Increase in Creatinine and Cardiovascular Risk in Patients with Systolic Dysfunction after Myocardial Infarction

Materials and Methods

Results

Patients who were excluded from this analysis because they were missing serum creatinine measurements (n = 356), had cardiovascular events (n = 21), or were not taking medication at 2 wk (n = 41) were older, were more likely to be female, had more hypertension, and were more frequent users of diuretics compared with those who were included in the analysis. Although there was a higher cardiovascular event rate in excluded patients, there was no significant difference in baseline creatinine, change in creatinine, or treatment assignment compared with the study cohort.

Of the 1854 patients who composed the study cohort, the change in creatinine from baseline to 2 wk was normally distributed, ranging from −1.2 to 2.8 mg/dl with a mean change in creatinine by treatment assignment of 0.05 ± 0.3 in both the placebo and captopril groups (P = 0.9). A total of 223 (12.0%) patient had WRF with no significant difference in time from onset of MI to enrollment into SAVE between patients with and without WRF (average 11 d after MI). Patients with WRF were older; were more likely to be female; and had a higher prevalence of diabetes, smoking, and use of diuretics but had fewer previous MI (Table 1). In the multivariate logistic regression model, only age (odds ratio [OR] 1.03; 95% CI 1.01 to 1.04) remained a significant predictor of WRF in this population of post-MI patients with systolic dysfunction.

Worsening renal function as early as 2 wk after acute MI was associated with a higher incidence of death and was an independent predictor of death (hazard ratio [HR] 1.46; 95% CI 1.05 to 2.02), cardiovascular death (HR 1.62; 95% CI 1.14 to 2.30), and the composite end point (HR 1.32; 95% CI 1.03 to 1.70; Table 2). The prognostic value of WRF not only remained significant after adjustment for covariates but also was a stronger predictor of cardiovascular outcomes than baseline creatinine (HR 1.31 [95% CI 1.01 to 1.70]; HR 1.21 [95% CI 0.91 to 1.61]; HR 1.15 [95% CI 0.95 to 1.39] for death, cardiovascular death, and composite end point, respectively). Although NS, there also was a higher incidence of recurrent MI and hospitalization for CHF in patients with WRF.

A total of 104 (5.7%) patient in the placebo group and 116 (6.4%) patients in the captopril group had WRF, with no difference in the rates of WRF between treatment groups (P = 0.38). When stratified by treatment, the risk for death in patients with WRF was higher in the placebo group (HR 1.63; 95% CI 1.05 to 2.52) than in patients who were taking captopril at 2 wk (HR 1.33; 95% CI 0.81 to 2.21). A similar attenuation of risk was observed for the other primary end points as well as hospitalization for CHF (Table 2, Figures 1 and 2). Tests for interaction, however, were NS (P = 0.49, 0.73, 0.49, and 0.37 for death, cardiovascular death, hospitalization for CHF, and the composite end point, respectively).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Event rate according to worsening renal function (WRF) and treatment group.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Kaplan-Meier curves for death stratified by WRF and treatment group.

Discussion

Although baseline renal dysfunction is considered a potent cardiovascular risk factor in post-MI patients, the predictive value of WRF has not yet been addressed in this population. Because ACE inhibition can precipitate acute renal failure in the presence of severe bilateral renal artery stenosis, it is common practice to withdraw ACE inhibitors when renal function deteriorates; renal artery stenosis is a common finding among patients who undergo coronary angiography (14–18) and among those with heart failure (19). However, this policy may deprive many patients of the potential benefits of these drugs. In this study of post-MI patients with systolic dysfunction, we observed that WRF, defined as a rise in creatinine of >0.3 mg/dl after 2 wk, was not uncommon (12.0%) and was associated with significantly increased risk for cardiovascular mortality and morbidity. The impact of WRF on outcomes persisted even after adjustment for known covariates and was a stronger predictor of events than baseline creatinine.

The majority of studies that have examined the prognostic value of reduced renal function have focused on serum creatinine values or estimated GFR at one point in time. The predictive value of WRF, however, has been less clear and has been determined primarily in patients who were hospitalized for heart failure, in which the incidence of WRF (27 and 28%) was consistently greater than in SAVE (12.0%) (9,10). More risk factors have proved predictive of WRF in the heart failure population (baseline creatinine, systolic BP, history of diabetes or CHF) than were seen in this population of post-MI patients with systolic dysfunction (age). Few trials have assessed the prognostic value of WRF in the setting of coronary artery disease. Shlipak et al. (20) found that WRF, defined as a change of creatinine of >0.3 mg/dl during 4 yr of follow-up, was not associated with outcome in postmenopausal women with stable coronary artery disease in the Heart and Estrogen/Progestin Replacement Study (HERS). No other study has examined the prognostic importance of WRF in the common and important setting of left ventricular dysfunction after MI.

