// ● Anti-tumor necrosis factor drugs may be disease-modifying antirheumatic biobadaser.ser.es/. Propensity score useful to treat kidney amyloidosis but drugs because of their efficacy and methods were applied to match pacould increase the risk and severity of rapid onset of action and tolerabiltients and controls according to sevinfections. ity. These treatments not only reeral baseline characteristics related ● Patient survival depends on the degree duce inflammatory activity but to amyloid A amyloidosis and inalso reduce the complications and of proteinuria at the beginning of antifection susceptibility:6 age, gender, excessive mortality associated tumor necrosis factor drug treatment. type and duration of rheumatic with some of these diseases.1,2 disease, and steroid use. First, ● The continuation rates with anti-tumor Furthermore, there is evidence patients with amyloidosis were necrosis factor drugs are similar in pathat these anti-tumor necrosis facexcluded from BIOBADASER tients with amyloid A amyloidosis and tor drugs may rapidly improve the controls. The matching variables controls without amyloidosis, although clinical status of patients with were then included in a multivarboth delay of treatment after the diagamyloid A amyloidosis secondary iate logistic regression to calcuto immune-mediated inflammanosis of amyloidosis and a high-grade late a predictive probability (protory diseases.3,4 Nevertheless, it proteinuria were associated with premapensity score) of both groups (with still remains unclear whether the and without amyloidosis) using ture interruption of treatment. effects of such therapies are susSTATA-specific commands. Fitained in the long-term or if they nally, this propensity score was reduce the mortality associated used to match amyloidosis cases with this severe complication. Indeed, because anti-tumor with the single nearest-neighbor controls without amynecrosis factor drugs also increase the incidence of serious loidosis in the BIOBADASER subsample.7 infections, especially in patients with associated comorbidity,2 it is important to determine whether any potential Follow-up and Amyloidosis Treatment reduction in mortality due to amyloidosis could be underProtocol mined by an increase in the mortality due to severe The following data were registered for each case: gender, infections. age, underlying inflammatory disease, time of evolution of Thus, as well as investigating the long-term effectiveness the underlying disease and of amyloidosis, diagnostic test of anti-tumor necrosis factor-␣ therapies in amyloid A amyused for amyloidosis, interstitial fibrosis in the kidney biloidosis associated with immune-mediated inflammatory opsy, organs affected, anti-tumor necrosis factor drug dose diseases, an important aim of this study was to assess the and time of use, cause of withdrawal, concomitant treatincidence of severe adverse events, especially infections, ments, adverse events, vital status, urinary protein excretion and their impact on patient mortality when compared with a in 24 hours, serum creatinine, creatinine clearance in 24 propensity score-matched population free of amyloidosis. hours, C-reactive protein, and erythrocyte sedimentation
PATIENTS AND METHODS Patients The patients included in this study were all patients with amyloid A amyloidosis secondary to an immune-mediated inflammatory disease who were treated and followed up prospectively in 4 Spanish hospitals from January 2001 to October 2006. This study represents an extension of a previous study.4 Amyloid A amyloidosis was confirmed histologically by Congo-red staining and immunohistochemistry. Informed consent was obtained from all patients, and the treatment was approved by the local health administration if
rate at baseline and periodically during the follow-up. The data were recorded at the visits scheduled at months 0, 3, 6, and 12, and then annually. All patients were screened, and if necessary, they were treated for active or latent tuberculosis following the recommendations of the European and Spanish Drug agencies.8 The anti-tumor necrosis factor drugs (infliximab, etanercept, or adalimumab) were chosen by each physician and initially administered according to the technical data sheets. There were no restrictions on the use of disease-modifying antirheumatic drugs, steroids, or supportive treatments. The treatments were readjusted or interrupted by the clinician according to the therapeutic response or the appearance of adverse events.
456
Evaluation Criteria All patients who received at least 1 dose of anti-tumor necrosis factor were evaluated. Treatment effectiveness was evaluated by the improvement of kidney outcomes (kidney response and progression), continuation rates of anti-tumor necrosis factor drugs, and the survival rate of patients with amyloidosis. Kidney outcomes were defined according to established consensus criteria.9 Kidney response was defined as follows: 50% or greater decrease (ⱖ0.5 g/d) of 24-hour urine protein excretion (proteinuria ⬎ 0.5 g/d pretreatment), and serum creatinine and creatinine clearance must not worsen by 25% above the baseline. Kidney progression was defined by a 50% increase in proteinuria (ⱖ1 g/d) of proteinuria to greater than 1 g/d (proteinuria must be ⬎ 0.5 g/d pretreatment) or 25% worsening of serum creatinine (minimum of 0.5 mg/dL) or creatinine clearance. Patients who do not fulfill the criteria for progressive disease or responsive disease are considered stable. Safety was evaluated through the relevant adverse events as assessed through physical examination and patient self-reporting at each visit. Our definition of a relevant adverse event was the same as that described in BIOBADASER.5
Statistical Analysis Intention-to-treat analysis was used. Incidence date was the first date of anti-tumor necrosis factor therapy after the diagnosis of amyloidosis. Quantitative descriptive data are presented as the mean (⫾ standard deviation), and the rates were expressed per patient-years (95% confidence interval). Normality was confirmed by the Kolmogorov test. For skewed data, the median (⫾ interquartile range, interquartile range) was used. The chi-square or Fisher exact 2-tailed test was used to compare categories. The efficacy in terms of proteinuria, renal function, and acute phase reactants at 0, 3, 6, 12, and 24 months was assessed by repeated-measures analysis of variance (Friedman test was used to detect differences in the proteinuria). Kaplan–Meier curves and the log-rank test were used to estimate the time to withdrawal of anti-tumor necrosis factor drugs and to compare the survival functions between cases and controls, respectively. Survival time was measured from the date of amyloidosis diagnosis to the date of death. A Cox regression was used to identify the prognostic factors of time to withdrawal of anti-tumor necrosis factor drug and death in the univariate and multivariate analyses (stepwise forward/Wald). All variables that reached a value of P less than .10 were included in the Cox multivariate model. A logistic regression analysis was used to identify the prognostic factors of kidney outcomes.
RESULTS From January 1, 2001, to October 31, 2006, a cohort of 36 patients with amyloid A amyloidosis was generated. The patients were treated with anti-tumor necrosis factor drugs for a total of 102.97 patient-years (median follow-up 34.6 months). Thirty-five propensity score-matched control subjects were chosen from the BIOBADASER registry. All
The American Journal of Medicine, Vol 123, No 5, May 2010 patients had highly active immune-mediated inflammatory diseases refractory to other treatments.
