Abstract
Background and objectives
Critically ill patients requiring RRT have higher circulating plasma concentrations of inflammatory and apoptosis markers that are associated with subsequent RRT dependence and death. Whether intensive dosing of RRT is associated with changes in specific mediators is unknown.
Design, setting, participants, & measurements
A multicenter, prospective, cohort study of 817 critically ill patients receiving RRT ancillary to the Veterans Affairs/National Institutes of Health Acute Renal Failure Trial Network study was conducted between November 2003 and July 2007. Plasma inflammatory (IL-6, IL-8, IL-10, IL-18, and macrophage migration inhibitory factor) and apoptosis (TNF receptor-I [TNFR-I], TNFR-II, and death receptor-5) biomarkers on days 1 and 8 were examined after initiation of intensive RRT. Whether intensive RRT, given day 1 biomarkers, is associated with RRT independence and lower mortality at day 60 was also examined.
Results
Overall, no differences were found in day 8 biomarker concentrations between intensive and less-intensive RRT groups. When adjusted for day 1 biomarkers and clinical variables, intensive RRT was not associated with renal recovery (adjusted odds ratio [OR], 0.80; 95% confidence interval, 0.56 to 1.14) or mortality (adjusted OR, 1.15; 95% confidence interval, 0.81 to 1.64). Use of intensive RRT, however, was associated with lower day 8 concentrations when day 1 plasma IL-6, macrophage migration inhibitory factor, and TNFR-I concentrations were high (interaction P value for all markers, <0.01). In contrast, day 8 marker concentrations were higher when day 1 levels were low (P<0.01). Elevated biomarker concentrations on day 8 among 476 participants were associated with lower renal recovery (adjusted OR range, 0.19–0.87) and higher mortality (adjusted OR range, 1.26–3.18).
Conclusions
Among critically ill patients receiving RRT, intensive dosing of RRT has variable association with biomarker concentration and no association with renal recovery and mortality. However, elevated concentrations of inflammatory and apoptosis markers on day 8 of RRT were associated with RRT dependence and death.
Keywords: kidney disease, apoptosis, renal failure
Introduction
Critically ill patients requiring RRT are at a higher risk of RRT dependence and death (1,2). By 2 months after acute illness, more than one half of patients die and a third of patients are dependent on RRT (1). This increased risk is not attenuated by intensive dosing of RRT compared with less-intensive RRT (1,2). Our recent work suggests that higher circulating concentrations of pro- and anti-inflammatory cytokines and apoptosis markers found in patients at the time of initiation of RRT are independently associated with nonrecovery of renal function and death. Specifically, we found that IL-6, IL-8, IL-10, IL-18, macrophage migration inhibitory factor (MIF), and TNF receptor-I (TNFR-I) and TNFR-II were associated with both reduced survival and renal recovery (3). Death receptor-5 (DR-5) was also associated with reduced survival (3). High-intensity RRT has been proposed to clear inflammatory markers in patients with AKI and improve outcomes (4,5). Whether intensive dosing of RRT modulates plasma cytokine concentration and whether modulation improves outcomes is unclear.
In this study, we first examined whether intensive, compared with less-intensive RRT, is associated with lower biomarker concentrations on day 8 relative to day 1. Second, we examined whether intensive RRT, given day 1 biomarker profile, is associated with renal recovery and lower mortality at day 60. Third, we examined the association between day 8 biomarker concentration and outcomes. Finally, we examined subgroups of patients on the basis of their biomarker levels at enrollment who would respond differently to the two RRT intensities.
Materials and Methods
Study Design and Selection of Participants
The Biological Markers of Recovery for the Kidney (BioMaRK) was a nested prospective observational cohort study conducted as an ancillary to the Veterans Affairs/National Institutes of Health Acute Renal Failure Trial Network (ATN) study between November 2003 and July 2007. The ATN study was a multicenter, clinical trial of two strategies of RRT in critically ill patients and is described elsewhere (1,6). Briefly, adults diagnosed with AKI requiring RRT with at least one nonrenal organ failure or sepsis were eligible. Patients with CKD were excluded. Participants were randomly assigned to receive either intensive or less-intensive RRT. The mean time interval between intensive care unit admission and initiation of RRT was 6.7±9.0 days, and the interval between the onset of AKI and randomization was 3.2±2.0 days.
Of participants receiving intensive RRT, intermittent hemodialysis was provided six times per week, and continuous venovenous hemodiafiltration was prescribed to provide an effluent flow rate of 35 ml/kg body wt per hour. In the less-intensive strategy, intermittent hemodialysis was provided three times per week, and continuous venovenous hemodiafiltration was prescribed to provide a total effluent flow rate of 20 ml/kg body wt per hour. Cellulose triacetate or synthetic membranes were used for all treatments, and the distribution of membranes was similar between the two RRT groups.