Most clinical studies have defined WRF as an increase in creatinine of >0.3 mg/dl, a threshold that has been found in previous studies to have the highest sensitivity (81%) and specificity (62%) for predicting mortality (13). Initial analyses in our cohort (data not shown) demonstrated a similar threshold effect, with higher event rates occurring in patients with a change in creatinine of >0.3 mg/dl. Lower magnitude changes may be less clinically relevant and represent transient changes or inherent variability in creatinine measurement. However, a more linear or J-shaped association between small changes in creatinine and risk for adverse cardiovascular outcomes also has been suggested (21,22).

The decision to withdraw ACE inhibition in patients who experience elevations in serum creatinine has been controversial (23). Physicians are reluctant to continue treatment with ACE inhibitors when creatinine begins to rise because of concerns about precipitating acute renal failure (24). In proteinuric kidney disease, ACE inhibitors reduce glomerular hyperfiltration by causing efferent glomerular arteriolar vasodilation, and short-term WRF is associated with long-term stability of kidney function (25–28). However, WRF in the setting of coronary artery disease and impaired systolic function probably is more likely to be due to reduced glomerular perfusion rather than glomerular hyperfiltration in the setting of ACE inhibition and therefore may carry a different prognosis. Recent data have demonstrated the benefits of ACE inhibition in patients with advanced renal insufficiency, even when a patient’s serum creatinine level continues to increase (29,30). Rapid and more extensive increases would suggest conditions such as bilateral renal artery stenosis or severe hypoperfusion, which warrant discontinuation of ACE inhibition. Butler et al. (31) did not find an association between WRF and ACE inhibitors in heart failure patients. Similarly, we did not observe a greater incidence of WRF in patients who received ACE inhibition in SAVE. Our data may suggest that the relationship between WRF and increased cardiovascular risk could be altered by ACE inhibitor therapy in patients after MI. Although a formal test for interaction was NS, most likely because of inadequate sample size and statistical power given the relatively small number of patients with WRF for each end point, the increased risk associated with WRF seemed consistently lower in the captopril group, raising the possibility that ACE inhibition may alter the relationship between elevation in creatinine and cardiovascular outcomes. This study has allowed for analysis of the effects of ACE inhibition on WRF and cardiovascular risk in a randomized, placebo-controlled setting. These hypothesis-generating results would argue against discontinuation of ACE inhibitor therapy after small increases in creatinine of <0.3 mg/dl.

A number of limitations of this analysis should be noted. Serum creatinine was measured 2 wk after randomization, which occurred on average 11 d after MI. The variation in time may have affected the results through either worsening or improvement of renal function. However, this variation in time not only underscores the prognostic value of WRF but also stresses the importance of repeating serum creatinine measurements from 1 to 3 wk after the acute event. The dosage of study medication (placebo or captopril) also was variable; although the study protocol did not specify that captopril should be stopped in the presence of WRF, titration regimens were left to the discretion of each patient’s physician.

Patients who tolerated the initial test dosage were included regardless of whether the study medication was not titrated fully to the target dosage of 25 mg three times daily by the end of 2 wk. The majority of patients were started at 12.5 mg and titrated upward, but there were patients who started at lower dosages or were down-titrated as a result of adverse drug reactions or cardiovascular events before 2 wk, which might have influenced the predictive value of increased creatinine. We would have expected to see an average decline in GFR for patients who started captopril but did not find this in SAVE. This finding most likely was due to variation in dose titration and patients’ not receiving the maximum therapeutic effect of captopril. The average change in GFR also may have been influenced by the 41 patients who dropped out of the trial as a result of adverse drug reactions from captopril. Our findings in the placebo group, however, would not have been influenced by dose titration. The results from this post hoc analysis were observed in post-MI patients with systolic dysfunction and cannot be extended to those with normal ventricular function. Finally, SAVE was conducted between 1988 and 1991. We cannot exclude the possibility that changes in the treatment of post-MI patients may have altered the current relation between WRF and cardiovascular risk. However, SAVE was conducted in a randomized, placebo-controlled setting and therefore allowed us to assess the true effect of ACE inhibition on cardiovascular outcomes in the setting of renal dysfunction using prospectively collected data.

Conclusion

We have shown that in patients with acute MI and systolic dysfunction, WRF defined as an increase in creatinine of >0.3 mg/dl within the first 2 wk is not uncommon (12.0%) and when present is associated with a significant increase in risk for cardiovascular outcomes and mortality. In this population of patients with systolic dysfunction after MI, WRF was a more robust predictor of events and death than baseline serum creatinine. The risk that is associated with WRF was most prominent in patients who received placebo and seems to be attenuated in patients who receive captopril. These findings suggest that close monitoring of renal function during the first few weeks after acute MI may aid in long-term risk stratification for cardiovascular events and may argue against discontinuation of ACE inhibitor therapy after small, nonprogressive increases in creatinine.