Patient Characteristics The baseline clinical status of the 36 patients and that of the controls was similar (Table 1). The amyloid A amyloidosis had been diagnosed 2.7 ⫾ 3.4 years (range, 0-13) before treatment started. Three of the cases were being treated by infliximab when they were diagnosed with amyloid A amyloidosis (one of whom had been treated for 29 months), whereas the remaining patients had not previously received biological therapy. Infliximab was the drug of first choice. The dose of infliximab was increased (from an initial mean of 3.6 mg/kg ⫾ 0.7 to 4.0 mg/kg ⫾ 0.9) or the interval between doses was reduced (from every 8 to every 6 weeks) in 13 of 29 cases (45%). Only in 2 cases did the interval between doses increase (to every 9 weeks and to every 11 weeks). Nevertheless, the majority of the subjects with amyloidosis continued to receive infliximab every 8 weeks. The dose and the interval of etanercept and adalimumab administration did not change in any patient. One of the cases of amyloidosis diagnosed with adult-onset Still disease received anakinra (100 mg/d) as a third option. Baseline clinical data related to amyloidosis diagnosis were collected (Table 2), and 56 biopsies were performed at 7 distinct locations to histologically confirm amyloidosis. The presence of a clinical parameter associated with altered kidney function was the initial indication of amyloidosis in most cases.
Effectiveness of Anti-Tumor Necrosis Factor against Amyloid A Amyloidosis The anti-tumor necrosis factor treatment reduced the median levels of proteinuria by 59.7% during the first 24 months of the follow-up period (median 0.67 g/d [interquartile range ⫽ 0.24-3.24] vs 0.40 g/d [interquartile range ⫽ 0.14-1.84]; Friedman test P ⬍ .001) (Figure 1). Both the mean serum creatinine and creatinine clearance remained stable. At the end of follow-up, kidney response was observed in 12 of 22 patients (54.5%), and kidney progression was observed in 6 of 36 patients (17%). Kidney amyloidosis remained stable in 16 of 36 patients (44%). The level of acute phase reactants diminished quickly but did not reach the normal level (Figure 2). The mean erythrocyte sedimentation rate decreased during the first 24 months from 70.7 mm/h (95% confidence interval (CI), 60.3-81.3) to 51.4 mm/h (95% CI, 39.7-63.1, F test P ⫽ .028), whereas the levels of C-reactive protein decreased from 39.5 mg/L (95% CI, 27.7-51.3) to 20.3 mg/L (95% CI, 12.7-27.9; F test P ⬍ .001). At the end of the follow-up period, normal values for the erythrocyte sedimentation rate and C-reactive protein were only evident in 7 patients (19%) and 9 patients (25%), respectively. The cumulative incidence of patients with amyloidosis who still received anti-tumor necrosis factor treatment after 1, 2, 3, and 4 or more years were 80%, 80%, 61%, and 52%,
Fernández-Nebro et al Table 1
Anti-TNF Drugs in Amyloid A Amyloidosis
457
Baseline Demographic and Clinical Characteristics of Cases and Controls
Age, y Women, n (%) Duration of disease, y Duration of follow-up, y Patient-y Underlying disease, n (%) Rheumatoid arthritis Spondyloarthropathy Psoriatic arthritis Juvenile idiopathic arthritis Adult-onset Still disease Antiphospholipid syndrome Anti-TNF treatment, n (%)c Infliximab Etanercept Adalimumab Concomitant disease-modifying antirheumatic drugs, n (%)b No disease-modifying antirheumatic drugs Methotrexate Leflunomide Chlorambucil Sulfasalazine Hydroxychloroquine Corticosteroids, n (%) No. of biologics therapies, n (%) 1 2 3
AA Amyloidosis Cases n ⫽ 36
Non-Amyloidosis Controls n ⫽ 35
P Valuea
55.0 ⫾ 13.1 23 (64) 14.3 ⫾ 10.3 2.9 ⫾ 1.7 102.97
57.3 ⫾ 12.8 22 (63) 14.7 ⫾ 12.2 2.9 ⫾ 1.8 102.11
.447 .561 .883 .963
21 8 4 1 1 1
18 (51) 11 (31) 5 (14) 1 (3) 0 0
.865b (58) (22) (11) (3) (3) (3)
.010b 29 (81) 7 (19) 0
19 (54) 10 (29) 6 (17) .613
17 12 3 2 1 1 14
(47) (33) (8) (6) (3) (3) (39)
22 (65) 11 (31) 1 (3) 0 0 (0) 0 12 (34)
27 (75) 6 (17) 3 (8)
30 (86) 4 (11) 1 (3)
.687 .242d
TNF ⫽ tumor necrosis factor. a Independent 2-sample t test or chi-square test. b Fisher exact test. c At incidence in the cohort. d Wilcoxon Mann–Whitney test.
respectively. The continuation rate with the drug was similar between cases and controls (mean [95% CI] of 4.1 years [3.3-4.9] vs 4.8 years [4.1-5.4]; log rank P ⫽ .106). Factors that predicted the withdrawal of anti-tumor necrosis factor treatment in the Cox regression models were identified (Table 3). The multivariate model demonstrated that the probability of premature interruption of the antitumor necrosis factor treatment in the patients with amyloidosis was 10% for each year in the delay of treatment onset after the diagnosis of amyloidosis and approximately 5 times higher if the levels of proteinuria at the onset of treatment were 2 g/d or more. In the multivariate analysis, the degree of kidney dysfunction did not seem to influence the continuation rate of the anti-tumor necrosis factor drug. In the logistic regression analysis, patients treated with prednisone and those with higher initial proteinuria were more likely to achieve a kidney response at the end of follow-up (Table 3). The progression of renal amyloidosis was not affected by any baseline variable.
The motives for the interruption of the first option biological therapy in these cases were toxicity (22%), ineffectiveness or escape phenomena (17%), improvement (3%), and preg-
Table 2 Baseline Clinical Data Related to Amyloid A Amyloidosis Diagnosis Clinical manifestations
N (%)
Kidney (proteinuria or renal failure) Proteinuria Nephrotic syndrome Renal failure Liver (high aminotransferases) Gastrointestinal tract (chronic diarrhoea) Urinary bladder (incoercible hematuria) Peripheral nerves (polyneuropathy) Bone marrow (pancytopenia)
34 31 9 26 3 4 1 1 1
(94) (86) (26) (72) (8) (11) (17) (17) (17)
458
The American Journal of Medicine, Vol 123, No 5, May 2010 failure, massive hemoptysis, sudden death, myocardiopathies, and pulmonary hypertension) and 3 (37.5%) with infection. One patient (6%) died of pneumonia in the control group (Fisher exact test P ⫽ .085). The proportion of patients with more organs clinically affected by amyloidosis was greater among the patients who died (4 [50%] vs 3 [11%], odds ratio 8.3; 95% CI, 1.3-182.0; Fisher exact test P ⫽ .030). The median survival of the cases from the date they were diagnosed with amyloid A amyloidosis was 12.2 years (95% CI, 9.1-15.2). In addition, the 5-year accumulated survival was 90.6% (95% CI, 80.2-100) and the 10year survival was 78.5% (95% CI, 59.5-97.5). The level of proteinuria at onset increased the risk of death by 10% for each gram of proteinuria in 24 hours (Table 3).