The ATN study found no overall difference in renal recovery or mortality between the intensive and less-intensive RRT groups. The BioMaRK study included all participants in the ATN study who gave additional written consent to blood collections for sample banking. We obtained approval from the institutional review boards of the University of Pittsburgh and all other participating sites.
Blood Sample Collection
Blood samples were collected after study randomization in the ATN trial (day 1) in 817 participants and before initiation of protocolized RRT dosing. Day 8 samples were collected in 568 participants who were alive and hospitalized irrespective of their RRT status. Details of the biomarker assays, detection threshold, and censoring are provided in Supplemental Table 1 and are also described in detail elsewhere (3).
Data Collection
We ascertained baseline characteristics, including demographics; cause of AKI; and other clinical, physiologic, and laboratory data at the time of initiation of RRT. Individual comorbid illnesses were assessed using the Charlson comorbidity score (7). Severity of illness was ascertained at enrollment using the Acute Physiology and Chronic Health Evaluation (APACHE-II) (8) and Cleveland Clinic Intensive Care Unit Acute Renal Failure score (9). We defined acute organ dysfunction as a new Sequential Organ Failure Assessment score of three or higher in any of the six organ systems (10). All participants were followed daily until hospital discharge, death, or day 28 after randomization, whichever occurred first.
Outcome Ascertainment
Our primary outcomes were renal recovery and mortality at day 60. Renal recovery was defined as being alive and independent from RRT by day 60, as specified a priori. Participants who became independent of RRT but died before day 60 were treated as nonrecovery throughout the study period. Outcomes were ascertained daily during hospitalization and at days 28 and 60 using telephone and/or mail follow-up (1).
Statistical Analyses
We first compared baseline characteristics by intensity of RRT in the BioMaRK and ATN cohorts. Continuous data were compared using the t test or Wilcoxon rank-sum test, and categorical data were compared using the chi-squared test or Fisher exact test. Left-censored biomarker data were imputed using the lower limit of detection. For all biomarker analyses, data were log transformed and analyzed in a natural logarithm scale. We compared biomarkers between intensive and less-intensive RRT groups on days 1 and 8. To examine the interaction between the day 1 concentration and the intensity of RRT, we fitted PROC GLM to model day 8 concentrations as a function of day 1 concentration. We then tested interactions using type III sum of squares.
To examine the association between intensity of RRT and clinical outcomes given day 1 biomarkers, we first examined the underlying relationship between individual markers and outcomes using the generalized additive model. We assessed linearity assumptions between biomarkers and outcomes using locally weighted scatterplot smoothing curves. We examined and report (Supplemental Table 2) the variation inflation factor for each biomarker, an index that measures how much the variance of an estimated regression coefficient is increased because of collinearity among markers.
We then calculated risk-adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs) using logistic regression adjusting for day 1 biomarkers. Because there are eight biomarkers and two outcomes, for all analyses we adjusted for multiple comparisons by setting the α value for the CI to 0.05/16=0.003, corresponding to a two-sided hypothesis test with a Bonferroni correction for 16 comparisons. Given that survivors versus nonsurvivors and those with recovery versus nonrecovery exhibit median marker concentrations approximately 2-fold different (e.g., IL-6, IL-8, MIF) (3), we calculated statistical power assuming that an OR of 2 (or 0.5) would be clinically meaningful. Using IL-6 as a standard, our sample is sufficient to detect even smaller ORs of 1.6 without Bonferroni correction or 1.8 with Bonferroni correction, assuming a power of 80%.
Because the relationship between the biomarkers and outcomes was nonlinear, we also examined the association between RRT intensity and outcomes within the stratum of log-quartiles of biomarker concentration. An exploratory interaction trees analysis (11) was also used to identify possible subgroups using day 1 marker concentrations as predictors and renal recovery and mortality as outcomes. We also examined association between day 8 individual markers and outcomes after adjusting for baseline characteristics and intensity of RRT using the aforementioned methods. Adjusted ORs were calculated for each natural log-unit increase in day 8 biomarker concentration. All statistical analyses were performed using SAS software version 9.3 (SAS Institute, Cary, NC) and R (R Core Team).
