Abstract
Background and objectives Proteinuric diabetic kidney disease frequently progresses to ESRD. Control of BP delays progression, but the optimal BP to improve outcomes remains unclear. The objective of this analysis was to evaluate the relationship between BP and renal outcomes in proteinuric diabetic kidney disease.
Design, setting, participants, & measurements BP data from all 1448 randomized participants in the Veterans Affairs Nephropathy in Diabetes Trial were included in a post hoc analysis. The associations of mean on–treatment BP with the primary end point (decline in eGFR, ESRD, or death), renal end point (decline in eGFR or ESRD), rate of eGFR decline, and mortality were measured.
Results The median (25th, 75th percentile) follow-up time was 2.2 (1.2, 3.0) years. There were 284 primary end points. In univariate analyses, both mean systolic and mean diastolic BPs were strongly associated (P<0.001) with the primary end point. After multivariate adjustment, the hazard of developing the primary end point became progressively higher as mean systolic BP rose from <120 to ≥150 mmHg (P=0.02), with a significantly higher hazard ratio for 140–149 versus 120–129 mmHg (1.51 [1.06, 2.15]; P=0.02). There was also a significant association of mean diastolic BP with the hazard of developing the primary end point (P<0.01), with a significantly higher hazard ratio when mean diastolic BP was 80–89 versus 70–79 mmHg (1.54 [1.05, 2.25]; P=0.03); there was also a strong trend when mean diastolic BP was <60 mmHg. Associations between BP and both renal end point and rate of eGFR decline were similar to those with the primary end point. No association of BP with mortality was observed, possibly because of the limited number of mortality events.
Conclusions In patients with proteinuric diabetic kidney disease, mean systolic BP ≥140 mmHg and mean diastolic BP ≥80 mmHg were associated with worse renal outcomes.
- diabetes mellitus
- proteinuria
- end-stage renal disease
- progression of chronic kidney disease
- blood pressure
- diabetic nephropathies
- follow-up studies
- glomerular filtration rate
- humans
- kidney failure, chronic
Introduction
Diabetes mellitus is the leading cause of ESRD in the United States (1). BP control is important in preventing adverse cardiovascular and kidney outcomes in patients with diabetes and hypertension (2). Most of the evidence that BP control slows renal progression in patients with diabetes and overt proteinuria (i.e., proteinuric diabetic kidney disease [DKD]) comes from small studies in patients with type I diabetes, in whom BP was poorly controlled by current standards (3–9). Subsequent larger studies in patients with type I (10) or II (11) diabetes found lesser effects of BP, probably because of improved BP control.
In a post hoc analysis of the Reduction of Endpoints in Non-Insulin Dependent Diabetes with the Angiotensin II Antagonist Losartan (RENAAL) Trial, proteinuria, increased serum creatinine, hypoalbuminemia, and anemia at baseline but not baseline BP were predictors of progression (12). In another analysis of the RENAAL Trial focusing on BP, last systolic BP (SBP) before the study end point of >140 mmHg predicted a higher risk for progressive nephropathy. Baseline SBP was also predictive of kidney outcomes, but only when it was >160 mmHg (13). In a post hoc analysis of the Irbesartan Diabetic Nephropathy Trial (IDNT), achieved SBP of 151–160 mmHg was associated with a significantly greater risk than an achieved SBP of 131–140 mmHg (14).
Previous guidelines suggesting that BP should be <130/80 mmHg in patients with diabetes and/or CKD were primarily on the basis of cardiovascular rather than kidney outcomes and only moderate quality evidence (15–17). Recently, the benefits of such tight BP control have been questioned (18,19). In the report from the Eighth Joint National Committee (JNC8) panel members published in 2014 (19), BP of <140/90 mmHg is recommended for both patients with diabetes and patients with CKD. There are few data with regard to target BP in proteinuric DKD, however, leaving physicians uncertain about how aggressively to lower BP in such patients. In an attempt to elucidate this issue, we analyzed data from the recently completed Veterans Affairs Nephropathy in Diabetes (VA NEPHRON-D) Study to determine the association of BP with renal end points in patients with proteinuric DKD.