Figure 1 Evolution of kidney parameter means in patients with amyloid A amyloidosis treated with anti-tumor necrosis factor drugs. Analysis of repeated-measure analysis of variance (Friedman test was used in the proteinuria) is limited to the first 24 months (the median of follow-up). Abscissa axis shows the estimated marginal means of proteinuria (g/d) (A), serum creatinine (mg/dL) (B), and creatinine clearance (mL/min) (C).
nancy (3%). The main supporting treatments used in the cases with severe nephropathy were hemodialysis in 4 cases (11%), angiotensin-converting enzyme inhibitors in 2 cases (6%), erythropoietin in 2 cases (6%), and other treatments in 5 cases (14%).
Survival During the follow-up period, 8 patients died (22%), 5 (62.5%) with progression of the amyloidosis (multiorgan
Figure 2 Evolution of acute phase reactant means in patients with amyloid A amyloidosis treated with anti-tumor necrosis factor drugs. Analysis of repeated-measure analysis of variance is limited to the first 24 months (median of follow-up). Post hoc test of analysis of variance (Bonferroni-Dunn test) did not demonstrate significant changes between each visit after the third month. Abscissa axes show the estimated marginal means of erythrocyte sedimentation rate (A) and C-reactive protein (B). ESR ⫽ erythrocyte sedimentation rate; CRP ⫽ C-reactive protein.
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
459
Table 3 Baseline Predictors of Time to Treatment Failure or Death in Patients with Secondary Amyloidosis Who Received at Least One Dose of Anti-Tumor Necrosis Factor Drug Univariate Predictor
Multivariate
Hazard Ratio (95% CI)
Outcome: time to treatment failure Duration of amyloidosis, y Proteinuria ⱖ 2 g/d Serum creatinine, mg/dL Outcome: death Proteinuria at incidence, g/d Involvement of ⬎ 1 organ Outcome: kidney response Prednisone Proteinuria, g/d
P Value
Hazard Ratio (95% CI)
P Value
1.1 (0.9-1.2) 4.1 (1.5-11.5) 1.4 (1.0-2.0)
.067 .004 .041
1.1 (1.0-1.3) 4.6 (1.6-13.3) —
.049 .005 —
1.1 (1.0-1.3) 3.8 (0.9-15.1)
.037 .044
1.1 (1.0-1.3) —
.042 —
6.0 (1.3-27.3) 1.3 (1.0-1.6)
.029a ⬍.001
8.4 (1.4-51.0) 1.3 (1.1-1.6)
.021 .015
CI ⫽ confidence interval. a Fisher exact test.
Adverse Events
before, shortly after starting anti-tumor necrosis factor therapy. The sites of infection were similar in both groups.
The rate of adverse events and the motifs underlying the principal adverse events in cases and controls were assessed (Table 4). The rate of adverse events was similar in both groups, although infections were 3 times as frequent among the cases and all the fatal adverse events were infectious. Although the total number of severe infections was not very different between the groups, only the patients with amyloid A amyloidosis had sepsis and a greater incidence of fatal infections (pulmonary aspergillosis, staphylococcal endocarditis, and septic shock). Two of the patients who died of a fatal infection had stopped taking chlorambucil 2 years
Table 4
DISCUSSION The present study showed the sustained efficacy of tumor necrosis factor-␣ antagonists in treating amyloid A amyloidosis secondary to rheumatic diseases. A continuous reduction in proteinuria with this therapy was associated with the stabilization of kidney function and a marked reduction of acute phase reactants. However, both infections and deaths were common.
Incidence of Adverse Events in Cases and Controls No./Incidence Rate
Total adverse events Non-severe adverse events Severe adverse events Fatal adverse events Total infections Total severe infections Sepsis Fatal infections Total deaths Site of infectionse Urologic Respiratory Cutaneous Joint
Casesa
Controlsb
Incidence Rate Ratio Point Estimate (95% CI)
P Value
36/0.35 22/0.21 14/0.14 3/0.03 35/0.34 7/0.07 4/0.04 3/0.03 8/0.08c
31/0.30 21/0.20 10/0.10 1/0.01 10/0.10 5/0.05 0 1/0.01 1/0.01d
1.15 1.04 1.39 2.79 3.47 1.39 — 2.97 7.93
.567 .902 .436 .379 ⬍.001 .591 .064 .379 .022
6 (17%) 5 (14%) 6 (17%) 0
1 2 1 2
— — — —
(3%) (6%) (3%) (6%)
CI ⫽ confidence interval. a 102.97 patient-y. b 102.11 patient-y. c Five of these patients died of progression of amyloidosis, and 3 patients died of infection. d This patient died of pneumonia. e P values were obtained using Fisher exact conditional test.
(0.71-1.86) (0.57-1.89) (0.62-3.11) (0.34-25.66) (1.79-6.71) (0.44-4.35) (0.34-25.66) (1.38-45.61)
.108 .431 .110 .225
460 Having found that tumor necrosis factor-␣ antagonists might be useful to treat amyloid A amyloidosis in studies with a short follow-up,4 we now confirm these results with a mean follow-up of 34.5 months. Similar benefits of antitumor necrosis factor drugs also were demonstrated in 2 studies;3,10 however, control groups were not included in either of these 2 studies. The total number of adverse events was not different between cases and controls, probably because these are patients with relatively complex conditions. Nevertheless, the rate of infection in our patients was 3 times that in the control group, and severe infections (sepsis and fatal infections) were more common among the cases. The small number of patients is the principal limitation of our study and explains the failure to find any difference in the total number of severe infections (especially sepsis and fatal infections), despite their clinical relevance. Nevertheless, although there was a clear tendency for infections to develop in our cohort, the rates found are fairly similar to those in other series of patients with amyloid A amyloidosis regardless of whether they are treated with immunosuppressors.11-15 It is probable that the background of immunosuppressant and the comorbidity, in addition to anti-tumor necrosis factor therapy, play an important role in the risk of infection in these patients, especially considering that 2 of our patients who died had received chlorambucil. The differences observed with the earlier studies mentioned above3,10 could reflect the different characteristics of these studies, such as the methods used, ethnic backgrounds, heterogeneous diagnoses, or drug doses administered. It is worth emphasizing that after 4 years of tumor necrosis factor-␣ antagonist treatment (including switching), 52% of the cases remained on treatment and the continuation rate of amyloid patients was similar to that of the controls. The factors that predicted the time of withdrawal of anti-tumor necrosis factor therapy were the lag in the institution of therapy after diagnosis of amyloidosis and proteinuria greater than 2 g/d. Thus, it is important to diagnose amyloidosis as soon as possible and not to let proteinuria progress. This study has several strengths, and indeed, the patient cohort size is reasonably large for this rare disease. Furthermore, it is a prospective cohort study with a robust control group. However, the absence of a placebo group prevents us from determining whether this treatment really influences patient survival. The inclusion of this type of control is impossible for ethical reasons. Although many risk factors for infection have been adjusted in a propensity score, this was not so for the type of anti-tumor necrosis factor agent, and we must bear in mind that the majority of the subjects with amyloidosis only received infliximab. We emphasize the importance of sustained annulment of inflammatory activity of any origin in such patients, including that caused by infection. However, this objective is apparently not achieved by all anti-tumor necrosis factor drugs because acute-phase reactants only become fully normalized in less than 25% of patients even though they might
The American Journal of Medicine, Vol 123, No 5, May 2010 experience a relevant reduction in 28% (erythrocyte sedimentation rate) and 50% (C-reactive protein) of cases. It is likely that the decrease in proteinuria was modest and only progressed slowly for this reason. Although the suppression of inflammatory activity should be pursued intensely, the complexity of these patients and the elevated risk of infection could be as important a limitation as the lack of alternative therapies.