Results
Characteristics of Study Participants
Participants in the BioMaRK and ATN cohorts are shown in Table 1. Of the 817 participants in the BioMaRK study, 402 received intensive RRT (49.2%). Baseline characteristics, including demographics, comorbid conditions, and etiology of AKI, were similar between intensive and less-intensive RRT groups. However, minor differences were seen in terms of APACHE-II and Sequential Organ Failure Assessment scores. Participants receiving intensive RRT compared with less-intensive RRT had similar renal recovery (12.2% versus 15.6%, respectively; P=0.20) and mortality (26.1% versus 21.2%, respectively; P=0.10) on day 8 and at 2 months (renal recovery, 34.6% versus 38.3%, P=0.27; and mortality, 52.5% versus 49.2%, P=0.34, respectively). Premorbid serum creatinine data were available only in 683 participants (83.6%). There was no difference in missing premorbid creatinine values between those who did and did not recover (18.8% versus 15%, respectively; P=0.16) and survivors and nonsurvivors (16.1% versus 16.6%, respectively; P=0.84).
Table 1.
Characteristics of participants in the Biologic Markers of Recovery for the Kidney and Acute Renal Failure Trial Network cohorts
| Characteristic | BioMaRK Cohort | ATN Cohort | |||
|---|---|---|---|---|---|
| Intensive RRT (n=402) | Less-Intensive RRT (n=415) | P Valuee | All Participants (n=817) | All Participants (n=1124) | |
| Age, (yr) | 60.4±15.9 | 60.3±15 | 0.97 | 60.3±15.4 | 59.7±15.3 |
| Male | 287 (71.4) | 280 (67.5) | 0.22 | 567 (69.4) | 793 (71) |
| Race | |||||
| White | 310 (77.1) | 316 (76.1) | 0.85 | 626 (76.6) | 835 (74) |
| Black | 61 (15.2) | 63 (15.2) | 124 (15.2) | 179 (16) | |
| Hispanic | 24 (6) | 25 (6) | 49 (6) | 77 (7) | |
| Other | 7 (1.7) | 11 (2.6) | 18 (2.2) | 16 (3) | |
| Charlson comorbidity indexa | 2.6±2.6 | 2.3±2.3 | 0.07 | 2.5±2.4 | 4.3±2.9 |
| Cardiovascular disease | 146 (36.3) | 163 (39.3) | 0.38 | 309 (37.8) | |
| Liver disease | 47 (11.7) | 47 (11.3) | 0.87 | 94 (11.5) | |
| Diabetes | 126 (31.3) | 112 (27.0) | 0.17 | 238 (29.1) | — |
| Malignancy | 88 (21.9) | 69 (16.6) | 0.06 | 157 (19.2) | |
| Immunocompromised | 66 (16.4) | 57 (13.7) | 0.28 | 123 (15.1) | |
| Cause of AKI | |||||
| Ischemia | 322 (80.3) | 327 (78.8) | 0.59 | 649 (79.5) | 871 (77) |
| Nephrotoxins | 98 (24.4) | 100 (24.1) | 0.91 | 198 (24.3) | 286 (25) |
| Sepsis | 216 (53.9) | 208 (50.1) | 0.28 | 424 (52) | 579 (52) |
| Multifactorial | 217 (54.1) | 223 (53.7) | 0.91 | 440 (53.9) | 626 (56) |
| APACHE-II scoreb | 26.8±7.2 | 25.7±7 | 0.03 | 26.2±7.1 | 26.4±7.3 |
| Cleveland Clinic ICU Acute Renal Failure scorec | 11.9±3.2 | 11.6±3.3 | 0.24 | 11.8±3.3 | 12.1±3.3 |
| SOFA scored | 14.1±3.8 | 13.4±3.9 | 0.01 | 13.8±3.9 | 14.5±3.7 |
| Outcome at day 60 | |||||
| Mortality | 211 (52.5) | 204 (49.2) | 0.34 | 415 (50.8) | 591 (52.6) |
| Alive and independent from RRT | 139 (34.6) | 159 (38.3) | 0.27 | 298 (36.5) | 398 (35.4) |
Values are mean±SD or n (%). APACHE-II, Acute Physiology and Chronic Health Evaluation; ICU, intensive care unit; SOFA, Sequential Organ Failure Assessment; BioMaRK, Biologic Markers of Recovery for the Kidney; ATN, Acute Renal Failure Trial Network.
According to Charlson et al. without the age (7).
APACHE-II includes initial values of 12 routine physiologic measurements, age, and previous health status ranging from 1 to 71; a higher score is closely correlated with the subsequent risk of hospital death (8).
The Cleveland Clinic ICU Acute Renal Failure score can range from 1 to 20, with higher scores predictive of higher risk of death (9).
SOFA score includes six organ systems with scores ranging from 0 to 4 for each organ system (with the renal system score included in the calculation of the total SOFA score); higher scores indicate more severe organ dysfunction (10).
Denotes comparison between intensive and less-intensive RRT groups in the BioMaRK cohort.