Materials and Methods
We performed a post hoc analysis designed to determine the association of mean achieved BP with the primary end point (decline in the eGFR, ESRD, or death), renal end point (decline in eGFR or ESRD), rate of eGFR decline, and mortality among participants in the VA NEPHRON-D Study. The VA NEPHRON-D Study was a multicenter, double–blind, randomized study designed to test the efficacy of the combination of losartan with lisinopril compared with losartan alone in slowing the progression of proteinuric DKD. The primary study results have been published previously (20). In brief, 1448 veteran patients with type II diabetes, an eGFR of 30–89.9 ml/min per 1.73 m2 calculated using the four–variable Modification of Diet in Renal Disease (MDRD) Equation (21), and a urinary albumin-to-creatinine ratio >300 mg/g (or protein-to-creatinine ratio of >500 mg/g) were evaluated. In addition to losartan at 100 mg daily, patients received lisinopril or matching placebo, which was titrated up to a maximum of 40 mg daily. After patients reached a stable maintenance dose, they were evaluated every 3 months, with BP medications adjusted to target an SBP of 110–130 mmHg and a diastolic BP (DBP) of <80 mmHg. The primary end point was the first occurrence of a decline in the eGFR (an absolute decrease of ≥30 ml/min per 1.73 m2 if the eGFR was ≥60 ml/min per 1.73 m2 at randomization or a relative decrease of ≥50% if the eGFR was <60 ml/min per 1.73 m2), ESRD (defined by the initiation of maintenance dialysis or an eGFR<15 ml/min per 1.73 m2), or death. The secondary renal end point was the first occurrence of a decline in the eGFR (as defined above) or ESRD. Randomization occurred between July of 2008 and September of 2012; the study was stopped early for safety reasons in October of 2012. The median follow-up time was 2.2 years.
BP Measurements
Trough sitting BP measurements were taken every visit during the screening, titration, randomization, and treatment phases. BP measurements were taken every 3 months after randomization, with all measurements taken after at least a 5-minute period in the sitting position. Two consecutive BP measurements were taken in the same arm at each visit, with the mean value of the two readings reported for each visit.
Statistical Analyses
All 1448 randomized participants were included in the analysis. BP values were categorized for SBP (<120, 120–129, 130–139, 140–149, and ≥150 mmHg) and DBP (<60, 60–69, 70–79, 80–89, and >90 mmHg). Baseline BP was defined as the BP value at the randomization visit after titration of monotherapy. Time–dependent last BP was defined as the last BP value before a primary end point or at the last study visit. Mean BP is the average of all on–treatment BPs from the randomization visit to the last BP. BPs obtained after the primary end point was reached were not included in the analysis.
In the calculation of mean BP, we used BPs collected during regular visits to maintain consistency among all participants. Because BP readings at the time of a primary event or close to the time of a primary event may be significantly different from the average BP, we did not include BPs that occurred within 60 days of a primary end point in the calculation of mean BP. For a similar reason, we took the last BP obtained at a regular visit at least 60 days before a primary end point as the last BP.
Baseline characteristics of participants stratified by categorical BP are summarized as means and SDs for normally distributed continuous variables and percentages for categorical variables. BP measurements from randomization to primary end point or exit from the study were summarized by the status of the primary end point in terms of means and 95% confidence intervals (95% CIs) at each time point.
To assess the association of SBP and DBP with study end points, we summarized the incidence rates of primary end point, renal end point, and mortality by categorical BP. We applied a univariate Cox regression evaluating time to each end point to estimate the hazard of each categorical BP group compared with the reference group and calculate its 95% CI.
To explore SBP and DBP as independent risk factors, we applied a multivariate Cox regression to evaluate the association of SBP and DBP with each end point after adjustment for age, race, treatment, baseline eGFR, albuminuria, coronary heart disease, congestive heart failure, diuretic usage, and body mass index (BMI). Continuous BP or categorical BP as appropriate was used to determine the hazard ratios of SBP and DBP. The proportional hazard assumption was tested at the 0.05 level. To evaluate whether there was an age dependency in the association between achieved BP and renal outcomes, we performed an additional analysis stratified by baseline age group (<65 versus ≥65 years old).
eGFR was summarized by categorical BP in terms of means and 95% CIs at each time point. A mixed model with repeated measures was applied to evaluate the association of BP with the rate of eGFR decline using the same covariates as in the Cox model.