CONCLUSIONS We have shown that anti-tumor necrosis factor drugs may be useful to treat kidney amyloidosis but could increase the risk and severity of infections. The long-term follow-up of these patients is essential to define the true effect of these drugs on patient survival. In view of the difficulties in recruiting patients to such trials, multicenter studies would be advisable.
ACKNOWLEDGMENT The authors thank the rheumatologists who contribute to the BIOBADASER registry.
References 1. Jacobsson LT, Turesson C, Nilsson JA, et al. Treatment with TNF blockers and mortality risk in patients with rheumatoid arthritis. Ann Rheum Dis. 2007;66:670-675. 2. Carmona L, Descalzo MA, Perez-Pampin E, et al. All-cause and cause-specific mortality in rheumatoid arthritis are not greater than expected when treated with tumour necrosis factor antagonists. Ann Rheum Dis. 2007;66:880-885. 3. Gottenberg JE, Merle-Vincent F, Bentaberry F, et al. Anti-tumor necrosis factor alpha therapy in fifteen patients with AA amyloidosis secondary to inflammatory arthritides: a followup report of tolerability and efficacy. Arthritis Rheum. 2003;48:2019-2024. 4. Fernández-Nebro A, Tomero E, Ortiz-Santamaría V, et al. Treatment of rheumatic inflammatory disease in 25 patients with secondary amyloidosis using tumor necrosis factor alpha antagonists. Am J Med. 2005;118:552-556. 5. Committee BS. BIOBADASER: Spanish registry of adverse events of biological therapies in rheumatic diseases. Rev Esp Reumatol. 2002; 29:7. 6. Lachmann HJ, Goodman HJ, Gilbertson JA, et al. Natural history and outcome in systemic AA amyloidosis. N Engl J Med. 2007;356:23612371. 7. Joffe MM, Rosenbaum PR. Invited commentary: propensity scores. Am J Epidemiol. 1999;150:327-333. 8. Gomez-Reino JJ, Carmona L, Valverde VR, et al. Treatment of rheumatoid arthritis with tumor necrosis factor inhibitors may predispose to significant increase in tuberculosis risk: a multicenter active-surveillance report. Arthritis Rheum. 2003;48:2122-2127. 9. Gertz MA, Comenzo R, Falk RH, et al. Definition of organ involvement and treatment response in immunoglobulin light chain amyloidosis (AL): a consensus opinion from the 10th International Symposium on Amyloid and Amyloidosis, Tours, France, 18-22 April 2004. Am J Hematol. 2005;79:319-328. 10. Nakamura T, Higashi S, Tomoda K, et al. Efficacy of etanercept in patients with AA amyloidosis secondary to rheumatoid arthritis. Clin Exp Rheumatol. 2007;25:518-522. 11. Joss N, McLaughlin K, Simpson K, Boulton-Jones JM. Presentation, survival and prognostic markers in AA amyloidosis. QJM. 2000;93: 535-542.
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
12. Reyners AK, Hazenberg BP, Reitsma WD, Smit AJ. Heart rate variability as a predictor of mortality in patients with AA and AL amyloidosis. Eur Heart J. 2002;23:157-161. 13. Gertz MA, Kyle RA. Secondary systemic amyloidosis: response and survival in 64 patients. Medicine (Baltimore). 1991;70:246-256. 14. Browning MJ, Banks RA, Tribe CR, et al. Ten years’ experience of an amyloid clinic—a clinicopathological survey. Q J Med. 1985;54:213-227. 15. Triger DR, Joekes AM. Renal amyloidosis—a fourteen-year followup. Q J Med. 1973;42:15-40.
Conflict of Interest: None of the authors have any conflicts of interest associated with the work presented in this manuscript; however, Antonio
461
Fernández-Nebro has earned $1500 or more per year as speaker or consultant (included all companies) for Abbott, Roche, Schering-Plough, and Bristol-Meyer-Squibb companies during the past 3 years, and Rosario García-Vicuña has earned $1500 or more per year as speaker or consultant (included all companies) for Abbott, Roche, Schering-Plough, and Wyeth companies during the past 3 years. Authorship: All authors had access to the data and played a role in writing this manuscript. Dr Fernández-Nebro 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: FernándezNebro, García-Vicuña, Olivé. Acquisition of data: Fernández-Nebro, Herranz, Riera, Castro. Analysis and interpretation of data: Fernández-Nebro, García de Yébenes, García-Vicuña, Olivé.
Long-term TNF-␣ Blockade in Patients with Amyloid A Amyloidosis Complicating Rheumatic Diseases Antonio Fernández-Nebro, MD, PhD,a Alejandro Olivé, MD, PhD,b María Carmen Castro, MD, PhD,c Angela Herranz Varela, MD,d Elena Riera, MD,b Maria V. Irigoyen, MD,a María Jesús García de Yébenes, MD, PhD,d Rosario García-Vicuña, MD, PhDe a Servicio de Reumatología y Departamento de Medicina de la Facultad de Medicina, Hospital Regional Universitario Carlos Haya, Málaga, Spain; bRheumatology Section, Hospital Universitario Germans Trias i Pujol, Badalona, Barcelona, Spain; c Servicio de Reumatología, Hospital Universitario Reina Sofía, Córdoba, Spain; dUnidad de Investigación, Fundación Española de Reumatología, Madrid, Spain; eRheumatology Service, Hospital Universitario de la Princesa, Universidad Autónoma de Madrid, Madrid, Spain.
ABSTRACT OBJECTIVE: To evaluate the effectiveness and safety of anti-tumor necrosis factor therapy in patients with amyloid A amyloidosis. METHODS: Multicenter, controlled, dynamic prospective cohort study of 36 patients with amyloid A amyloidosis (94% kidney involvement) treated with anti-tumor necrosis factor agents (drug exposure of 102.97 patient-years). As an external control group, 35 propensity score-matched non-amyloid patients were chosen from the Base de Datos de Productos Biológicos de la Sociedad Española de Reumatología registry. The end points were kidney response and progression, anti-tumor necrosis factor continuation rate, patient survival, and adverse events. RESULTS: At the end of follow-up, a kidney response was observed in 12 of 22 patients (54.5%) and a kidney progression was observed in 6 of 36 patients (17%). The kidney amyloidosis remained stable in 16 of 36 patients (44%). The level of acute phase reactants diminished but did not reach the normal level. The continuation rates of anti-tumor necrosis factor drugs among patients with amyloid A amyloidosis after 1, 2, 3, and 4 or more years were 80%, 80%, 61%, and 52%, respectively, comparable to controls. The 5-year cumulative survival of amyloid A amyloidosis cases was 90.6%, and the 10-year survival was 78.5%. In a multivariate Cox regression analysis, the duration of amyloidosis and the level of proteinuria at the onset of anti-tumor necrosis factor treatment were independent predictors of treatment failure, whereas the level of proteinuria was the only factor that predicts mortality. Most adverse events were similar in both groups, although the number of infections was 3 times higher in amyloid A amyloidosis cases. CONCLUSION: Anti-tumor necrosis factor drugs are effective in treating amyloid A amyloidosis, although they might increase the risk of infection. © 2010 Elsevier Inc. All rights reserved. • The American Journal of Medicine (2010) 123, 454-461 KEYWORDS: Amyloid A amyloidosis; Anti-tumor necrosis factor-alpha blockade; Effectiveness; Immune-mediated inflammatory diseases; Infections
Funding: None. The drugs administered to our patients were supplied by the sanitary service as part of the medical care that they normally received. Conflict of Interest: See last page of article. Authorship: See last page of article.
