The

n e w e ng l a n d j o u r na l

of

m e dic i n e

original article

Bardoxolone Methyl and Kidney Function in CKD with Type 2 Diabetes

Pablo E. Pergola, M.D., Ph.D., Philip Raskin, M.D., Robert D. Toto, M.D., Colin J. Meyer, M.D., J. Warren Huff, J.D., Eric B. Grossman, M.D., Melissa Krauth, M.B.A., Stacey Ruiz, Ph.D., Paul Audhya, M.D., Heidi Christ-Schmidt, M.S.E., Janet Wittes, Ph.D., and David G. Warnock, M.D., for the BEAM Study Investigators*

A bs t r ac t
Background

Chronic kidney disease (CKD) associated with type 2 diabetes is the leading cause of kidney failure, with both inflammation and oxidative stress contributing to disease progression. Bardoxolone methyl, an oral antioxidant inflammation modulator, has shown efficacy in patients with CKD and type 2 diabetes in short-term studies, but longer-term effects and dose response have not been determined.
Methods

In this phase 2, double-blind, randomized, placebo-controlled trial, we assigned 227 adults with CKD (defined as an estimated glomerular filtration rate [GFR] of 20 to 45 ml per minute per 1.73 m2 of body-surface area) in a 1:1:1:1 ratio to receive placebo or bardoxolone methyl at a target dose of 25, 75, or 150 mg once daily. The primary outcome was the change from baseline in the estimated GFR with bardoxolone methyl, as compared with placebo, at 24 weeks; a secondary outcome was the change at 52 weeks.
Results

From Renal Associates, San Antonio, TX (P.E.P.); the Department of Medicine, University of Texas Southwestern, Dallas (P.R., R.D.T.); Reata Pharmaceuticals, Irving, TX (C.JM., J.W.H., E.B.G., M.K., S.R., P.A.); Statistics Collaborative, Washington, DC (H.C.-S., J.W.); and the Department of Medicine, University of Alabama at Birmingham, Birmingham (D.G.W.). Address reprint requests to Dr. Pergola at Renal Associates, 215 E. Quincy, Suite 610, San Antonio, TX 78215, or at ppergola@raparesearch.com. * The investigators in the 52-Week Bardoxolone Methyl Treatment: Renal Function in CKD/Type 2 Diabetes (BEAM) study are listed in the Supplementary Appendix, available at NEJM.org. This article (10.1056/NEJMoa1105351) was published on June 24, 2011, at NEJM.org. N Engl J Med 2011;365:327-36.
Copyright 2011 Massachusetts Medical Society.

Patients receiving bardoxolone methyl had significant increases in the mean (SD) estimated GFR, as compared with placebo, at 24 weeks (with between-group differences per minute per 1.73 m2 of 8.21.5 ml in the 25-mg group, 11.41.5 ml in the 75-mg group, and 10.41.5 ml in the 150-mg group; P<0.001). The increases were maintained through week 52, with significant differences per minute per 1.73 m2 of 5.81.8 ml, 10.51.8 ml, and 9.31.9 ml, respectively. Muscle spasms, the most frequent adverse event in the bardoxolone methyl groups, were generally mild and dose-related. Hypomagnesemia, mild increases in alanine aminotransferase levels, and gastrointestinal effects were more common among patients receiving bardoxolone methyl.
Conclusions

Bardoxolone methyl was associated with improvement in the estimated GFR in patients with advanced CKD and type 2 diabetes at 24 weeks. The improvement persisted at 52 weeks, suggesting that bardoxolone methyl may have promise for the treatment of CKD. (Funded by Reata Pharmaceuticals; BEAM ClinicalTrials.gov number, NCT00811889.)

n engl j med 365;4

nejm.org

july 28, 2011

327

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

iabetes mellitus is a major cause of chronic kidney disease (CKD) worldwide.1,2 The complications of CKD (e.g., cardiovascular disease and death) occur before kidney failure develops and are independent of known risk factors (i.e., hypertension and proteinuria).3,4 Although CKD progression is slowed by the use of angiotensin-convertingenzyme (ACE) inhibitors and angiotensin-receptor blockers (ARBs), in many patients, the condition progresses to kidney failure.5-9 CKD in patients with diabetes is associated with chronic inflammation and oxidative stress.10-12 Effects of these processes result in glomerular endothelial dysfunction and mesangial-cell contraction and, with time, glomerular fibrosis and mesangial expansion. These effects result in a decline in kidney function.13-17 Bardoxolone methyl, an antioxidant inflammation modulator, activates the Keap1Nrf2 pathway, which plays an important role in maintaining kidney function and structure.18-22 The chemical and biologic characteristics of bardoxolone methyl, a derivative of the natural product oleanolic acid, have been reviewed recently (see Fig. 1 in the Supplementary Appendix, available with the full text of this article at NEJM.org).23 Bardoxolone methyl interacts with cysteine residues on Keap1, allowing Nrf2 translocation to the nucleus and subsequent up-regulation of a multitude of cytoprotective genes.18,22,23 The structure and activity profile of bardoxolone methyl resemble those of the cyclopentenone prostaglandins, endogenous Nrf2 activators that promote the resolution of inflammation.24 Like cyclopentenone prostaglandins, bardoxolone methyl exerts antiinflammatory effects by inhibiting the proinflammatory nuclear factor B pathway.25-27 In a previous phase 2 trial, we found that daily administration of bardoxolone methyl for 8 weeks significantly increased the estimated glomerular filtration rate (GFR).28 In the randomized, placebocontrolled 52-Week Bardoxolone Methyl Treatment: Renal Function in CKD/Type 2 Diabetes (BEAM) study, the results of which are reported here, we assessed the effects of three doses of bardoxolone methyl on the estimated GFR at 24 and 52 weeks in patients with CKD and type 2 diabetes.

