Creatinine Fluctuation in Patients With Lupus Nephritis: Considerations for Clinical Trial Endpoints

Lupus nephritis (LN) is a common manifestation of systemic lupus erythematosus (SLE) and a major driver of morbidity and mortality. Proliferative forms of LN are typically managed with immunosuppressive therapy, with the aim of attenuating renal inflammation and preserving kidney function.1Almaani S. Meara A. Rovin B.H. Update on lupus nephritis.Clin J Am Soc Nephrol. 2017; 12: 825-835Crossref PubMed Scopus (306) Google Scholar Unfortunately, despite several clinical trials of LN conducted over the past 30 to 40 years, none has translated into new Food and Drug Administration (FDA)−approved therapies. Multiple issues in clinical trial design likely contributed to these negative outcomes, such as confounding background medications (especially high-dose glucocorticoids), trial duration, and the choice of trial endpoints. Most trials incorporated composite endpoints to define clinical response based on proteinuria and kidney function. Kidney function was generally assessed as a change in estimated glomerular filtration rate or serum creatinine (sCr) as compared to the patients’ values at trial entry.2Mackay M. Dall'Era M. Fishbein J. et al.Establishing surrogate kidney end points for lupus nephritis clinical trials: development and validation of a novel approach to predict future kidney outcomes.Arthritis Rheum. 2019; 71: 411-419Crossref PubMed Scopus (16) Google Scholar Thresholds were then chosen to define the extent of the clinical response (complete, partial, or no response). However, evidence supporting these thresholds is not robust. For example, most studies used a proteinuria cutoff of <0.3 to 0.5 g/d to describe complete response; however, a post hoc analysis of 2 large LN trials demonstrated that proteinuria levels of <0.7 to 0.8 g/d after 1 year of treatment predicted favorable long-term kidney outcomes, suggesting that a less stringent proteinuria cutoff may be reasonable. Similarly, an sCr value <15% above baseline has often been required for complete response in LN trials.S1−S5 However, day-to-day variations in sCr measurements are routinely observed in clinical practice in patients with and without chronic kidney disease, even when measured within a 24-hour period.3Hilderink J.M. van der Linden N. Kimenai D.M. et al.Biological variation of creatinine, cystatin C, and eGFR over 24 hours.Clin Chem. 2018; 64: 851-860Crossref PubMed Scopus (9) Google Scholar To determine a threshold of kidney function that accounts for expected day-to-day variations, we investigated the fluctuation of sCr in a cohort of patients with LN who were complete renal responders. Individual data from 574 patients participating in 3 clinical trials and several real-world observational cohorts were reviewed.2Mackay M. Dall'Era M. Fishbein J. et al.Establishing surrogate kidney end points for lupus nephritis clinical trials: development and validation of a novel approach to predict future kidney outcomes.Arthritis Rheum. 2019; 71: 411-419Crossref PubMed Scopus (16) Google Scholar After excluding 283 patients because they did not have at least 3 sCr measurements after having achieved a complete renal response (defined as proteinuria ≤0.5 g/d), data from 291 patients were analyzed. Patients were classified as upward fluctuators and nonfluctuators based on having at least 1 sCr measurement >115% of baseline or no sCr measurement >115% of baseline, respectively (Table 1). There were no differences between the 2 groups based on race/ethnicity, sex, biopsy histopathologic class, proteinuria levels, or induction immunosuppressive regimen. The upward fluctuators had a lower baseline creatinine (0.70 vs. 0.86, P < 0.0001) and were younger (28.5 vs. 33, P = 0.006) than the nonfluctuators. The median number of sCr measurements per patient was slightly higher in the upward fluctuator group (8 vs. 7, P = 0.02), and the upward fluctuators had a numerically longer follow-up (Table 2). In the fluctuator group, 33%, 20%, and 12% of each patient’s sCr measurements were 15%, 20%, and 25% above their baseline sCr, respectively (Table 3). The variables