Association between Intensity of RRT and Biomarker Concentration
Figure 1 shows the median biomarker concentrations and interquartile ranges in natural logarithm scale, on day 1 (n=817) and day 8 (n=568) of RRT, stratified by RRT intensity. Table 2 shows raw median biomarker concentrations. Overall, use of intensive RRT compared with less-intensive RRT was not associated with lower biomarker concentrations on day 8. We also found no difference in day 8 biomarker concentrations by RRT intensity within the subgroups of patients who only received intermittent hemodialysis or continuous RRT (Supplemental Table 3).
Figure 1.
Circulating plasma inflammatory and apoptosis biomarker concentrations on day 1 and day 8, stratified by intensity of RRT. Boxplot summaries of plasma inflammatory and apoptosis biomarker concentrations are displayed in natural logarithm scale. The vertical box represents the 25th percentile (bottom line), median (middle line), and 75th percentile (top line) values. The lowest datum (lower whisker) represents 1.5 times the interquartile range of the lower quartile, and the highest datum (upper whisker) represent 1.5 times the interquartile range of the upper quartile. Circles represent outliers. MIF, macrophage migration inhibitory factor; TNFR-I, TNF receptor-I; TNFR-II, TNF receptor-II.
Table 2.
Biomarker concentration stratified by intensity of RRT
| Biomarkers | Day 1 | Day 8 | ||
|---|---|---|---|---|
| Intensive RRT (n=402) | Less-Intensive RRT (n=415) | Intensive RRT (n=270) | Less-Intensive RRT (n=298) | |
| Inflammation | ||||
| IL-6 | 176 (74–604) | 152 (72.2–467) | 85.9 (42–184) | 85.6 (45.6–172) |
| IL-8 | 108 (53.2–349) | 91.8 (42.7–234) | 72.5 (36.5–150) | 64 (39.3–143) |
| IL-10 | 16.9 (7.3–44.6) | 15 (7.2–37) | 10 (5.6–19.2) | 9.3 (4.8–22.7) |
| IL-18 | 102 (43.4–201) | 101 (46.8–221) | 80.9 (35.4–157) | 90.1 (41.8–179) |
| MIF | 271 (85–746) | 240 (99.7–788) | 113 (45.6–256) | 127 (45.9–270) |
| Apoptosis | ||||
| TNFR-I | 13,491 (9504–18,823) | 12,448 (9716–17,106) | 13,875 (9207–17,996) | 13,230 (9062–17,750) |
| TNFR-II | 5697 (4283–7900) | 5411 (4063–7377) | 5491 (3840–7387) | 5061 (3510–7155) |
| DR-5 | 228 (134–393) | 230 (135–412) | 197 (108–369) | 198 (108–380) |
Values are medians (interquartile ranges). Median biomarker concentrations are expressed in pictograms per milliliter. Day 1 biomarkers were assayed in 817 participants except DR-5, which was assayed in 816 participants because of insufficient specimen. Day 8 markers were assayed in 568 participants for IL-6, IL-8, IL-10, TNF-I, and TNF-II. IL-18 and MIF were assayed on 567 participants and DR-5 on 566 participants. MIF, macrophage migration inhibitory factor; TNFR-I, TNF receptor-I; TNFR-II, TNF receptor-II; DR-5, death receptor-5.
Nevertheless, depending on the day 1 concentration, there was a statistically significant interaction between RRT intensity and day 8 concentrations (Figure 2). In participants with higher day 1 biomarker concentrations, use of intensive RRT compared with less-intensive RRT was associated with lower day 8 concentrations of plasma IL-6, MIF, and TNFR-I (interaction P<0.01 for all three markers) (Supplemental Figure 1). In contrast, among participants with lower day 1 biomarker concentration, use of intensive RRT was associated with higher day 8 concentrations (interaction P<0.01 for all three markers) (Supplemental Figure 1).
Figure 2.
Plots showing interaction between plasma biomarker concentration and intensity of RRT. Model-predicted curves were generated using day 1 and day 8 biomarker concentrations in the intensive and less-intensive RRT groups. The x axis represents day 1 and the y axis represents day 8 biomarker concentrations in natural logarithm scale. The solid line represents biomarker concentrations in the intensive RRT group, and the dashed line represents biomarker concentrations in the less-intensive RRT group. In participants with high day 1 biomarker concentrations, use of intensive RRT compared with less-intensive RRT was associated with lower day 8 concentrations for IL-6, MIF, and TNFR-I (P<0.01). In contrast, among participants with lower baseline day 1 biomarker concentrations of IL-6, MIF, and TNFR-I, use of intensive RRT compared with less-intensive RRT was associated with higher day 8 concentrations (P<0.01). MIF, macrophage migration inhibitory factor; TNFR-I, TNF receptor-I.