had an estimated GFR of 20 to 45 ml per minute per 1.73 m2 of body-surface area, calculated as the mean of two measurements (differing by 25%) in blood samples collected less than 5 days apart within a 3-week screening period. Treatment with a stable dose of an ACE inhibitor, an ARB, or both for at least 8 weeks before screening was required, unless such therapy was not tolerated. Major exclusion criteria were type 1 diabetes, nondiabetic kidney disease, a glycated hemoglobin level of more than 10%, hepatic dysfunction, or a cardiovascular event within the previous 3 months.
Study Design

We screened 573 patients at 43 sites in the United States. Of these patients, 227 were randomly assigned in a 1:1:1:1 ratio to receive placebo or oral bardoxolone methyl at a dose of 25, 75, or 150 mg once daily for 52 weeks (Fig. 2 in the Supplementary Appendix). The study had four periods: a 21day screening period; an 8-week period of dose adjustment (every 4 weeks) that could be extended up to 20 weeks if needed to reach the randomly assigned dose of bardoxolone methyl; dose maintenance from the end of the dose-adjustment period through week 52; and follow-up for 4 weeks after the last dose. Patients were stratified at randomization according to the estimated GFR (<30 ml or 30 ml per minute per 1.73 m2), the urinary albumin-to-creatinine ratio (ACR, 300 or >300 [with albumin measured in milligrams and creatinine in grams]), and the glycated hemoglobin level (<7% or 7%). The use of concomitant medications, including antihypertensive agents, was adjusted by the treating physician according to established guidelines. Patient safety was assessed at each visit and monitored by an independent data and safety monitoring board. The study protocol (which is available at NEJM.org) was approved by the institutional review board associated with each study center. All patients provided written informed consent.
Study Oversight

This study was sponsored by Reata Pharmaceuticals and was designed by the first author and representatives of the sponsor. Study investigators and coordinators jointly managed the study with the sponsor. The main database was held by a contract research organization (Numoda). Employees Me thods of Statistics Collaborative performed the statistical Patients analyses. Data were reviewed by the data and safeAdults with moderate-to-severe CKD and type 2 ty monitoring board during the study. An employee diabetes were eligible for study enrollment if they of the sponsor wrote the first draft of the manun engl j med 365;4 nejm.org july 28, 2011

328

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Bardoxolone Methyl in CKD with Type 2 Diabetes

script. All the authors participated in manuscript revision, made the decision to submit the manuscript for publication, agreed to maintain data confidentiality pending publication, and vouch for the completeness and accuracy of the data.

from the study, regardless of the status of studydrug administration and categorical variables for treatment, week, and treatment-by-week interaction. Within-patient correlation was modeled with a Toeplitz structure.31 We used KaplanMeier methods to estimate the proportion of patients Procedures and Outcomes remaining event-free at time points of interest. Measurement of the estimated GFR and routine The study groups were compared with the use laboratory testing were performed during screen- of unstratified log-rank tests with unadjusted ing and every 4 weeks by a central laboratory. Cre- two-sided type I error rates. atinine assays were calibrated to an isotope-dilution standard for mass spectrometry. The GFR was esR e sult s timated with the use of the four-variable equation used in the Modification of Diet in Renal Disease Patients study.29 No dietary interventions were included, Dose increases to achieve the randomly assigned and 24-hour urine collections were not obtained. dose occurred during the first 20 weeks. Among The primary outcome was the change from patients receiving bardoxolone methyl, the proporbaseline in the estimated GFR with bardoxolone tions who received the randomly assigned dose at methyl, as compared with placebo, at 24 weeks; a 52 weeks were 81% in the 25-mg group, 42% in secondary outcome was the change at 52 weeks. the 75-mg group, and 25% in the 150-mg group. Exploratory outcomes included the cumulative The four study groups were similar with respect distribution of the change in the estimated GFR to baseline variables (Table 1); the mean age was from baseline to 52 weeks; the time to and per- 67 years. Blood glucose levels were generally well centage of patients with a reduction in the esti- controlled. A total of 98% of patients were receivmated GFR of 25% or more; the change in the ing an ACE inhibitor, an ARB, or both; the reestimated GFR 4 weeks after the last administra- maining patients could not tolerate such medication of the study drug; and the changes from tions. The mean estimated GFR was 32.46.9 ml baseline in the ACR and in levels of serum creati- per minute per 1.73 m2. The ACR was less than nine, blood urea nitrogen, serum phosphorus, uric 30 (normoalbuminuria) in 37% of patients, 30 to acid, magnesium, potassium, bicarbonate, glycated 300 (microalbuminuria) in 29%, and more than hemoglobin, and intact parathyroid hormone with 300 (macroalbuminuria) in 34%. bardoxolone methyl, as compared with placebo, Primary and Secondary End Points at 24 and 52 weeks. The estimated GFR increased within 4 weeks afStatistical Analysis ter the initiation of treatment in the three bardoxoWe calculated that 55 patients per group in a lon- lone methyl groups. The value peaked at 12 weeks gitudinal analysis would provide a power of ap- and remained relatively stable through 52 weeks proximately 90% to determine a mean (SD) differ- (Fig. 1A). At 24 weeks, there was significant imence in the change from baseline in the estimated provement in the primary end point (change GFR of 4.335.77 ml per minute per 1.73 m2 be- from baseline in the estimated GFR) in all bartween any dose of bardoxolone methyl and pla- doxolone methyl groups, as compared with the cebo. Our calculations assumed a rate of dropout placebo group, with mean differences per min(permanent discontinuation of a study drug be- ute per 1.73 m2 of 8.21.5 ml in the 25-mg group, fore 24 weeks) of 25%. The type I error rate was 11.41.5 ml in the 75-mg group, and 10.41.5 ml controlled at 0.05 through DunnettTamhane tests in the 150-mg group (P<0.001 for all compariat 0.019 for three comparisons with placebo.30 sons). The difference in the change between the We used a repeated-measures model to analyze 25-mg group and the 75-mg group was signifithe primary and secondary outcomes, with month- cant (P = 0.04), but the difference between the 75ly measurement of the estimated GFR through mg group and the 150-mg group was not signifi52 weeks as the response and with the baseline cant (P = 0.54). At 52 weeks, differences in the estimated GFR, ACR, and glycated hemoglobin changes in the estimated GFR, as compared with level as continuous covariates. The model included placebo (the secondary end point), continued to all measurements up to the time of withdrawal favor the bardoxolone methyl groups, with difn engl j med 365;4 nejm.org july 28, 2011