0002-9343/$ -see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2009.11.010
Reprint requests should be addressed to Antonio Fernández-Nebro, MD, PhD, Servicio de Reumatología, Hospital Regional Universitario Carlos Haya, Pabellón C: “Hospital Civil,” Plaza del Hospital Civil s/n, 29009, Málaga, Spain. E-mail address: [email protected]
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
455
The discovery that clinically unrelated conditions, such as biological therapy was not officially accepted for the underrheumatoid arthritis and Crohn’s disease, share similar allying disease. terations in the immune system has led to a shift in the management of immune-mediated inflammatory diseases External Controls from organ-based symptom relief to mechanism-based Controls were extracted from the Base de Datos de Protreatment. Currently, anti-tumor ductos Biológicos de la Sociedad necrosis factor-␣ drugs offer a Española de Reumatología regissuitable treatment for many imtry (BIOBADASER), a detailed CLINICAL SIGNIFICANCE mune-mediated inflammatory disdescription of which has been pubeases refractory to nonbiological lished5 and is available at http:// ● Anti-tumor necrosis factor drugs may be disease-modifying antirheumatic biobadaser.ser.es/. Propensity score useful to treat kidney amyloidosis but drugs because of their efficacy and methods were applied to match pacould increase the risk and severity of rapid onset of action and tolerabiltients and controls according to sevinfections. ity. These treatments not only reeral baseline characteristics related ● Patient survival depends on the degree duce inflammatory activity but to amyloid A amyloidosis and inalso reduce the complications and of proteinuria at the beginning of antifection susceptibility:6 age, gender, excessive mortality associated tumor necrosis factor drug treatment. type and duration of rheumatic with some of these diseases.1,2 disease, and steroid use. First, ● The continuation rates with anti-tumor Furthermore, there is evidence patients with amyloidosis were necrosis factor drugs are similar in pathat these anti-tumor necrosis facexcluded from BIOBADASER tients with amyloid A amyloidosis and tor drugs may rapidly improve the controls. The matching variables controls without amyloidosis, although clinical status of patients with were then included in a multivarboth delay of treatment after the diagamyloid A amyloidosis secondary iate logistic regression to calcuto immune-mediated inflammanosis of amyloidosis and a high-grade late a predictive probability (protory diseases.3,4 Nevertheless, it proteinuria were associated with premapensity score) of both groups (with still remains unclear whether the and without amyloidosis) using ture interruption of treatment. effects of such therapies are susSTATA-specific commands. Fitained in the long-term or if they nally, this propensity score was reduce the mortality associated used to match amyloidosis cases with this severe complication. Indeed, because anti-tumor with the single nearest-neighbor controls without amynecrosis factor drugs also increase the incidence of serious loidosis in the BIOBADASER subsample.7 infections, especially in patients with associated comorbidity,2 it is important to determine whether any potential Follow-up and Amyloidosis Treatment reduction in mortality due to amyloidosis could be underProtocol mined by an increase in the mortality due to severe The following data were registered for each case: gender, infections. age, underlying inflammatory disease, time of evolution of Thus, as well as investigating the long-term effectiveness the underlying disease and of amyloidosis, diagnostic test of anti-tumor necrosis factor-␣ therapies in amyloid A amyused for amyloidosis, interstitial fibrosis in the kidney biloidosis associated with immune-mediated inflammatory opsy, organs affected, anti-tumor necrosis factor drug dose diseases, an important aim of this study was to assess the and time of use, cause of withdrawal, concomitant treatincidence of severe adverse events, especially infections, ments, adverse events, vital status, urinary protein excretion and their impact on patient mortality when compared with a in 24 hours, serum creatinine, creatinine clearance in 24 propensity score-matched population free of amyloidosis. hours, C-reactive protein, and erythrocyte sedimentation
PATIENTS AND METHODS Patients The patients included in this study were all patients with amyloid A amyloidosis secondary to an immune-mediated inflammatory disease who were treated and followed up prospectively in 4 Spanish hospitals from January 2001 to October 2006. This study represents an extension of a previous study.4 Amyloid A amyloidosis was confirmed histologically by Congo-red staining and immunohistochemistry. Informed consent was obtained from all patients, and the treatment was approved by the local health administration if
rate at baseline and periodically during the follow-up. The data were recorded at the visits scheduled at months 0, 3, 6, and 12, and then annually. All patients were screened, and if necessary, they were treated for active or latent tuberculosis following the recommendations of the European and Spanish Drug agencies.8 The anti-tumor necrosis factor drugs (infliximab, etanercept, or adalimumab) were chosen by each physician and initially administered according to the technical data sheets. There were no restrictions on the use of disease-modifying antirheumatic drugs, steroids, or supportive treatments. The treatments were readjusted or interrupted by the clinician according to the therapeutic response or the appearance of adverse events.
456
Evaluation Criteria All patients who received at least 1 dose of anti-tumor necrosis factor were evaluated. Treatment effectiveness was evaluated by the improvement of kidney outcomes (kidney response and progression), continuation rates of anti-tumor necrosis factor drugs, and the survival rate of patients with amyloidosis. Kidney outcomes were defined according to established consensus criteria.9 Kidney response was defined as follows: 50% or greater decrease (ⱖ0.5 g/d) of 24-hour urine protein excretion (proteinuria ⬎ 0.5 g/d pretreatment), and serum creatinine and creatinine clearance must not worsen by 25% above the baseline. Kidney progression was defined by a 50% increase in proteinuria (ⱖ1 g/d) of proteinuria to greater than 1 g/d (proteinuria must be ⬎ 0.5 g/d pretreatment) or 25% worsening of serum creatinine (minimum of 0.5 mg/dL) or creatinine clearance. Patients who do not fulfill the criteria for progressive disease or responsive disease are considered stable. Safety was evaluated through the relevant adverse events as assessed through physical examination and patient self-reporting at each visit. Our definition of a relevant adverse event was the same as that described in BIOBADASER.5
Statistical Analysis Intention-to-treat analysis was used. Incidence date was the first date of anti-tumor necrosis factor therapy after the diagnosis of amyloidosis. Quantitative descriptive data are presented as the mean (⫾ standard deviation), and the rates were expressed per patient-years (95% confidence interval). Normality was confirmed by the Kolmogorov test. For skewed data, the median (⫾ interquartile range, interquartile range) was used. The chi-square or Fisher exact 2-tailed test was used to compare categories. The efficacy in terms of proteinuria, renal function, and acute phase reactants at 0, 3, 6, 12, and 24 months was assessed by repeated-measures analysis of variance (Friedman test was used to detect differences in the proteinuria). Kaplan–Meier curves and the log-rank test were used to estimate the time to withdrawal of anti-tumor necrosis factor drugs and to compare the survival functions between cases and controls, respectively. Survival time was measured from the date of amyloidosis diagnosis to the date of death. A Cox regression was used to identify the prognostic factors of time to withdrawal of anti-tumor necrosis factor drug and death in the univariate and multivariate analyses (stepwise forward/Wald). All variables that reached a value of P less than .10 were included in the Cox multivariate model. A logistic regression analysis was used to identify the prognostic factors of kidney outcomes.