Association between Intensity of RRT and Renal Recovery and Mortality Given Day 1 Marker Concentrations
Overall, use of intensive RRT was not associated with renal recovery or mortality (Table 1). Table 3 shows unadjusted and adjusted ORs for intensive RRT compared with less-intensive RRT for renal recovery and mortality. When adjusted for differences in day 1 biomarker concentration and age, sex, race, Charlson comorbidity score, premorbid serum creatinine, chronic hypoxia, presence of liver disease, immunosuppressed state, nephrotoxin exposure, diagnosis of sepsis, presence of oliguria, use of mechanical ventilation, and APACHE-II score, intensive RRT was not associated with renal recovery (adjusted OR, 0.80; 95% CI, 0.56 to 1.14) or mortality (adjusted OR, 1.15; 95% CI, 0.81 to 1.64).
Table 3.
Association between intensive RRT and clinical outcomes
| Characteristic | Odds Ratios for Intensive versus Less-Intensive RRT (95% CI) (n=682) | |
|---|---|---|
| Renal Recovery | Mortality | |
| Unadjusted | 0.78 (0.57–1.06) | 1.22 (0.91–1.65) |
| Adjusted for day 1 biomarkersa | 0.79 (0.57–1.10) | 1.21 (0.88–1.66) |
| Adjusted for day 1 biomarkersa and clinical variablesb | 0.80 (0.56–1.14) | 1.15 (0.81–1.64) |
Multivariable logistic regression models with corresponding odds ratios were constructed to examine association between intensive versus less-intensive RRT on clinical outcomes. For the recovery model, an odds ratio <1 indicates that intensive RRT compared with less-intensive RRT is associated with lower renal recovery. For the mortality model, an odds ratio >1 indicates that intensive RRT compared with less-intensive RRT is associated with higher mortality. The models included 682 participants because of missing premorbid creatinine data in 134 participants and day 1 death receptor-5 marker in one participant. 95% CI, 95% confidence interval.
Day 1 biomarkers include IL-6, IL-8, IL-10, IL-18, macrophage migration inhibitory factor, TNF receptor-I, TNF receptor-II, and death receptor-5.
Adjusted for differences in baseline covariates including age, race, sex, Charlson comorbidity score, chronic hypoxia, liver disease, immunocompromised state, Acute Physiology and Chronic Health Evaluation score, premorbid serum creatinine, oliguria, mechanical ventilation, nephrotoxic cause of AKI, and presence of sepsis.
Subgroup analyses by quartiles of day 1 biomarker concentration revealed variable results (Supplemental Tables 4 and 5). Overall, no clear pattern emerged linking RRT intensity with outcomes. Finally, these results were confirmed in exploratory tree analyses (data not shown) and after correction for multiple comparisons (Supplemental Table 6) in which we were unable to find any association between RRT intensity and day 1 biomarker concentration with respect to renal recovery or mortality.
Association between Day 8 Biomarker Concentration and Renal Recovery and Mortality
Table 4 shows adjusted ORs for associations between day 8 biomarker concentrations and outcomes. Of the 568 participants with day 8 biomarker data, two participants had missing values for DR-5, and 90 participants had unknown premorbid creatinine. Of the remaining 476 participants with complete data, when adjusted for differences in age, sex, race, Charlson comorbidity score, premorbid serum creatinine, chronic hypoxia, presence of liver disease, immunosuppressed state, nephrotoxin exposure, diagnosis of sepsis, presence of oliguria, use of mechanical ventilation, intensity of RRT, and APACHE-II score, each natural log unit increase in biomarker concentrations was associated with lower odds of renal recovery (adjusted OR range for all markers, 0.19–0.87) and higher mortality (adjusted OR range for all markers, 1.26–3.18).
Table 4.