329

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 1. Demographic and Clinical Characteristics of the Patients.* Variable Placebo (N = 57) 25 mg (N = 57) Demographic characteristic Male sex no. (%) Age yr Race or ethnic group no. (%) White Black Other Hispanic or Latino Clinical characteristic Time from diabetes diagnosis to randomization yr Weight kg Body-mass index Blood pressure mm Hg Systolic Diastolic Glycated hemoglobin % Medications no. (%) ACE inhibitor, ARB, or both Statin Calcium-channel blocker Diuretic agent Antidiabetic drug Laboratory results Estimated GFR ml/min/1.73 m2 Serum creatinine mg/dl Albumin-to-creatinine ratio Median Interquartile range Level no. (%) <30 30300 >300 21 (37) 18 (32) 18 (32) 19 (33) 18 (32) 20 (35) 22 (39) 14 (25) 21 (37) 21 (38) 16 (29) 19 (34) 83 (37) 66 (29) 78 (34) 76 7549 126 11502 84 6766 94 6985 95 7678 31.26.3 2.00.5 32.97.0 2.00.5 33.06.6 2.00.5 32.37.6 2.00.6 32.46.9 2.00.5 57 (100) 45 (79) 35 (61) 51 (89) 54 (95) 55 (96) 46 (81) 31 (54) 48 (84) 56 (98) 57 (100) 49 (86) 34 (60) 47 (82) 56 (98) 54 (96) 49 (88) 31 (55) 51 (91) 54 (96) 223 (98) 189 (83) 131 (58) 197 (87) 223 (98) 130.513.5 67.39.0 7.21.2 129.512.6 68.78.3 7.20.9 130.213.2 69.78.1 7.31.0 130.511.8 68.68.7 7.11.1 130.212.7 68.68.5 7.21.0 17.19.9 95.222.8 34.48.0 18.210.8 103.022.9 36.37.8 17.89.8 97.723.1 35.07.6 18.69.8 103.625.3 35.87.3 18.010.0 99.823.7 35.47.7 44 (77) 10 (18) 3 (5) 18 (32) 44 (77) 10 (18) 3 (5) 12 (21) 41 (72) 11 (19) 5 (9) 21 (37) 44 (79) 11 (20) 1 (2) 11 (20) 173 (76) 42 (19) 12 (5) 62 (27) 28 (49) 67.710.0 34 (60) 66.99.2 33 (58) 66.18.7 33 (59) 66.79.2 128 (56) 66.89.3 Bardoxolone Methyl 75 mg (N = 57) 150 mg (N = 56) All Patients (N = 227)

* Plusminus values are means SD. Percentages may not total 100 because of rounding. To convert the values for creatinine to micromoles per liter, multiply by 88.4. ACE denotes angiotensin-converting enzyme, ARB angiotensin-receptor blocker, and GFR glomerular filtration rate. Race or ethnic group was self-reported. Patients of any race could identify themselves as Hispanic or Latino. The body-mass index is the weight in kilograms divided by the square of the height in meters. The ratio is based on measurement of albumin in milligrams and creatinine in grams.

ferences per minute per 1.73 m2 of 5.81.8 ml in 150-mg group (P<0.001). There were no signifithe 25-mg group (P = 0.002), 10.51.8 ml in the cant changes from baseline in the estimated GFR 75-mg group (P<0.001), and 9.31.9 ml in the in the placebo group at 24 or 52 weeks (Fig. 1A).
330
n engl j med 365;4 nejm.org july 28, 2011

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Bardoxolone Methyl in CKD with Type 2 Diabetes