Data management and statistical analysis used SAS programming language, version 9.3. All statistical tests were two sided. A P value of <0.05 was considered statistically significant, with no adjustment for multiple comparisons.
Results
Demographics
The demographic characteristics by categorical mean SBP and DBP are given in Tables 1 and 2. Mean BP was on the basis of a mean of 7.7±4.3 readings per participant. There were no differences in age or sex among the different SBP categories. Mean SBP was higher in black compared with white patients. SBP was also higher in those with higher LDL cholesterol or albuminuria and lower in those with greater BMI or coronary artery disease. Mean DBP was higher in black patients and those with higher LDL cholesterol and lower in older patients and those with coronary artery disease, congestive heart failure, or retinopathy.
Baseline characteristics by mean systolic BP during the study period
Baseline characteristics by mean diastolic BP during the study period
There was a −3.2-mmHg change in SBP and −2.1-mmHg change in DBP from baseline overall. In the combination group, the values were −3.5 and −2.3 mmHg, and in the monotherapy group, the values were −2.9 and −1.8 mmHg, respectively. The mean SBPs and DBPs were both slightly but not significantly lower in the combination therapy versus the monotherapy group (1.1 and 0.8 mmHg, respectively; P=0.06 for each comparison).
Study End Points
Primary End Point.
Overall, there were 284 primary study end point events (132 in the combination therapy group and 152 in the monotherapy group). The BP among patients who did and did not reach the primary end point is depicted in Figure 1. There was a strong association between mean SBP and the primary end point (P<0.001), with the hazard ratio progressively increasing relative to the reference group (mean SBP of 120–129 mmHg) (Figure 2A). There was also a significant association between mean DBP and the primary end point (P<0.001); however, a significantly higher hazard ratio occurred only with mean DBP ≥80 mmHg (Figure 2B). There was no association of baseline BP or last BP with the incidence of the primary end point (Supplemental Figures 1 and 2).
BP among patients who did and did not reach the primary end point. The figures depict the (A) systolic BP (SBP) and (B) diastolic BP (DBP) at baseline (Pre), at randomization (Rand), and during the course of the study in patients who did and did not develop the primary end point. The error bars denote means and 95% confidence intervals.
Univariate hazard ratios of the primary end point, secondary end point, and mortality for categorical mean systolic BP (SBP) and diastolic BP (DBP). (A) Mean SBP: P<0.001 for equal hazard of primary end point among all SBP groups, P<0.001 for equal hazard of secondary end point among all SBP groups, and P=0.11 for equal hazard of mortality among all SBP groups. (B) Mean DBP: P<0.001 for equal hazard of primary end point among all DBP groups, P<0.001 for equal hazard of secondary end point among all DBP groups, and P=0.22 for equal hazard of mortality among all DBP groups. The error bars denote 95% confidence intervals of the hazard ratios.
Additionally, Table 3 shows the independent associations of mean SBP and DBP with the primary end point after adjustment for treatment group, age, race, baseline eGFR, albuminuria, coronary heart disease, congestive heart failure, diuretic usage, and BMI. When mean SBP was analyzed as a categorical variable, the hazard of developing the primary end point was higher among patients with higher BP (P=0.02). Relative to a reference value of 120–129 mmHg, the hazard ratios for the primary end point were significantly higher for the 140–149 and >150 mmHg categories. When analyzed as a continuous variable, for every 10-mmHg increase of mean SBP, there was a 22% (95% CI, 7% to 33%) increase in the hazard of developing a primary end point (P<0.001). There was also a significant association of mean DBP with the hazard of developing the primary end point (P<0.01), with a U-shaped relationship noted. Relative to a mean DBP of 70–79 mmHg, the hazard ratios for the primary end point were significantly higher for the 80–89 and ≥90 mmHg categories. There was also a trend when DBP was <60 mmHg. There was no significant age group association with the primary end point, and there was not a significant age group by BP group interaction (data not shown). We also found similar results when we restricted analysis to patients treated with losartan alone (data not shown). The results using a fully adjusted model are shown in Supplemental Table 1.
Results of multivariate Cox regression for the hazard of primary end point
Secondary (Renal) End Point.