Association between individual day 8 biomarker concentration and clinical outcomes (n=476)
| Biomarker | Renal Recovery | Mortality |
|---|---|---|
| Inflammation | ||
| IL-6 | 0.73 (0.62–0.87) | 1.46 (1.23–1.73) |
| IL-8 | 0.61 (0.51–0.74) | 1.88 (1.54–2.28) |
| IL-10 | 0.73 (0.60–0.89) | 1.46 (1.20–1.77) |
| IL-18 | 0.73 (0.61–0.88) | 1.46 (1.20–1.76) |
| MIF | 0.87 (0.75–1.01) | 1.26 (1.08–1.47) |
| Apoptosis | ||
| TNFR-I | 0.19 (0.12–0.30) | 3.18 (2.06–4.89) |
| TNFR-II | 0.50 (0.35–0.73) | 1.65 (1.12–2.42) |
| DR-5 | 0.71 (0.55–0.91) | 1.48 (1.16–1.91) |
Values are adjusted odds ratios (95% confidence intervals). Odds ratios for individual day 8 biomarkers are adjusted for differences in baseline covariates, including age, race, sex, Charlson comorbidity score, chronic hypoxia, liver disease, immunocompromised state, APACHE-II score, premorbid serum creatinine, oliguria, mechanical ventilation, nephrotoxic cause of AKI, presence of sepsis, and intensity of RRT (but not adjusted for day 1 biomarker results). For the recovery model, an odds ratio >1 indicates that natural log increase in day 8 biomarker concentrations is associated with higher renal recovery, and an odds ratio <1 indicates nonrecovery. For the mortality model, an odds ratio >1 indicates that natural log increase in day 8 biomarker concentrations is associated with higher mortality, and odds ratio <1 indicates lower mortality. The models included only 476 participants with complete data because of missing premorbid creatinine in 90 participants and DR-5 in two participants. MIF, macrophage migration inhibitory factor; TNFR-I, TNF receptor-I; TNFR-II, TNF receptor-II; DR-5, death receptor-5.
When adjusted for multiple comparisons, higher concentrations of IL-6, IL-8, IL-18, IL-10, TNFR-I, and TNFR-II were associated with lower renal recovery, and concentrations of IL-6, IL-8, IL-10, IL-18, MIF, TNFR-I, and DR-5 were associated with mortality (Supplemental Table 7). There was no difference in day 8 biomarker concentrations between 90 participants with missing premorbid creatinine data, two participants with missing DR-5 data, and 476 participants with complete covariate data (data not shown). The number of patients who died and did not contribute to day 8 samples was similar between intensive and less-intensive RRT groups (26.1% versus 21.2%, respectively; P=0.09).
Discussion
Among critically ill individuals requiring RRT, we found that use of intensive RRT was not associated with overall lower plasma concentrations of inflammatory and apoptosis biomarkers on day 8 after initiation of RRT. However, there was an interaction between baseline mediator concentrations and intensive RRT where intensive RRT was associated with lower mediator concentrations when baseline levels were high (Figure 2). Although this interaction only reaches statistical significance for IL-6, MIF, and TNFR-I, it always favors intensive RRT. However, despite this interaction, use of intensive RRT was not associated with renal recovery or mortality given baseline marker profiles, not individually or when patterns were sought using a tree analysis. In contrast, elevated concentrations of biomarkers on day 8 were associated with lower renal recovery and higher mortality.
Taken together, these results indicate that although there is evidence that intensive RRT does influence mediator concentrations resulting in a reduction when baseline levels are high (Figure 2), this effect is limited with no overall difference (Figure 1). One possible explanation is that the dosing of intensive RRT might have been insufficient to modulate biomarker concentrations. Studies that have reported lower cytokine concentrations in critically ill patients have used higher dialysis dose or hemofiltration volume than that used in our study (12). Second, despite this interaction, no patient subgroup appeared to benefit from intensive RRT. Finally, day 8 mediator concentrations remain associated with outcomes such that higher levels portent worse prognosis.
In our study intensive RRT was associated with lower concentration of markers when circulating concentrations were high. Although we did not specifically examine clearance of cytokines in effluent fluids, our findings are in agreement with other studies (13). We speculate several mechanisms that could have contributed to this finding. First, higher ultrafiltration rates and more frequent dialysis in the intensive RRT group could have contributed to cytokine clearance. Second, removal of unknown mediators (biologically active substances, such as prostaglandins, leukotrienes, chemokines, or other cytokines) might have led to lower systemic inflammation. Third, our finding is in agreement with the peak concentration hypothesis that suggests the higher the circulating concentration, the better the clearance of cytokines with intensive RRT dosing (4).
Another unique finding of our study is that among participants with low cytokine concentrations, use of intensive RRT was associated with an increase in proinflammatory (IL-6 and MIF) and apoptosis (TNFR-I) markers. Excessive removal of solutes and proteins in minimally inflamed patients could have resulted in hypotension and increased immune response. Removal of small- and medium-size mediators, such as antimicrobial agents, micronutrients, and phosphate, could have potentially exacerbated inflammatory response, attenuating any potential benefit of intensive RRT.
Our finding of differential effect of intensive RRT on marker concentration might have influenced study outcomes. For instance, a potential beneficial effect of lowering cytokines might have been mitigated by the increased cytokine response associated with intensive RRT among patients who presented with lower cytokine levels. These findings suggest that delivery of a standard dose of intensive RRT for all patients is unlikely to be beneficial. Whether a dosing strategy on the basis of risk stratification using serial biomarker concentrations over the course of RRT treatment improves outcomes might warrant further investigation.