Exploratory Outcomes

The cumulative distribution of changes in the estimated GFR at 52 weeks in patients completing 52 weeks of therapy showed decreases in 54% of patients in the placebo group but in only 20% of those in the group receiving 25 mg of bardoxolone methyl, 21% of those in the 75-mg group, and 27% of those in the 150-mg group (Fig. 1B). The proportion of patients with a reduction in the estimated GFR of 25% or more at 24 weeks was 13% (95% confidence interval [CI], 6 to 25) in the placebo group and 2% (95% CI, 1 to 6) in the combined bardoxolone methyl groups (P = 0.05) (Fig. 1C). Four weeks after the last administration of the study drug (at 56 weeks), the estimated GFR remained above the baseline value in patients receiving bardoxolone methyl, with differences from baseline per minute per 1.73 m2 of 0.71.6 in the 25-mg group, 2.51.6 ml in the 75-mg group, and 2.31.7 ml in the 150-mg group. In the placebo group, patients had a modest decrease of 0.61.1 ml per minute per 1.73 m2 from baseline (Fig. 1D). The change in the estimated GFR at 56 weeks was significantly correlated with the change at the end of treatment, at 52 weeks (r = 0.84, P<0.001) (Fig. 1E). In a post hoc sensitivity analysis that was based on the dose of bardoxolone methyl that was actually received at 52 weeks rather than the randomly assigned dose, increases in the estimated GFR were sustained for 52 weeks in the 75-mg and 150-mg groups but not in the 25-mg group (Fig. 3 in the Supplementary Appendix). At 24 weeks in all three bardoxolone methyl groups, there were significant decreases in levels of blood urea nitrogen, serum phosphorus, uric acid, and magnesium, as compared with placebo; these decreases had an inverse correlation with changes in the estimated GFR (Table 1 in the Supplementary Appendix). At 52 weeks, changes in these solutes continued to have an inverse correlation with changes in the estimated GFR (Fig. 4 in the Supplementary Appendix). In the 75-mg and 150-mg groups, there was a slight but significant increase in the ACR at 24 and 52 weeks (Table 1 in the Supplementary Appendix). Changes in the ACR and estimated GFR were positively correlated at 52 weeks (Fig. 5A in the Supplementary Appendix). At 56 weeks, the ACR generally returned to the baseline level (Fig. 5B in the Supplementary Appendix). Changes in the levels of potassium, bicarbonate, intact parathyroid hormone, and glycated hemoglobin did not differ
n engl j med 365;4

significantly among the study groups (Table 1 in the Supplementary Appendix).

Adverse Events

Adverse events were more frequent in the bardoxolone methyl groups than in the placebo group (Table 2, and Table 2 in the Supplementary Appendix). Most of the adverse events were mild to moderate in severity. The most common adverse event among patients receiving bardoxolone methyl was muscle spasm, with an incidence of 42% in the 25-mg group, 61% in the 75-mg group, and 59% in the 150-mg group, as compared with 18% in the placebo group. Muscle spasms, which were most frequent during the first 12 weeks of the study, primarily involved the calf and generally resolved without discontinuation of the study drug. Levels of lactate dehydrogenase were not significantly increased. Although hypomagnesemia was more common among patients receiving bardoxolone methyl than among those receiving placebo, hypomagnesemia was not present before or at the time of the spasms in 70% of the patients in the bardoxolone methyl groups who had muscle spasms (data not shown). Other adverse events that were more common in patients receiving bardoxolone methyl were mild elevations in alanine aminotransferase levels and gastrointestinal effects. Of the 170 patients who received bardoxolone methyl, 120 (71%) had transient aminotransferase elevations that peaked within 2 to 4 weeks after the initiation of treatment or an increase in the dose and generally resolved while the patients continued to take the drug. A total of 18 patients (11%) had an alanine aminotransferase elevation of more than three times the upper limit of the normal range. Persistent elevations at such levels were not observed, and such elevations did not recur once they had resolved. Total bilirubin levels were unchanged in patients with aminotransferase elevations, except in one patient with a posthepatic obstruction that was deemed to be unrelated to the study drug. Signs and symptoms of hepatic injury (i.e., severe, persistent, and concomitant pain in the right upper quadrant, jaundice, rash, vomiting, or malaise) were not observed. There was no clear dose-related trend in blood pressure; blood-pressure values were variable, and doses of antihypertensive medications were adjusted during the study in most patients. Despite increases in blood pressure that were noted in some patients, there was no correlation between
july 28, 2011

nejm.org

331

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Mean Estimated GFR (ml/min/1.73 m2)

No. at Risk
55 52 53 52 54 52 54 51 54 51 53 49 53 49 52 47 53 48 51 48 53 48 52 48 48 44 48 46 48 46 50 47 50 47 48 48 51 48 52 52 52 52 53 53 51 54 51 55 0
30

Proportion of Patients with a Given Change in Estimated GFR

Patients with 25% Reduction in Estimated GFR (%)

Change in Estimated GFR (ml/min/1.73 m2) at 56 Wk

Change in Estimated GFR (ml/min/1.73m2) at 52 Wk

332
B
46 40 Placebo 25 mg 75 mg 150 mg
0 4 8 12 16 20 24 28 32 36 40 44 48 52

34 100 28 0 75

Week 52

50 Placebo (n=48) 25 25 mg (n=45) 75 mg (n=39) 150 mg (n=40) The

15 0 15 30 45 60

100 90 80 70 60 50 40 30 20 10 0
8 52 12 16 20 24 28 32 36 40 44 48

Week

57 57

56 53

57

55

Change from Baseline in Estimated GFR (ml/min/1.73 m2)

Placebo Bardoxolone methyl, 25 mg Bardoxolone methyl, 75 mg Bardoxolone methyl, 150 mg