age, follow-up time, sex, number of sCr measurements, and baseline sCr were evaluated in a multivariable logistic regression model. Future fluctuation of sCr was significantly associated with the number of sCr measurements made (odds ratio, 1.09; 95% confidence interval, 1.01−1.16 for each additional measurement), and a lower baseline sCr (odds ratio, 0.02; 95% confidence interval, 0.00−0.05 for each 1-mg/dl increase in baseline sCr).Table 1Patient and disease characteristicsCharacteristicUpward fluctuatorNonfluctuatorP valueTotal84 (28.9)207 (71.1)Ethnicity0.19 Caucasian44 (52.4)104 (50.2) Asian21 (25.0)71 (34.3) African descent9 (10.7)20 (9.7) Other10 (11.9)12 (5.8)Sex0.14 Female79 (94.0)181 (87.4) Male5 (6.0)26 (12.6)Age28.5 (22−37)33 (25−41)0.01Lupus nephritis class0.79 Proliferative61 (76.3)160 (79.6) Pure membranous10 (12.5)22 (10.9) Mixed9 (11.3)19 (9.5)Baseline C0.7 (0.6−0.8)0.86 (0.8−1.01)<0.001Baseline proteinuria0.3 (0.18−0.40)0.3 (0.17−0.4)0.37Induction regimenaMAINTAIN Nephritis Trial (Mycophenolate Mofetil Versus Azathioprine for Maintenance Therapy of Lupus Nephritis) data not included.0.42 AZA2 (2.9)3 (1.9) MMF32 (47.1)62 (39.0) CYC34 (50.0)94 (59.1)AZA, azathioprine; Cr, creatinine; CYC, cyclophosphamide; MMF, mycophenolate mofetil.Data are n (%) or median (interquartile range).By convention, P values that are less than 0.05 are considered statistically significant and bolded.a MAINTAIN Nephritis Trial (Mycophenolate Mofetil Versus Azathioprine for Maintenance Therapy of Lupus Nephritis) data not included. Open table in a new tab Table 2Sample breakdown in each groupMeasurement characteristicsUpward fluctuatorNonfluctuatorP valueTotal no. of sCr measurements7411598No. of sCr measurements per patient, median (IQR)8 (5−11.25)7 (4.5−10)0.022Mo of response, median (IQR)29.75 (18.5−37.5)24.1 (16.5−36)0.054IQR, interquartile range; sCr, serum creatinine.By convention, P value that is less than 0.05 is considered statistically significant and is bolded. Open table in a new tab Table 3Serum creatinine fluctuates widely in the upward fluctuator groupsCr Measurement characteristicsTotal, nPer patient, median (IQR)% of Patient’s no. of follow-up samples, median (IQR)Total no. of sCr measurements7418 (5−11.25)—sCr measurements >115% of baseline2242 (1−4)33.3% (17.8−50.0)sCr measurements >120% of baseline1421 (1−2)20.0% (10.8−33.3)sCr measurements >125% of baseline1091 (0−2)11.8% (0.0−27.0)IQR, interquartile range; sCr, serum creatinine. Open table in a new tab AZA, azathioprine; Cr, creatinine; CYC, cyclophosphamide; MMF, mycophenolate mofetil. Data are n (%) or median (interquartile range). By convention, P values that are less than 0.05 are considered statistically significant and bolded. IQR, interquartile range; sCr, serum creatinine. By convention, P value that is less than 0.05 is considered statistically significant and is bolded. IQR, interquartile range; sCr, serum creatinine. To assess whether upward fluctuators developed a persistent change in sCr over time, suggestive of a progressive decline or improvement in kidney function, each individual’s serial sCr values were plotted and analyzed with simple linear regression (Figure 1) and analyzed with simple linear regression. Only 7 of the 84 upward fluctuators had a positive β with an unadjusted P value < 0.05, suggesting that the vast majority of patients did not have a progressive increase or decrease in serum creatinine. The median β of those 7 patients was 0.087 (interquartile range, 0.06−0.13), translating into an increase in serum creatinine of 0.087 mg/dl for every year of follow-up. A complete renal response is often the endpoint in clinical trials of new therapeutics for LN. Although there is no consensus as to what constitutes a complete renal response, based on recent trials, patients whose final sCr is 10% to 30% higher than their baseline sCr are not considered complete responders.2Mackay M. Dall'Era M. Fishbein J. et al.Establishing surrogate kidney end points for lupus nephritis clinical trials: development and validation of a novel approach to predict future kidney outcomes.Arthritis Rheum. 