Renal dysfunction in older age groups and mortality after acute myocardial infarction

Risk stratification is an important step in the management of patients with acute myocardial infarction (MI) because those at higher risk may benefit the most from more active and effective therapy. Whatever the stage and regardless of the underlying cause, renal failure is an added risk factor for patients with acute MI.^1,2 The higher mortality rate in these patients may be explained by the presence of comorbid conditions³ and underuse of effective treatments.^3,4 We evaluated the association between renal function and mortality rates among patients with acute MI, taking into consideration the patient’s level of risk on hospital admission and the acute treatment strategy used.

Patients and methods

Between October 2000 and March 2001, patients admitted with acute MI were included in a prospective registry. During their stay, their serum creatinine levels were taken at least 2 times. Renal function was assessed by glomerular filtration rate (GFR), which was determined using the abbreviated Modification of Diet and Renal Disease Study formula.⁵

Patients were divided into the following 3 groups based on their GFR, using the guidelines of the National Kidney Foundation for staging chronic kidney disease,⁶ with the modification of combining stages 1 and 2, and stages 4 and 5: GFR1 group, > 59 mL/min/1.73 m²; GFR2 group, > 29 to < 60 mL/min/1.73 m²; and GFR3 group, < 30 mL/min/1.73 m². We used the Thrombolysis in Myocardial Infarction (TIMI) risk index to evaluate patients’ risk levels.⁷

The quality of care index was calculated according to the use of effective and guidelines-recommended treatments, when indicated, including statins; angiotensin-converting enzyme (ACE) inhibitors; aspirin or clopidogrel (Plavix); beta blockers; coronary angiography or revascularization, or both, within 48 hours; glycoprotein (GP) IIb/IIIa inhibitors; and reperfusion therapy. The details of this index and its relation to survival have previously been published.⁸

Results

Of 754 patients admitted with a diagnosis of MI, 56% (n = 421) had non—ST-segment elevation MI (NSTEMI), and 44% (n = 333) had ST-segment elevation MI (STEMI), with a median GFR of 48 mL/min/1.73 m² (range, 35-65 mL/min/1.73 m²). The percentages of patients in the GFR1, GFR2, and GFR3 groups were 28.5%, 43.5%, and 28%, respectively. A total of 130 patients in the GFR3 group had a GFR < 15 mL/min/1.73 m², and 20 patients had known renal failure.

At baseline, as GFR decreased, age tended to increase, along with a history of peripheral artery disease, hypertension, and stroke. Smoking and hypercholesterolemia tended to decrease as GFR decreased (

). Similarly, with decreased GFR, the TIMI risk index increased: 34.5 ± 14 in the GFR1 group, 50.0 ± 16 in the GFR2 group, and 60.1 ± 20 in the GFR3 group.

Table 1

Guidelines-recommended treatments, including beta blockers, statins, GP IIb/IIIa inhibitors, and reperfusion treatments, were underused in patients with decreased GFR, even though there were no contraindications to treatment and the indications were clear (

). As a result, the quality of care index decreased significantly with reduced GFR: 6.81 ± 2.53 in the GFR1 group, 6.23 ± 2.75 in the GFR2 group, and 4.58 ± 2.73 in the GFR3 group.

Table 2

At 1 year, 11.5% of patients (n = 87) died. As the quartiles of the TIMI risk index increased, so did the mortality rate: 2/188 (1.1%), 9/167 (5.4%), 32/182 (17.6%), and 44/201 (21.9%), for the first, second, third, and fourth quartiles, respectively. As shown in

, the 1-year mortality rate also increased markedly for all 3 groups: GFR1 group, 5/215 (2.3%); GFR2 group, 31/328 (9.4%); and GFR3 group, 51/211 (24.2%).

Table 2

TIMI risk index, type of infarction (STEMI or NSTEMI), history of diabetes, quality of care index, and GFR groups were analyzed using a stepwise multivariable logistic regression. After analysis, 3 variables remained in the model: the TIMI risk index (odds ratio [OR] = 1.01 [1.02; 1.04] per unit increase), GFR (OR = 2.4 [1.1; 7.5] from GFR1 to GFR2 and OR = 4.2 [1.6; 13] from GFR1 to GFR3), and the quality of care index (OR = 0.77 [0.69; 0.86] per 10% increase).

As shown in

, the model demonstrated a satisfactory discriminant accuracy, with a c statistic of c = 0.807 and an adequate model calibration. The c statistic value declined markedly when the GFR was eliminated from the model.

Table 3

Discussion

These data confirm the relation between the degree of renal dysfunction and mortality after acute MI. Renal dysfunction was associated with a higher risk score on admission and underuse of guidelines-recommended treatment. After adjusting for the quality of care index and the TIMI risk index, the prognostic value of GFR on mortality remained significant.

Many studies have evaluated the value of risk stratification systems in predicting mortality. Although all have used information about age and hemodynamic conditions, only some have incorporated renal function into the model.¹⁰ Our study was the only one to include renal function, especially the GFR, in the risk estimate. We also used a validated risk score and accounted for the use of effective therapy.

A large segment of our participants had renal impairment, with only 29% having a GFR > 60 mL/min/1.73 m² and 17.2% having previously undiagnosed renal dysfunction, with a GFR < 15 mL/min/1.73 m². Our results, which showed that a decreased GFR correlated with such comorbid conditions as a history of peripheral artery disease, previous MI, hypertension, and older age, were in agreement with other studies.^1,2

Our participants with a decreased GFR had higher risk scores at baseline and did not receive as many effective treatments,^3,11 or statins,³ even though there were no contraindications to treatment. This was in concurrence with previous studies. Despite the fact that there were no contraindications to revascularization and coronary angiography, these procedures were not used as often in patients with renal dysfunction.^12,13 Renal function was shown on multivariable logistic regression analysis to be an independent predictor of 1-year mortality, despite potential confounding factors, such as higher TIMI risk and quality of care index.

Conclusion

Our results verify the frequency of renal dysfunction in patients admitted to the hospital with acute MI and the predictive value of renal dysfunction on mortality, even at a moderate stage. The higher risk score and underuse of guidelines-recommended treatments only partially explain the higher mortality in patients with impaired renal function. Based on the results of our study, renal function assessments should be incorporated into risk scoring systems, and guidelines-recommended treatments should be used in patients with renal dysfunction.