56

55

n e w e ng l a n d j o u r na l

n engl j med 365;4

C
25 20 15 10 5 0
0 4 8 12 16 20 24 28 32 36 40 44 48 52

D
Placebo 25 mg 30 25 20 15 10 5 0 5 10 15 20 25 75 mg 150 mg

E
40 30 20 10 0 10 20

100

90

nejm.org
12 16 20 24 28 32 36 40 44 48 52

80

70

60

of

50

40

july 28, 2011

30

20

m e dic i n e

10

Week

No. at Risk

57 55 54 52 52 51 48 48 47 47 45 45 44 32 57 56 54 53 52 51 49 49 48 47 44 44 43 35

r=0.84 P<0.001

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.
P c la eb o 25 m g 75 m g 0 15 m g
20 10 0 10 20 30

40

57 57 56 54 53 51 51 50 49 48 47 45 45 34

Placebo Bardoxolone methyl, 25 mg Bardoxolone methyl, 75 mg Bardoxolone methyl, 150 mg

56 55 54 54 54 53 52 51 49 47 44 43 41 33

Change in Estimated GFR (ml/min/1.73 m2) at 56 Wk

Bardoxolone Methyl in CKD with Type 2 Diabetes

Figure 1 (facing page). Effects of Bardoxolone Methyl on the Estimated Glomerular Filtration Rate (GFR). Panel A shows a repeated-measures model of the mean estimated GFR over time in the intention-to-treat population, with a limited range shown in the inset graph to highlight differences between groups. Significant differences in the change from baseline in the estimated GFR in each bardoxolone methyl group, as compared with the placebo group, were observed at 24 weeks (P<0.001 for all comparisons) and at 52 weeks (P = 0.002 for the comparison with the 25-mg group and P<0.001 for the comparisons with the 75-mg and 150-mg groups). The I bars indicate standard errors. Panel B shows the cumulative distribution of changes in the estimated GFR at 52 weeks for patients who completed 52 weeks of study. Panel C shows the time to the first reduction in the estimated GFR of 25% or more through 52 weeks, with a limited range shown in the inset graph to highlight between-group differences. The numbers of patients at risk (i.e., patients who were still being followed and did not have a 25% decline in the estimated GFR) during each 4-week period are presented below the x axis. The difference between all 170 patients receiving bardoxolone methyl and the 57 patients receiving placebo had borderline significance (P = 0.05 by the log-rank test); the differences between the dose-specific groups and the placebo group were not significant (25-mg group, P = 0.13; 75-mg group, P = 0.12; and 150-mg group, P = 0.22). Points at which data were censored are indicated by solid circles. Panel D shows the distribution of the change from baseline in the mean estimated GFR at 56 weeks (4 weeks after study-drug discontinuation) for 48 patients in the placebo group, 45 patients in the 25-mg group, 39 patients in the 75-mg group, and 40 patients in the 150-mg group. The open box displays the interquartile range. The bar within the box is the median, and the solid square is the mean. The lower whisker indicates the smallest observed value that is at least 1.5 times the interquartile range below the 25th percentile. The upper whisker indicates the largest observed value that is at least 1.5 times the interquartile range above the 25th percentile. Panel E shows the correlation between the change in the estimated GFR at 52 weeks and the change at 56 weeks (r = 0.84, P<0.001).

was similar in all study groups (Table 2, and Table 5 in the Supplementary Appendix). Discontinuations and withdrawals occurred primarily during the first 24 weeks of the study (Fig. 7 in the Supplementary Appendix). Sixteen patients discontinued bardoxolone methyl because of adverse events or changes in their medical status, with nine patients having more than one event. Events leading to discontinuation that were adjudicated to be possibly related to bardoxolone methyl included muscle spasms in six patients (4%), nausea and vomiting in three patients (2%), weight change in three patients (2%), and decreased appetite in two patients (1%) (Table 6 in the Supplementary Appendix). These events resolved after the discontinuation of bardoxolone methyl in all patients. One death occurred after cardiac-bypass surgery in a patient in the 75-mg group.

Discussion
Therapeutic interventions that suppress inflammation and oxidative stress may address both short-term (dynamic) and long-term (structural) contributors to a decline in the GFR in patients with CKD and could possibly stabilize or even improve kidney function.12 Bardoxolone methyl activates the Keap1Nrf2 pathway, which regulates inflammation and oxidative stress.18 The deletion of Nrf2 in murine models results in structural and functional kidney defects that are associated with increased inflammation and oxidative stress.19-21,32,33 In a model of acute kidney injury, bardoxolone methyl induced Nrf2 expression (messenger RNA and protein), as well as its activity, which was measured indirectly through the expression of heme oxygenase 1, which is known to decrease the amount of reactive oxygen species and decrease inflammation.34 In this model, the use of bardoxolone methyl led to increased Nrf2 expression in glomerular endothelial cells and cortical peritubular capillaries, as well as to increased expression of heme oxygenase 1 in kidney tubules and interstitial leukocytes. A concomitant reduction in inflammation, decrease in the blood urea nitrogen level, and amelioration of both glomerular and tubular injury was observed.34 In our study, treatment with bardoxolone methyl for 52 weeks led to sustained, significant improvements in the estimated GFR among patients receiving standard medical care for CKD
july 28, 2011

a change in blood pressure and the estimated GFR (Fig. 6 in the Supplementary Appendix). Weight reduction was observed in all the study groups (Table 1 in the Supplementary Appendix) and was more pronounced in patients with an increased body-mass index at baseline than in other patients (Table 3 in the Supplementary Appendix). Among patients receiving bardoxolone methyl, reductions in weight were independent of the magnitude of change in the estimated GFR (Table 4 in the Supplementary Appendix). The total incidence of serious adverse events
n engl j med 365;4