RESULTS From January 1, 2001, to October 31, 2006, a cohort of 36 patients with amyloid A amyloidosis was generated. The patients were treated with anti-tumor necrosis factor drugs for a total of 102.97 patient-years (median follow-up 34.6 months). Thirty-five propensity score-matched control subjects were chosen from the BIOBADASER registry. All
The American Journal of Medicine, Vol 123, No 5, May 2010 patients had highly active immune-mediated inflammatory diseases refractory to other treatments.
Patient Characteristics The baseline clinical status of the 36 patients and that of the controls was similar (Table 1). The amyloid A amyloidosis had been diagnosed 2.7 ⫾ 3.4 years (range, 0-13) before treatment started. Three of the cases were being treated by infliximab when they were diagnosed with amyloid A amyloidosis (one of whom had been treated for 29 months), whereas the remaining patients had not previously received biological therapy. Infliximab was the drug of first choice. The dose of infliximab was increased (from an initial mean of 3.6 mg/kg ⫾ 0.7 to 4.0 mg/kg ⫾ 0.9) or the interval between doses was reduced (from every 8 to every 6 weeks) in 13 of 29 cases (45%). Only in 2 cases did the interval between doses increase (to every 9 weeks and to every 11 weeks). Nevertheless, the majority of the subjects with amyloidosis continued to receive infliximab every 8 weeks. The dose and the interval of etanercept and adalimumab administration did not change in any patient. One of the cases of amyloidosis diagnosed with adult-onset Still disease received anakinra (100 mg/d) as a third option. Baseline clinical data related to amyloidosis diagnosis were collected (Table 2), and 56 biopsies were performed at 7 distinct locations to histologically confirm amyloidosis. The presence of a clinical parameter associated with altered kidney function was the initial indication of amyloidosis in most cases.
Effectiveness of Anti-Tumor Necrosis Factor against Amyloid A Amyloidosis The anti-tumor necrosis factor treatment reduced the median levels of proteinuria by 59.7% during the first 24 months of the follow-up period (median 0.67 g/d [interquartile range ⫽ 0.24-3.24] vs 0.40 g/d [interquartile range ⫽ 0.14-1.84]; Friedman test P ⬍ .001) (Figure 1). Both the mean serum creatinine and creatinine clearance remained stable. At the end of follow-up, kidney response was observed in 12 of 22 patients (54.5%), and kidney progression was observed in 6 of 36 patients (17%). Kidney amyloidosis remained stable in 16 of 36 patients (44%). The level of acute phase reactants diminished quickly but did not reach the normal level (Figure 2). The mean erythrocyte sedimentation rate decreased during the first 24 months from 70.7 mm/h (95% confidence interval (CI), 60.3-81.3) to 51.4 mm/h (95% CI, 39.7-63.1, F test P ⫽ .028), whereas the levels of C-reactive protein decreased from 39.5 mg/L (95% CI, 27.7-51.3) to 20.3 mg/L (95% CI, 12.7-27.9; F test P ⬍ .001). At the end of the follow-up period, normal values for the erythrocyte sedimentation rate and C-reactive protein were only evident in 7 patients (19%) and 9 patients (25%), respectively. The cumulative incidence of patients with amyloidosis who still received anti-tumor necrosis factor treatment after 1, 2, 3, and 4 or more years were 80%, 80%, 61%, and 52%,
Fernández-Nebro et al Table 1
Anti-TNF Drugs in Amyloid A Amyloidosis
457
Baseline Demographic and Clinical Characteristics of Cases and Controls
Age, y Women, n (%) Duration of disease, y Duration of follow-up, y Patient-y Underlying disease, n (%) Rheumatoid arthritis Spondyloarthropathy Psoriatic arthritis Juvenile idiopathic arthritis Adult-onset Still disease Antiphospholipid syndrome Anti-TNF treatment, n (%)c Infliximab Etanercept Adalimumab Concomitant disease-modifying antirheumatic drugs, n (%)b No disease-modifying antirheumatic drugs Methotrexate Leflunomide Chlorambucil Sulfasalazine Hydroxychloroquine Corticosteroids, n (%) No. of biologics therapies, n (%) 1 2 3
AA Amyloidosis Cases n ⫽ 36
Non-Amyloidosis Controls n ⫽ 35
P Valuea
55.0 ⫾ 13.1 23 (64) 14.3 ⫾ 10.3 2.9 ⫾ 1.7 102.97
57.3 ⫾ 12.8 22 (63) 14.7 ⫾ 12.2 2.9 ⫾ 1.8 102.11
.447 .561 .883 .963
21 8 4 1 1 1
18 (51) 11 (31) 5 (14) 1 (3) 0 0
.865b (58) (22) (11) (3) (3) (3)
.010b 29 (81) 7 (19) 0
19 (54) 10 (29) 6 (17) .613
17 12 3 2 1 1 14
(47) (33) (8) (6) (3) (3) (39)
22 (65) 11 (31) 1 (3) 0 0 (0) 0 12 (34)
27 (75) 6 (17) 3 (8)
30 (86) 4 (11) 1 (3)
.687 .242d
TNF ⫽ tumor necrosis factor. a Independent 2-sample t test or chi-square test. b Fisher exact test. c At incidence in the cohort. d Wilcoxon Mann–Whitney test.
respectively. The continuation rate with the drug was similar between cases and controls (mean [95% CI] of 4.1 years [3.3-4.9] vs 4.8 years [4.1-5.4]; log rank P ⫽ .106). Factors that predicted the withdrawal of anti-tumor necrosis factor treatment in the Cox regression models were identified (Table 3). The multivariate model demonstrated that the probability of premature interruption of the antitumor necrosis factor treatment in the patients with amyloidosis was 10% for each year in the delay of treatment onset after the diagnosis of amyloidosis and approximately 5 times higher if the levels of proteinuria at the onset of treatment were 2 g/d or more. In the multivariate analysis, the degree of kidney dysfunction did not seem to influence the continuation rate of the anti-tumor necrosis factor drug. In the logistic regression analysis, patients treated with prednisone and those with higher initial proteinuria were more likely to achieve a kidney response at the end of follow-up (Table 3). The progression of renal amyloidosis was not affected by any baseline variable.