A proinflammatory milieu in our study was associated with nonrecovery of renal function and death. Our findings are similar to cytokine signatures that are associated with poor survival after sepsis (14,15) and cancer (16). We also found that elevated TNFR-I on day 8 was strongly associated with mortality, and lower concentrations were associated with renal recovery. Our data support other studies that found that exposure of renal tubular epithelium to TNFR-II increases apoptosis (17), and high plasma concentrations of TNFR-I and TNFR-II predict development of ESRD (18). Although, our data do not prove causality, they suggest that interventions that modulate persistent inflammation needs to be examined further.
Our study has important limitations. First, although the ATN study was interventional, our ancillary study was observational. Patients were randomized to intensive or less-intensive therapy but not on the basis of their inflammatory phenotype. Therefore, although we can attribute a causal relationship of RRT intensity on meditators and outcomes, the link between mediators and outcomes is purely observational and therefore causality cannot be inferred.
Second, we did not examine biomarker concentrations immediately before or after initiation of RRT sessions because our goal is to examine overall effect of intensity on marker concentration rather than efficacy of individual RRT sessions. Third, we did not examine local (i.e., renal tissue or ultrafiltrate) concentrations of these markers, which could have been different from plasma levels, and we are unable to ascertain the precise mechanisms by which intensive RRT, when baseline levels were high, is associated with lower marker concentrations.
Fourth, we did not account for membrane-flux characteristics because of variation in types of membranes used for intermittent hemodialysis sessions in the ATN trial. Fifth, day 8 biomarkers were sampled only in 568 patients who were alive (informative censoring). Finally, although this is the largest study to date, we remain underpowered to detect small differences that may have existed in clinical outcomes or biomarkers between RRT groups. On the basis of day 1 IL-6 concentrations we had sufficient power to detect 2-fold change; however, to detect very small differences, such as 25% change, we would need a sample size of approximately 3000 patients.
Our study has several major strengths. First, being a large multicenter prospective cohort study, our findings are highly generalizable. Second, we measured biomarkers in two domains that have been implicated in renal injury and recovery and examined the association between intensity of RRT and long-term outcomes. Third, we were able to detect the differential effect of intensity on outcomes given various marker profiles. Fourth, we measured biomarker concentration at two different time points and were able to examine the temporal relationship between intensive RRT and marker concentrations.
In summary, our results show that use of intensive RRT is not associated with changes in overall plasma biomarker concentrations in critically ill patients or in improvements in renal recovery or mortality. Nevertheless, there is an interaction between RRT intensity and plasma biomarker concentration depending on baseline concentration. Elevated concentrations of plasma inflammatory and apoptosis markers on day 8 were associated with RRT dependence and death.
Disclosures
JAK discloses grant support and/or consulting fees from Fresenius, Gambro, Baxter, Astute Medical, Alere, AM Pharma, Spectral, Grifols, Cytosorbents, Alung, Atox Bio, Bard, and Kaneka. None of the remaining authors have any financial conflicts to disclose.
Supplementary Material
Acknowledgments
Because Dr. Palevsky is a Deputy Editor of CJASN, he was not involved in the peer-review process for this manuscript. Another editor oversaw the peer-review and decision-making process for this manuscript.
We thank the nurses, respiratory therapists, phlebotomists, physicians, and other health care professionals who participated in the ATN and BioMaRK studies. We also thank the patients and their families.
The BioMaRK study was conducted by the BioMaRK investigators and supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (grant no. R01-DK070910) and the National Center For Advancing Translational Sciences of the National Institutes of Health (award no. KL2-TR000146). The Veterans Affairs/National Institutes of Health ATN study was supported by the Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development (CSP no. 530) and by NIDDK by an interagency agreement (no. Y1-DK-3508).
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIDDK or National Institutes of Health. This article does not necessarily reflect the opinions or views of the ATN study, NIDDK Central Repositories, or the NIDDK.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.04560514/-/DCSupplemental.