nejm.org

333

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 2. Most Common Adverse and Serious Adverse Events by Week 56.* Adverse Event Placebo (N = 57) Bardoxolone Methyl 25 mg (N = 57) Any adverse event Hypoglycemia Upper respiratory tract infection Peripheral edema Muscle spasms Hyperkalemia Diarrhea Fatigue Arthralgia Hypertension Nausea Hypomagnesemia Decreased appetite Serious adverse event Renal failure (acute or chronic) Congestive heart failure Pneumonia Hypoglycemia Noncardiac chest pain Anemia Hypertension Syncope Asthenia Osteoarthritis Atrial fibrillation Viral infection Costochondritis Renal failure 3 (5) 3 (5) 2 (4) 1 (2) 1 (2) 1 (2) 1 (2) 1 (2) 1 (2) 0 0 0 0 0 3 (5) 2 (4) 1 (2) 2 (4) 1 (2) 1 (2) 0 0 0 1 (2) 2 (4) 2 (4) 2 (4) 2 (4) 2 (4) 2 (4) 2 (4) 1 (2) 2 (4) 0 1 (2) 0 0 2 (4) 0 0 0 0 2 (4) 2 (4) 1 (2) 2 (4) 1 (2) 3 (5) 2 (4) 2 (4) 2 (4) 1 (2) 1 (2) 0 0 0 30 (53) 13 (23) 11 (19) 10 (18) 10 (18) 9 (16) 7 (12) 7 (12) 6 (11) 5 (9) 3 (5) 2 (4) 29 (51) 9 (16) 11 (19) 24 (42) 12 (21) 14 (25) 8 (14) 6 (11) 10 (18) 10 (18) 12 (21) 8 (14) 30 (53) 5 (9) 10 (18) 35 (61) 8 (14) 13 (23) 6 (11) 12 (21) 12 (21) 15 (26) 14 (25) 13 (23) 36 (64) 7 (12) 11 (20) 33 (59) 13 (23) 6 (11) 13 (23) 5 (9) 11 (20) 12 (21) 18 (32) 9 (16) 75 mg (N = 57) 150 mg (N = 56)

number of patients (percent)

* Listed are all adverse events that occurred in at least 20% of patients in any group. Serious adverse events were reported in at least two patients in any group. All events are ranked according to the incidence in the placebo group.

and type 2 diabetes. The estimated GFR increased in most patients receiving bardoxolone methyl but in less than half the patients in the placebo group. The time course of changes in the estimated GFR among patients receiving bardoxolone methyl suggests that there was a shortterm effect during the first 12 weeks, which was followed by a longer-term effect. The estimated GFR remained above baseline levels 4 weeks after
334
n engl j med 365;4

the discontinuation of bardoxolone methyl (i.e., approximately 17 half-lives of the drug). Although additional analyses are necessary, we speculate that such sustained increases in the estimated GFR after the discontinuation of bardoxolone methyl could be due to a decrease in the inflammation and oxidative stress associated with CKD. Moreover, increases in the estimated GFR occurred independently of changes in albumin excretion, a finding similar to that reported for patients with diabetic nephropathy who were treated with pirfenidone, a drug that inhibits inflammatory cytokines.35 The absence of a significant decline in the estimated GFR among patients receiving placebo may reflect the fact that most patients did not have macroalbuminuria at baseline and their important risk factors (e.g., blood pressure and glycated hemoglobin levels) were well controlled with standard medical care. In a previous study, therapy with bardoxolone methyl was not associated with a short-term increase in biomarkers of renal injury (i.e., N-acetyl-d-glucosaminidase and neutrophil gelatinaseassociated lipocalin), nor did it alter creatinine production or excretion.28 We did not assess these biomarkers in our study. Concurrent, correlated improvements in levels of blood urea nitrogen, uric acid, phosphate, and magnesium suggest that bardoxolone methyl has effects on multiple blood chemical values. Among these measures, blood urea nitrogen, which is not actively secreted by the kidney tubules, had the highest correlation with the estimated GFR. Although we did not directly assess tubular function, the decreased correlation between the estimated GFR and levels of uric acid, phosphate, and magnesium (which are handled by the kidney tubules) indicate that bardoxolone methyl may have a direct effect on solute transport in the tubules. Patients receiving bardoxolone methyl had slight but significant increases in the mean ACR, for which the mechanism is unknown. We speculate that such increases may result from increased glomerular filtration and decreased tubular resorption of protein due to an increased rate of tubular transit. Decreases in the ACR 4 weeks after the discontinuation of bardoxolone methyl may indicate that changes in the ACR are reversible. There was no correlation between changes in the estimated GFR and in blood pressure. In addition, some patients had an increase in the estimated GFR without an increase in blood pressure. Although increases in blood pressure may
july 28, 2011

nejm.org

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Bardoxolone Methyl in CKD with Type 2 Diabetes