The motives for the interruption of the first option biological therapy in these cases were toxicity (22%), ineffectiveness or escape phenomena (17%), improvement (3%), and preg-
Table 2 Baseline Clinical Data Related to Amyloid A Amyloidosis Diagnosis Clinical manifestations
N (%)
Kidney (proteinuria or renal failure) Proteinuria Nephrotic syndrome Renal failure Liver (high aminotransferases) Gastrointestinal tract (chronic diarrhoea) Urinary bladder (incoercible hematuria) Peripheral nerves (polyneuropathy) Bone marrow (pancytopenia)
34 31 9 26 3 4 1 1 1
(94) (86) (26) (72) (8) (11) (17) (17) (17)
458
The American Journal of Medicine, Vol 123, No 5, May 2010 failure, massive hemoptysis, sudden death, myocardiopathies, and pulmonary hypertension) and 3 (37.5%) with infection. One patient (6%) died of pneumonia in the control group (Fisher exact test P ⫽ .085). The proportion of patients with more organs clinically affected by amyloidosis was greater among the patients who died (4 [50%] vs 3 [11%], odds ratio 8.3; 95% CI, 1.3-182.0; Fisher exact test P ⫽ .030). The median survival of the cases from the date they were diagnosed with amyloid A amyloidosis was 12.2 years (95% CI, 9.1-15.2). In addition, the 5-year accumulated survival was 90.6% (95% CI, 80.2-100) and the 10year survival was 78.5% (95% CI, 59.5-97.5). The level of proteinuria at onset increased the risk of death by 10% for each gram of proteinuria in 24 hours (Table 3).
Figure 1 Evolution of kidney parameter means in patients with amyloid A amyloidosis treated with anti-tumor necrosis factor drugs. Analysis of repeated-measure analysis of variance (Friedman test was used in the proteinuria) is limited to the first 24 months (the median of follow-up). Abscissa axis shows the estimated marginal means of proteinuria (g/d) (A), serum creatinine (mg/dL) (B), and creatinine clearance (mL/min) (C).
nancy (3%). The main supporting treatments used in the cases with severe nephropathy were hemodialysis in 4 cases (11%), angiotensin-converting enzyme inhibitors in 2 cases (6%), erythropoietin in 2 cases (6%), and other treatments in 5 cases (14%).
Survival During the follow-up period, 8 patients died (22%), 5 (62.5%) with progression of the amyloidosis (multiorgan
Figure 2 Evolution of acute phase reactant means in patients with amyloid A amyloidosis treated with anti-tumor necrosis factor drugs. Analysis of repeated-measure analysis of variance is limited to the first 24 months (median of follow-up). Post hoc test of analysis of variance (Bonferroni-Dunn test) did not demonstrate significant changes between each visit after the third month. Abscissa axes show the estimated marginal means of erythrocyte sedimentation rate (A) and C-reactive protein (B). ESR ⫽ erythrocyte sedimentation rate; CRP ⫽ C-reactive protein.
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
459
Table 3 Baseline Predictors of Time to Treatment Failure or Death in Patients with Secondary Amyloidosis Who Received at Least One Dose of Anti-Tumor Necrosis Factor Drug Univariate Predictor
Multivariate
Hazard Ratio (95% CI)
Outcome: time to treatment failure Duration of amyloidosis, y Proteinuria ⱖ 2 g/d Serum creatinine, mg/dL Outcome: death Proteinuria at incidence, g/d Involvement of ⬎ 1 organ Outcome: kidney response Prednisone Proteinuria, g/d
P Value
Hazard Ratio (95% CI)
P Value
1.1 (0.9-1.2) 4.1 (1.5-11.5) 1.4 (1.0-2.0)
.067 .004 .041
1.1 (1.0-1.3) 4.6 (1.6-13.3) —
.049 .005 —
1.1 (1.0-1.3) 3.8 (0.9-15.1)
.037 .044
1.1 (1.0-1.3) —
.042 —
6.0 (1.3-27.3) 1.3 (1.0-1.6)
.029a ⬍.001
8.4 (1.4-51.0) 1.3 (1.1-1.6)
.021 .015
CI ⫽ confidence interval. a Fisher exact test.
Adverse Events
before, shortly after starting anti-tumor necrosis factor therapy. The sites of infection were similar in both groups.
The rate of adverse events and the motifs underlying the principal adverse events in cases and controls were assessed (Table 4). The rate of adverse events was similar in both groups, although infections were 3 times as frequent among the cases and all the fatal adverse events were infectious. Although the total number of severe infections was not very different between the groups, only the patients with amyloid A amyloidosis had sepsis and a greater incidence of fatal infections (pulmonary aspergillosis, staphylococcal endocarditis, and septic shock). Two of the patients who died of a fatal infection had stopped taking chlorambucil 2 years
Table 4
DISCUSSION The present study showed the sustained efficacy of tumor necrosis factor-␣ antagonists in treating amyloid A amyloidosis secondary to rheumatic diseases. A continuous reduction in proteinuria with this therapy was associated with the stabilization of kidney function and a marked reduction of acute phase reactants. However, both infections and deaths were common.
Incidence of Adverse Events in Cases and Controls No./Incidence Rate
Total adverse events Non-severe adverse events Severe adverse events Fatal adverse events Total infections Total severe infections Sepsis Fatal infections Total deaths Site of infectionse Urologic Respiratory Cutaneous Joint
Casesa
Controlsb
Incidence Rate Ratio Point Estimate (95% CI)
P Value
36/0.35 22/0.21 14/0.14 3/0.03 35/0.34 7/0.07 4/0.04 3/0.03 8/0.08c
31/0.30 21/0.20 10/0.10 1/0.01 10/0.10 5/0.05 0 1/0.01 1/0.01d
1.15 1.04 1.39 2.79 3.47 1.39 — 2.97 7.93
.567 .902 .436 .379 ⬍.001 .591 .064 .379 .022
6 (17%) 5 (14%) 6 (17%) 0
1 2 1 2
— — — —
(3%) (6%) (3%) (6%)
CI ⫽ confidence interval. a 102.97 patient-y. b 102.11 patient-y. c Five of these patients died of progression of amyloidosis, and 3 patients died of infection. d This patient died of pneumonia. e P values were obtained using Fisher exact conditional test.