2019; 71: 411-419Crossref PubMed Scopus (16) Google Scholar Using a large cohort of prospectively followed LN patients who achieved and maintained a complete clinical renal response, we found considerable variability in sCr over time. About 30% of patients had episodic increases in sCr of ≥15% without clinical consequences, and over 10% of patients had episodic increases in SCr of more than 25%. These natural and clinically inconsequential fluctuations of sCr could result in patients being labeled as partial responders or nonresponders, potentially affecting the outcome of a trial. This is especially important, because many trials measure sCr only once at the conclusion of the trial. The likelihood of having sCr fluctuations ≥15% increased in patients with low baseline sCr values. This is an important issue in LN, in which most patients are young women who often have low sCr levels normally. In such patients, small changes in sCr translate into higher percent changes and could be misinterpreted as a decline in kidney function. Intraindividual variability of sCr in healthy patients can be due to biological variability and technical variability of the assay used to measure sCr. In healthy individuals, the biological variability of sCr is about 4.5%.4Ricos C. Alvarez V. Cava F. et al.Current databases on biological variation: pros, cons and progress.Scand J Clin Lab Invest. 1999; 59: 491-500Crossref PubMed Scopus (874) Google Scholar,5Carobene A. Marino I. Coskun A. et al.The EuBIVAS Project: within- and between-subject biological variation data for serum creatinine using enzymatic and alkaline picrate methods and implications for monitoring.Clin Chem. 2017; 63: 1527-1536Crossref PubMed Scopus (49) Google Scholar Biological variability may be higher in a kidney that has suffered prior injury, such as flares of LN. The analytic variability of the Jaffe method, a commonly used technique for measuring sCr, is 5.5%.4Ricos C. Alvarez V. Cava F. et al.Current databases on biological variation: pros, cons and progress.Scand J Clin Lab Invest. 1999; 59: 491-500Crossref PubMed Scopus (874) Google Scholar Between biological and analytical variability, the smallest change between 2 sCr measurements in an individual that warrants clinical concern is 19%.6Delanaye P. Cavalier E. Pottel H. Serum creatinine: not so simple!.Nephron. 2017; 136: 302-308Crossref PubMed Scopus (87) Google Scholar Commonly used medications, such sulfamethoxazole/trimethoprim, angiotensin-converting enzyme inhibitors, and angiotensin blockers, can also cause fluctuations in sCr. We cannot exclude that a change in medications affected the measurements seen in our patients. In summary, variability in sCr measurements is commonly observed in clinical practice and does not necessarily indicate a decline or improvement in kidney function. In a cohort of lupus patients who have had at least 1 episode of LN, a large proportion of patients who otherwise appear to have achieved a stable complete renal response based on proteinuria criteria have sCr fluctuations of 15% or more, suggesting that a 15% cutoff to define the success of a trial may be overly conservative. To define complete renal response, we recommend that a 25% cutoff for the upper limit of change in sCr might be more appropriate. Furthermore, a single measurement of sCr at the end of a trial cannot be put into an appropriate context, and sCr should be measured on at least 2 occasions. SA reports receiving personal fees from Aurinia Pharmaceuticals Inc. outside of the submitted work. CA reports receiving nonfinancial support from Bristol Myers Squibb and Glaxo-Smith Kline and a grant from Exagen, outside of the submitted work. NS is an employee and shareholder at Aurinia Pharmaceuticals Inc. BHR reports personal fees from Aurinia, Callidatis, Chemocentryx, Retrophin, Novartis, Morphosys, EMD Serono, Bristol Myers Squibb, Janssen, AstraZeneca; and Omeros; nonfinancial support from Lupus Foundation of America; and grants from the National Institutes of Health, outside the submitted work. All the other authors declared no competing interests. Download .pdf (.07 MB) Help with pdf files Supplementary File (PDF)

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.