have partly accounted for the increase in the estimated GFR in some patients, we believe it was not the predominant mechanism. This lack of correlation between blood-pressure changes and estimated GFR changes does not support the possibility of a short-term change in glomerular hemodynamics (hyperfiltration), followed by an accelerated decline in the GFR. Our study did not assess muscle mass, but the changes in the estimated GFR do not appear to be dependent on changes in body weight among patients receiving bardoxolone methyl. Several of the most common adverse events were consistent with known pharmacologic effects of bardoxolone methyl. Because Nrf2 expression is ubiquitous, multiple physiological and biochemical changes may be observed in response to bardoxolone methyl and may be explained, at least in part, by the presumed mechanism of action. Transient, self-limited elevations in aminotransferase levels in patients receiving bardoxolone methyl resolved without drug withdrawal and were asymptomatic. Liver biopsies were performed in two patients on the basis of increased alanine aminotransferase levels; each patient had chronic disease that was deemed to be unrelated to the use of bardoxolone methyl (data not shown). Nrf2 has been shown to increase alanine aminotransferase levels transcriptionally.36 In preclinical studies in a model of hepatic injury, mild increases in aminotransferase levels that were caused by both genetic and bardoxolone methyl analoguemediated Nrf2 activation were not associated with hepatic toxic effects. Nrf2 activation, in fact, inhibits acute inflammatory hepatic injury, despite the increase in aminotransferase levels.37
References
1. Kidney Disease Outcomes Quality Ini-

Among patients receiving bardoxolone methyl, muscle spasms were not associated with increases in lactate dehydrogenase levels, a marker of muscle damage, or with hypomagnesemia. The most common site of muscle spasms was the calf. Preclinical data showed that bardoxolone methyl increased glucose uptake in the calf muscle in mice with diabetes.38 Furthermore, insulin-mediated glucose uptake is associated with muscle spasms in humans. We hypothesize that muscle spasms in patients receiving bardoxolone methyl may be associated with a similar mechanism. Our study has several limitations. The doseadjustment design and the intention-to-treat analysis precluded a full assessment of dose-related efficacy, since many patients did not reach the target dose. Another limitation is the surrogate primary end point (the change in the estimated GFR). Measurement of the GFR to validate these results would be important. The confirmation of clinical benefit will require a larger, long-term study involving the assessment of clinical outcomes. In conclusion, patients with advanced CKD and type 2 diabetes who received treatment with bardoxolone methyl had sustained increases in the estimated GFR throughout the 52-week study period, a finding that is consistent with an improvement in kidney function. Thus, bardoxolone methyl appears to be an attractive therapeutic candidate for further study in patients with CKD.
Supported by Reata Pharmaceuticals. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank Drs. Michael Sporn, Gordon Gribble, Tadashi Honda, Nanjoo Suh, and Karen Liby for their discovery and characterization of bardoxolone methyl; and Dr. Edmund Doherty, Danielle Wrolstad, and Ritu Rajan for their assistance in the preparation of the manuscript.

tiative. KDOQI clinical practice guidelines and clinical practice recommendations for diabetes and chronic kidney disease. Am J Kidney Dis 2007;49:Suppl 2:S12-S154. 2. Stenvinkel P. Chronic kidney disease: a public health priority and harbinger of premature cardiovascular disease. J Intern Med 2010;268:456-67. 3. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu C. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296-305. 4. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification,

and stratification. Am J Kidney Dis 2002; 39:Suppl 1:S1-S266. 5. Barnett AH, Bain SC, Bouter P, et al. Angiotensin-receptor blockade versus convertingenzyme inhibition in type 2 diabetes and nephropathy. N Engl J Med 2004; 351:1952-61. [Erratum, N Engl J Med 2005;352:1731.] 6. Mann JF, Schmieder RE, McQueen M, et al. Renal outcomes with telmisartan, ramipril, or both, in people at high vascular risk (the ONTARGET study): a multicentre, randomised, double-blind, controlled trial. Lancet 2008;372:547-53. 7. Navaneethan SD, Nigwekar SU, Sehgal AR, Strippoli GF. Aldosterone antagonists for preventing the progression of chronic kidney disease: a systematic ren engl j med 365;4 nejm.org july 28, 2011

view and meta-analysis. Clin J Am Soc Nephrol 2009;4:542-51. 8. Onuigbo MA. Can ACE inhibitors and angiotensin receptor blockers be detrimental in CKD patients? Nephron Clin Pract 2011;118:c407-c419. 9. Rahman M, Pressel S, Davis BR, et al. Renal outcomes in high-risk hypertensive patients treated with an angiotensinconverting enzyme inhibitor or a calcium channel blocker vs a diuretic: a report from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Arch Intern Med 2005;165:93646. 10. Garrido AM, Griendling KK. NADPH oxidases and angiotensin II receptor signaling. Mol Cell Endocrinol 2009;302:148-58.

335

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Bardoxolone Methyl in CKD with Type 2 Diabetes