(0.71-1.86) (0.57-1.89) (0.62-3.11) (0.34-25.66) (1.79-6.71) (0.44-4.35) (0.34-25.66) (1.38-45.61)
.108 .431 .110 .225
460 Having found that tumor necrosis factor-␣ antagonists might be useful to treat amyloid A amyloidosis in studies with a short follow-up,4 we now confirm these results with a mean follow-up of 34.5 months. Similar benefits of antitumor necrosis factor drugs also were demonstrated in 2 studies;3,10 however, control groups were not included in either of these 2 studies. The total number of adverse events was not different between cases and controls, probably because these are patients with relatively complex conditions. Nevertheless, the rate of infection in our patients was 3 times that in the control group, and severe infections (sepsis and fatal infections) were more common among the cases. The small number of patients is the principal limitation of our study and explains the failure to find any difference in the total number of severe infections (especially sepsis and fatal infections), despite their clinical relevance. Nevertheless, although there was a clear tendency for infections to develop in our cohort, the rates found are fairly similar to those in other series of patients with amyloid A amyloidosis regardless of whether they are treated with immunosuppressors.11-15 It is probable that the background of immunosuppressant and the comorbidity, in addition to anti-tumor necrosis factor therapy, play an important role in the risk of infection in these patients, especially considering that 2 of our patients who died had received chlorambucil. The differences observed with the earlier studies mentioned above3,10 could reflect the different characteristics of these studies, such as the methods used, ethnic backgrounds, heterogeneous diagnoses, or drug doses administered. It is worth emphasizing that after 4 years of tumor necrosis factor-␣ antagonist treatment (including switching), 52% of the cases remained on treatment and the continuation rate of amyloid patients was similar to that of the controls. The factors that predicted the time of withdrawal of anti-tumor necrosis factor therapy were the lag in the institution of therapy after diagnosis of amyloidosis and proteinuria greater than 2 g/d. Thus, it is important to diagnose amyloidosis as soon as possible and not to let proteinuria progress. This study has several strengths, and indeed, the patient cohort size is reasonably large for this rare disease. Furthermore, it is a prospective cohort study with a robust control group. However, the absence of a placebo group prevents us from determining whether this treatment really influences patient survival. The inclusion of this type of control is impossible for ethical reasons. Although many risk factors for infection have been adjusted in a propensity score, this was not so for the type of anti-tumor necrosis factor agent, and we must bear in mind that the majority of the subjects with amyloidosis only received infliximab. We emphasize the importance of sustained annulment of inflammatory activity of any origin in such patients, including that caused by infection. However, this objective is apparently not achieved by all anti-tumor necrosis factor drugs because acute-phase reactants only become fully normalized in less than 25% of patients even though they might
The American Journal of Medicine, Vol 123, No 5, May 2010 experience a relevant reduction in 28% (erythrocyte sedimentation rate) and 50% (C-reactive protein) of cases. It is likely that the decrease in proteinuria was modest and only progressed slowly for this reason. Although the suppression of inflammatory activity should be pursued intensely, the complexity of these patients and the elevated risk of infection could be as important a limitation as the lack of alternative therapies.
CONCLUSIONS We have shown that anti-tumor necrosis factor drugs may be useful to treat kidney amyloidosis but could increase the risk and severity of infections. The long-term follow-up of these patients is essential to define the true effect of these drugs on patient survival. In view of the difficulties in recruiting patients to such trials, multicenter studies would be advisable.
ACKNOWLEDGMENT The authors thank the rheumatologists who contribute to the BIOBADASER registry.
References 1. Jacobsson LT, Turesson C, Nilsson JA, et al. Treatment with TNF blockers and mortality risk in patients with rheumatoid arthritis. Ann Rheum Dis. 2007;66:670-675. 2. Carmona L, Descalzo MA, Perez-Pampin E, et al. All-cause and cause-specific mortality in rheumatoid arthritis are not greater than expected when treated with tumour necrosis factor antagonists. Ann Rheum Dis. 2007;66:880-885. 3. Gottenberg JE, Merle-Vincent F, Bentaberry F, et al. Anti-tumor necrosis factor alpha therapy in fifteen patients with AA amyloidosis secondary to inflammatory arthritides: a followup report of tolerability and efficacy. Arthritis Rheum. 2003;48:2019-2024. 4. Fernández-Nebro A, Tomero E, Ortiz-Santamaría V, et al. Treatment of rheumatic inflammatory disease in 25 patients with secondary amyloidosis using tumor necrosis factor alpha antagonists. Am J Med. 2005;118:552-556. 5. Committee BS. BIOBADASER: Spanish registry of adverse events of biological therapies in rheumatic diseases. Rev Esp Reumatol. 2002; 29:7. 6. Lachmann HJ, Goodman HJ, Gilbertson JA, et al. Natural history and outcome in systemic AA amyloidosis. N Engl J Med. 2007;356:23612371. 7. Joffe MM, Rosenbaum PR. Invited commentary: propensity scores. Am J Epidemiol. 1999;150:327-333. 8. Gomez-Reino JJ, Carmona L, Valverde VR, et al. Treatment of rheumatoid arthritis with tumor necrosis factor inhibitors may predispose to significant increase in tuberculosis risk: a multicenter active-surveillance report. Arthritis Rheum. 2003;48:2122-2127. 9. Gertz MA, Comenzo R, Falk RH, et al. Definition of organ involvement and treatment response in immunoglobulin light chain amyloidosis (AL): a consensus opinion from the 10th International Symposium on Amyloid and Amyloidosis, Tours, France, 18-22 April 2004. Am J Hematol. 2005;79:319-328. 10. Nakamura T, Higashi S, Tomoda K, et al. Efficacy of etanercept in patients with AA amyloidosis secondary to rheumatoid arthritis. Clin Exp Rheumatol. 2007;25:518-522. 11. Joss N, McLaughlin K, Simpson K, Boulton-Jones JM. Presentation, survival and prognostic markers in AA amyloidosis. QJM. 2000;93: 535-542.
Fernández-Nebro et al
Anti-TNF Drugs in Amyloid A Amyloidosis
12. Reyners AK, Hazenberg BP, Reitsma WD, Smit AJ. Heart rate variability as a predictor of mortality in patients with AA and AL amyloidosis. Eur Heart J. 2002;23:157-161. 13. Gertz MA, Kyle RA. Secondary systemic amyloidosis: response and survival in 64 patients. Medicine (Baltimore). 1991;70:246-256. 14. Browning MJ, Banks RA, Tribe CR, et al. Ten years’ experience of an amyloid clinic—a clinicopathological survey. Q J Med. 1985;54:213-227. 15. Triger DR, Joekes AM. Renal amyloidosis—a fourteen-year followup. Q J Med. 1973;42:15-40.
Conflict of Interest: None of the authors have any conflicts of interest associated with the work presented in this manuscript; however, Antonio
461
Fernández-Nebro has earned $1500 or more per year as speaker or consultant (included all companies) for Abbott, Roche, Schering-Plough, and Bristol-Meyer-Squibb companies during the past 3 years, and Rosario García-Vicuña has earned $1500 or more per year as speaker or consultant (included all companies) for Abbott, Roche, Schering-Plough, and Wyeth companies during the past 3 years. Authorship: All authors had access to the data and played a role in writing this manuscript. Dr Fernández-Nebro 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: FernándezNebro, García-Vicuña, Olivé. Acquisition of data: Fernández-Nebro, Herranz, Riera, Castro. Analysis and interpretation of data: Fernández-Nebro, García de Yébenes, García-Vicuña, Olivé.