References
- 1.Palevsky PM, Zhang JH, O’Connor TZ, Chertow GM, Crowley ST, Choudhury D, Finkel K, Kellum JA, Paganini E, Schein RM, Smith MW, Swanson KM, Thompson BT, Vijayan A, Watnick S, Star RA, Peduzzi P, VA/NIH Acute Renal Failure Trial Network : Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med 359: 7–20, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bellomo R, Cass A, Cole L, Finfer S, Gallagher M, Lo S, McArthur C, McGuinness S, Myburgh J, Norton R, Scheinkestel C, Su S, RENAL Replacement Therapy Study Investigators : Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med 361: 1627–1638, 2009 [DOI] [PubMed] [Google Scholar]
- 3.Murugan R, Wen X, Shah N, Lee M, Kong L, Pike F, Keener C, Unruh M, Finkel K, Vijayan A, Palevsky PM, Paganini E, Carter M, Elder M, Kellum JA, Biological Markers for Recovery of Kidney (BioMaRK) Study Investigators : Plasma inflammatory and apoptosis markers are associated with dialysis dependence and death among critically ill patients receiving renal replacement therapy. Nephrol Dial Transplant 29: 1854–1864, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ronco C, Recent evolution of renal replacement therapy in the critically ill patient. Crit Care 10:123, 2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of different doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: A prospective randomised trial. Lancet 356: 26–30, 2000 [DOI] [PubMed] [Google Scholar]
- 6.Demirjian S, Chertow GM, Zhang JH, O’Connor TZ, Vitale J, Paganini EP, Palevsky PM, VA/NIH Acute Renal Failure Trial Network : Model to predict mortality in critically ill adults with acute kidney injury. Clin J Am Soc Nephrol 6: 2114–2120, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Charlson ME, Pompei P, Ales KL, MacKenzie CR: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40: 373–383, 1987 [DOI] [PubMed] [Google Scholar]
- 8.Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med 13: 818–829, 1985 [PubMed] [Google Scholar]
- 9.Paganini EP, Halstenberg WK, Goormastic M: Risk modeling in acute renal failure requiring dialysis: The introduction of a new model. Clin Nephrol 46: 206–211, 1996 [PubMed] [Google Scholar]
- 10.Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG: The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 22: 707–710, 1996 [DOI] [PubMed] [Google Scholar]
- 11.Su X, Tsai CL, Wang H, Nickerson DM, Li B: Subgroup analysis via recursive partitioning. J Mach Learn Res 10: 141–158, 2009 [Google Scholar]
- 12.Wang CT, Ren HS, Jiang JJ, Zhang JC, Meng M, Yu JB, Chu YF, Ding M: [Study the effects of high-volume hemofiltration and fluid resuscitation on removing blood lactic acid and pro-inflammatory cytokines in patients with refractory septic shock] [Chinese]. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 21: 421–424, 2009 [PubMed] [Google Scholar]
- 13.Morgera S, Slowinski T, Melzer C, Sobottke V, Vargas-Hein O, Volk T, Zuckermann-Becker H, Wegner B, Müller JM, Baumann G, Kox WJ, Bellomo R, Neumayer HH: Renal replacement therapy with high-cutoff hemofilters: Impact of convection and diffusion on cytokine clearances and protein status. Am J Kidney Dis 43: 444–453, 2004 [DOI] [PubMed] [Google Scholar]
- 14.Kellum JA, Kong L, Fink MP, Weissfeld LA, Yealy DM, Pinsky MR, Fine J, Krichevsky A, Delude RL, Angus DC, GenIMS Investigators : Understanding the inflammatory cytokine response in pneumonia and sepsis: results of the Genetic and Inflammatory Markers of Sepsis (GenIMS) Study. Arch Intern Med 167: 1655–1663, 2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yende S, D’Angelo G, Kellum JA, Weissfeld L, Fine J, Welch RD, Kong L, Carter M, Angus DC, GenIMS Investigators : Inflammatory markers at hospital discharge predict subsequent mortality after pneumonia and sepsis. Am J Respir Crit Care Med 177: 1242–1247, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Casasnovas R-O, Mounier N, Brice P, Divine M, Morschhauser F, Gabarre J, Blay JY, Voillat L, Lederlin P, Stamatoullas A, Bienvenu J, Guiguet M, Intrator L, Grandjean M, Brière J, Ferme C, Salles G, Groupe d’Etude des Lymphomes de l’Adulte : Plasma cytokine and soluble receptor signature predicts outcome of patients with classical Hodgkin’s lymphoma: A study from the Groupe d’Etude des Lymphomes de l’Adulte. J Clin Oncol 25: 1732–1740, 2007 [DOI] [PubMed] [Google Scholar]
- 17.Al-Lamki RS, Wang J, Vandenabeele P, Bradley JA, Thiru S, Luo D, Min W, Pober JS, Bradley JR: TNFR1- and TNFR2-mediated signaling pathways in human kidney are cell type-specific and differentially contribute to renal injury. FASEB J 19: 1637–1645, 2005 [DOI] [PubMed] [Google Scholar]
- 18.Niewczas MA, Gohda T, Skupien J, Smiles AM, Walker WH, Rosetti F, Cullere X, Eckfeldt JH, Doria A, Mayadas TN, Warram JH, Krolewski AS: Circulating TNF receptors 1 and 2 predict ESRD in type 2 diabetes. J Am Soc Nephrol 23: 507–515, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.