11. Marrero MB, Fulton D, Stepp D, Stern 21. Ma Q, Battelli L, Hubbs AF. Multior-

DM. Angiotensin II-induced signaling pathways in diabetes. Curr Diabetes Rev 2005;1:197-202. 12. Cachofeiro V, Goicochea M, de Vinuesa SG, Oubia P, Lahera V, Luo J. Oxidative stress and inflammation, a link between chronic kidney disease and cardiovascular disease. Kidney Int Suppl 2008;111:S4-S9. 13. Lemley KV, Abdullah I, Myers BD, et al. Evolution of incipient nephropathy in type 2 diabetes mellitus. Kidney Int 2000; 58:1228-37. [Erratum, Kidney Int 2000;58: 2257.] 14. Hayashi H, Karasawa R, Inn H, et al. An electron microscopic study of glomeruli in Japanese patients with non-insulin dependent diabetes mellitus. Kidney Int 1992;41:749-57. 15. Mauer SM, Steffes MW, Ellis EN, Sutherland DE, Brown DM, Goetz FC. Structural-functional relationships in diabetic nephropathy. J Clin Invest 1984; 74:1143-55. 16. Fioretto P, Caramori ML, Mauer M. The kidney in diabetes: dynamic pathways of injury and repair: the Camillo Golgi Lecture 2007. Diabetologia 2008; 51:1347-55. 17. Raptis AE, Viberti G. Pathogenesis of diabetic nephropathy. Exp Clin Endocrinol Diabetes 2001;109:Suppl 2:S424-S437. 18. Dinkova-Kostova AT, Liby KT, Stephenson KK, et al. Extremely potent triterpenoid inducers of the phase 2 response: correlations of protection against oxidant and inflammatory stress. Proc Natl Acad Sci U S A 2005;102:4584-9. 19. Li J, Stein TD, Johnson JA. Genetic dissection of systemic autoimmune disease in Nrf2-deficient mice. Physiol Genomics 2004;18:261-72. 20. Yoh K, Itoh K, Enomoto A, et al. Nrf2deficient female mice develop lupus-like autoimmune nephritis. Kidney Int 2001; 60:1343-53.

gan autoimmune inflammation, enhanced lymphoproliferation, and impaired homeostasis of reactive oxygen species in mice lacking the antioxidant-activated transcription factor Nrf2. Am J Pathol 2006;168: 1960-74. 22. Yates MS, Tauchi M, Katsuoka F, et al. Pharmacodynamic characterization of chemopreventive triterpenoids as exceptionally potent inducers of Nrf2-regulated genes. Mol Cancer Ther 2007;6:154-62. 23. Sporn MB, Liby KT, Yore MM, Fu L, Lopchuk JM, Gribble GW. New synthetic triterpenoids: potent agents for prevention and treatment of tissue injury caused by inflammatory and oxidative stress. J Nat Prod 2011;74:537-45. 24. Levonen AL, Landar A, Ramachandran A, et al. Cellular mechanisms of redox cell signalling: role of cysteine modification in controlling antioxidant defences in response to electrophilic lipid oxidation products. Biochem J 2004;378:373-82. 25. Ahmad R, Raina D, Meyer C, Kharbanda S, Kufe D. Triterpenoid CDDO-Me blocks the NF-{kappa}B pathway by direct inhibition of IKKbeta on Cys-179. J Biol Chem 2006;281:35764-9. 26. Straus DS, Pascual G, Li M, et al. 15-Deoxy-delta 12,14-prostaglandin J2 inhibits multiple steps in the NF-kappa B signaling pathway. Proc Natl Acad Sci U S A 2000;97:4844-9. 27. Serhan CN, Chiang N, Van Dyke TE. Resolving inflammation: dual anti-inflammatory and pro-resolution lipid mediators. Nat Rev Immunol 2008;8:349-61. 28. Pergola P, Krauth M, Huff JW, et al. Effect of bardoxolone on kidney function in patients with T2D and stage 3b-4 CKD. Am J Nephrol 2011;33:469-76. 29. Levey AS, Coresh J, Greene T, et al. Using standardized serum creatinine values in the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate. Ann Intern Med

2006;145:247-54. [Erratum, Ann Intern Med 2008;149:519.] 30. Dunnett CW, Tamhane ACA. Step-up multiple test procedure. J Am Stat Assoc 1992;87:162-70. 31. Woflinger R. Covariance structure selection in general mixed models. Comm Statist Simulation Comput 1993;22:1079106. 32. Itoh K, Mochizuki M, Ishii Y, et al. Transcription factor Nrf2 regulates inflammation by mediating the effect of 15-deoxy-Delta(12,14)-prostaglandin j(2). Mol Cell Biol 2004;24:36-45. 33. Kobayashi M, Yamamoto M. Molecular mechanisms activating the Nrf2-Keap1 pathway of antioxidant gene regulation. Antioxid Redox Signal 2005;7:385-94. 34. Wu QQ, Wang Y, Senitko M, et al. Bardoxolone methyl (BARD) ameliorates ischemic AKI and increases expression of protective genes Nrf2, PPAR{gamma}, and HO-1. Am J Physiol Renal Physiol 2011;300:F1180-F1192. 35. Sharma K, Ix JH, Mathew AV, et al. Pirfenidone for diabetic nephropathy. J Am Soc Nephrol 2011;22:1144-51. 36. Osburn WO, Yates MS, Dolan PD, et al. Genetic or pharmacologic amplification of nrf2 signaling inhibits acute inflammatory liver injury in mice. Toxicol Sci 2008;104:218-27. 37. Malhotra D, Portales-Casamar E, Singh A, et al. Global mapping of binding sites for Nrf2 identifies novel targets in cell survival response through ChIP-Seq profiling and network analysis. Nucleic Acids Res 2010;38:5718-34. 38. Saha PK, Reddy VT, Konopleva M, Andreeff M, Chan L. The triterpenoid 2-cyano3,12-dioxooleana-1,9-dien-28-oic-acid methyl ester has potent antidiabetic effects in diet-induced diabetic mice and Lepr(db/ db) mice. J Biol Chem 2010;285:40581-92.
Copyright 2011 Massachusetts Medical Society.

336

n engl j med 365;4

nejm.org

july 28, 2011

The New England Journal of Medicine Downloaded from nejm.org by emmanuel ricky on November 4, 2011. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.