Overall, there were 178 secondary (renal) end point events (77 in the combination therapy group and 101 in the monotherapy group). Similar to the primary end point, there was a strong association between mean SBP and the renal end point (P<0.001), with a progressively higher hazard ratio with higher BP (Figure 2A). There was also a significant association between mean DBP and the renal end point (P=0.001) (Figure 2B). There was a significant association of baseline DBP but no association of baseline SBP with the incidence of the secondary end point (Supplemental Figure 1). There was no association of last BP with the incidence of the secondary end point (Supplemental Figure 2).
Additionally, Table 4 shows the independent associations of SBP and DBP with the renal end point after multivariate adjustment. When analyzed as categorical variables, there was a greater hazard of developing the renal end point for both mean SBP (P=0.03) and DBP (P=0.02). Relative to a reference value for mean SBP of 120–129 mmHg, the hazard ratio for the primary end point was significantly higher than for the 140–149 mmHg category. When analyzed as a continuous variable, for every 10-mmHg increase of mean SBP, there was a 23% (95% CI, 4% to 34%) increase in the hazard of developing a renal end point (P=0.004). Similar to the primary end point, there was a U-shaped relationship between mean DBP and the renal end point, although there were no significant differences between the DBP groups. There was no significant age group association with the secondary end point, and there was not a significant age group by BP group interaction.
Results of multivariate Cox regression for the hazard of renal end point
Mortality
There was a total of 119 patient deaths during the study. There was no significant association between mean SBP or DBP and mortality (Figure 2), which was not affected by multivariate adjustment (Table 5). Also, there was not an association with baseline BP or last BP before the end point on mortality (Supplemental Figures 1 and 2). There was no significant age group association with mortality, and there was not a significant age group by BP group interaction.
Results of multivariate Cox regression for the hazard of mortality
eGFR
The observed relationship of eGFR over time by BP group is depicted in Figure 3. To determine the relationship of BP with renal progression, we evaluated the association of SBP and DBP with eGFR slope (Table 6). The interaction term time × mean of SBP reached statistical significance (P<0.001), showing that different SBP groups had different slopes of eGFR decline, with a progressively greater decline of eGFR at higher BP. The interaction term time × mean of DBP also reached statistical significance (P<0.01), showing that different DBP groups also had different slopes of eGFR decline. There was a U-shaped relationship between rate of eGFR decline and mean DBP, with the lowest rate of eGFR decline seen when DBP was in the 70- to 79-mmHg range.
Summary of eGFR at randomization (Rand) and during the course of the study for categorical mean systolic BP (SBP) and diastolic BP (DBP). (A) SBP. (B) DBP. The error bars denote the 95% confidence intervals of the means.
Association of BP with eGFR slope
Discussion
Most previous post hoc analyses examining the association of BP with renal outcomes in DKD have focused on measurements made at a single time point, such as BP at baseline, at randomization, or before the study end point (12,13). However, such single measurements may not reflect the BP load experienced by the patient over a prolonged period. This may especially be true in diabetes, which is associated with a large intraindividual BP variability (22). Mean BP calculated from multiple BP measurements over a long duration should represent the general BP status better than single measurements. In this analysis of BP data from the VA NEPHRON-D Trial, mean achieved BP (both SBP and DBP) was significantly associated with both the primary end point (decline in eGFR, ESRD, or death) and the renal end point (decline in eGFR or ESRD). In addition, there was a progressively greater rate of eGFR decline at higher SBP. Neither baseline BP nor last BP was significantly associated with the primary end point, although there was an association between baseline DBP and renal end point.
In the post hoc analysis of IDNT, Pohl et al. (14) analyzed the association of mean achieved BP with outcomes, which was also done in our study. However, in IDNT, BP was not as well controlled as in the VA NEPHRON-D Trial; baseline BP in IDNT was 159/87 mmHg, and only 30% of patients achieved an SBP<135 mmHg, whereas in the VA NEPHRON-D Trial, baseline BP was 137/73 mmHg, 39% achieved an SBP<130 mmHg, and 74% achieved an SBP<140 mmHg. In the IDNT analysis, achieved SBP of 151–160 mmHg was associated with a significantly greater risk than an achieved SBP of 131–140 mmHg. In our study, achieved SBP of 140–149 mmHg was associated with worse outcomes than achieved SBP of 120–129 mmHg. Mean DBP ≥80 mmHg was also associated with worse outcomes, and there was a strong trend toward worse outcomes with low (<60 mmHg) mean DBP. None of the BP measurements were significantly associated with mortality, possibly because of the limited number of mortality events.
Similar to the data for primary and renal end points, there was a monotonic relationship between mean SBP and eGFR slope, suggesting that the lower the SBP, the better the renal outcome. However, there was a U-shaped relationship between mean DBP and eGFR slope and a significantly greater rate of eGFR decline when mean DBP was <60 mmHg. From a clinical standpoint, achieving a low SBP (e.g., <120–130 mmHg) without lowering DBP to <60 mmHg may not be possible, especially in older patients with decreased arterial compliance.
Previous guidelines for BP control in both diabetes and CKD were to attempt to achieve a BP of <130/80 mmHg (15). However, the recent JNC8 report liberalized BP goals in such patients to <140/90 mmHg (19). This recommendation was primarily on basis of the findings of the Action to Control Cardiovascular Risk in Diabetes Blood Pressure Trial (23), the African American Study of Kidney Disease and Hypertension Study (24), the MDRD Study (25), and the Ramipril Efficacy in Nephropathy Study (26). None of these trials showed that treatment to a lower BP goal significantly lowered kidney or cardiovascular disease end points compared with a goal of <140/90 mmHg. In patients with proteinuria (>3 g/24 h), post hoc analyses from the MDRD Study indicated benefit from treatment to a lower BP goal (<130/80 mmHg) with respect to kidney outcomes. However, only about 5% of patients with the MDRD Study had diabetes.
Tight BP control may not improve renal outcome in patients with diabetes who are nonproteinuric. In the normotensive Appropriate Blood Pressure Control in Diabetes Trial, intensive BP control (128/75 mmHg) versus moderate BP control (137/81 mmHg) did not affect rate of renal function decline (27). In a retrospective observational study of patients with type II diabetes and CKD, BP was not a predictor of rate of decline of eGFR (28). Similarly, recent large analyses in predominantly nonproteinuric type II diabetes support a goal SBP of <140 mmHg (29,30).
What are the current guidelines in patients with proteinuric DKD? The 2012 Kidney Disease Improving Global Outcomes guidelines recommend a goal BP of <130/80 mmHg (31,32). However, these recommendations were on the basis of weak evidence. In the only major randomized clinical trial that specifically targeted different levels of BP in DKD, there was no difference in renal outcome between a mean arterial pressure of <92 mmHg versus a mean arterial pressure of 100–107 mmHg (10).
Our study does have limitations. The VA NEPHRON-D Trial was designed to compare two treatment regimens and not a randomized study focused on BP control. The findings of our post hoc analysis seem robust, in that the adjusted and nonadjusted analyses showed very similar results. Nevertheless, we cannot exclude residual confounding that could have affected the relationship between BP and outcomes. For instance, it is possible that poorer control of BP reflected poorer compliance or an increase in comorbid conditions, either of which could have independently affected outcomes. Also, we cannot completely exclude the possibility of reverse causality (i.e., worsening renal function could have resulted in higher BP). Most of the patients were men, and therefore, the results may or may not apply to women. Finally, we recognize that results of post hoc observational studies cannot be used to make treatment recommendations or provide definitive data about target BPs.
In conclusion, our analysis supports the importance of BP control in slowing the progression of proteinuric DKD. Mean SBP >140 mmHg and mean DBP ≥80 mmHg were both associated with worse renal outcomes.
Disclosures
None of the authors declare actual or potential conflict of interest relevant to this article. N.V.E. is a speaker for Merck Pharmaceuticals.
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.
This study was supported by the Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development.
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.02850315/-/DCSupplemental.
- Received March 11, 2015.
- Accepted September 8, 2015.
- Copyright © 2015 by the American Society of Nephrology
References
In this issue
Jump to section
More in this TOC Section
Original Articles
Cited By...
- 11. Microvascular Complications and Foot Care: Standards of Medical Care in Diabetes--2021
- 11. Microvascular Complications and Foot Care: Standards of Medical Care in Diabetes-2020
- 11. Microvascular Complications and Foot Care: Standards of Medical Care in Diabetes--2019
- 10. Microvascular Complications and Foot Care: Standards of Medical Care in Diabetes--2018
- Blood Pressure-Related Outcomes in a Diabetic Population