A total of 4363 patients were considered for inclusion in the study. Of these, 84 were excluded due to recurrent ischemia or cardiogenic shock. Of the 4279 patients finally analyzed, 69% had normal or slightly reduced renal function, 26% presented moderate renal dysfunction, and 5% had severe renal dysfunction. The observed in-hospital mortality rate was 1.8%, 8.4% and 9.6%, respectively (p<0.0001) (Fig. 1).

Figure 1.

In-hospital mortality and routine invasive strategy (RIS) utilization according to baseline risk.

(0.12MB).

In-hospital mortality could not be determined in 287 patients with normal renal function (9.8%) and in 130 patients with renal dysfunction (9.7%). The demographic characteristics, comorbidities and initial severity of the patients lost to follow-up did not differ appreciably from those of the patients that completed follow-up (see Table S3 of the supplementary material available in the online version).

The percentage of patients subjected to RIS in the subgroups of normal renal function, moderate renal dysfunction and severe renal dysfunction was 52%, 42% and 33%, respectively (p<0.0001) (Fig. 1).

The baseline characteristics of the patient treated on an invasive or conservative basis were very different (Table 1). The conservative management group was characterized by patients preferentially admitted to hospitals without possibilities for hemodynamic testing, individuals of older age, with poorer renal function, an increased prevalence of diabetes and known ischemic heart disease, a lesser prevalence of high risk ECG findings, and greater initial severity as established from the Killip, TIMI or GRACE scores.

Propensity score matching of the cohorts yielded 1144 matched pairs with normal renal function and 499 with renal dysfunction. The distribution of the propensity scores in the invasive and conservative management groups showed adequate overlap (see Figure S1 of the supplementary material available in the online version). Propensity score matching was largely able to homogenize both populations (see Table S2 of the supplementary material available in the online version), though standardized differences of over 10 persisted in relation to some variables of little prognostic impact in both the subgroup with renal dysfunction and in the subgroup with normal renal function.

Regarding the primary endpoint, the effect of RIS was similar in patients with moderate-severe dysfunction (adjusted OR 0.52; 95%CI 0.32–0.87) and normal renal function (adjusted OR 0.38; 95%CI 0.17–0.81) (interaction test p=0.4779). Likewise, no interaction was observed between RIS and the degree of renal function on stratifying into three groups (interaction test p=0.76) (see Figure S3 of the supplementary material available in the online version).

Given the different baseline risk, the risk difference was greater in the group of patients with renal dysfunction (RD −5.1%; 95%CI −8.1 to −2.1) than in the group with preserved renal function (RD −1.6%; 95%CI −2.6 to −0.6) (interaction test p=0.0335) (Tables 3 and 4; Fig. 2).

Figure 2.

Raw and adjusted in-hospital mortality differences.

(0.13MB).

Regarding the secondary endpoints, RIS was associated to a decrease in mortality in the ICU and after 30 days, with no significant reduction in the risk of infarction/reinfarction, acute renal failure or moderate-severe hemorrhage. The effect for these 5 endpoints was similar (interaction test p=nonsignificant) for the patients with and without renal dysfunction (Table 3).

The analysis of subgroups showed that within the stratum of patients with a GRACE score >140, RIS reduced in-hospital mortality both in patients with renal dysfunction and in those with normal renal function. In the patients with a GRACE score ≤140, RIS did not significantly reduce in-hospital mortality. The interaction test p-value was not statistically significant, however. The effectiveness of RIS in patients with normal renal function was shown to be greater in centers without a Hemodynamics Unit than in those where hemodynamic testing was available (interaction test p=0.0133). There was no evidence of interaction between RIS application and patient gender in either patients with normal renal function or in patients with altered renal function (see Table S5 of the supplementary material available in the online version).

Our results confirm the persistent scant use of RIS in patients with NSTE-ACS and renal dysfunction, and show that this strategy is associated to a relative reduction of in-hospital mortality and mortality after 30 days in these patients, being comparable to that of patients with normal renal function.5,6,8,15,32 In absolute terms, however, the short-term reduction in mortality among patients with renal failure is greater than in those with normal renal function: on treating the same number of patients, the number of deaths avoided by RIS is greater in patients with renal dysfunction than in individuals with normal renal function.

On the other hand, we detected no association between the application of RIS and the incidence of infarction, renal failure or moderate-severe hemorrhage in patients with renal dysfunction. The absence of a detectable effect upon the risk of infarction is consistent with the observations of most published studies, which report an initial increase in infarction risk that is only compensated by a decrease in reinfarction risk over the middle to long term. In turn, the negative results referred to the risk of hemorrhage and renal failure could be due to both a lack of statistical power of the study and to improved preventive practices in this field (e.g., generalization of the radial access, or the prevention of contrast-induced nephropathy).

With regard to the reasons underlying lesser utilization of RIS in patients with renal dysfunction, concerns about causing contrast-induced nephropathy and additional impairment of renal function could dissuade clinicians from applying RIS in such cases. However, the impact of iodine contrast administration upon important endpoints or outcomes (such as death or the need for dialysis) has not been clearly established,33 and the existing evidence points to a benefit-risk ratio favorable to the use of RIS in patients with renal failure.9,10

From a broader perspective, the infra utilization of RIS in these patients could be regarded as a peculiar manifestation of a tendency among clinicians to avoid using invasive procedures in high risk patients.11,12,14,17 In our concrete case, the utilization of RIS predominantly corresponded to the subgroup of patients with normal renal function and a GRACE score <140, which is precisely the group that stands to benefit least from such treatment. The possible influence of difficulties in moving the patient to the catheterization room, and the limitation of vital support measures in the more seriously ill patients, requires more in depth studies.

Our study has limitations, including particularly its observational design. Although propensity score matching resulted in a uniform distribution of the main covariables between the experimental group and the control group, we cannot rule out the existence of residual confounding effects attributable to certain variables not included in the registry. This residual confounding effect could explain the differences between the odds ratios estimated by observational studies and those estimated from randomized trials–with clearly more favorable values in the former. An alternative explanation could be the strong cross-over presence in the randomized studies (patients in the control group subjected to percutaneous coronary intervention), which would contribute to underestimate the benefit of RIS. In any case, rather than establishing a point estimation of the effect of RIS, our study sought to explore the possible heterogeneity of effects according to the degree of renal functional impairment.

Another limitation of our study refers to the losses to follow-up and the lack of middle-long term results. The credibility of our findings is strengthened by the absence of systematic differences between the patients lost to follow-up and those included in the study, as well as by the consistency of our results with those of other studies. Nevertheless, the possible existence of selection bias cannot be ruled out entirely.

Lastly, although we detected no subgroup effect in the patients with severe renal dysfunction, a total of 16 events were recorded among these subjects. Our study therefore has a problem of imprecision in estimating the effect in patients with severe renal dysfunction.34

In conclusion, the results obtained indicate that the effectiveness of RIS is similar in patients with normal or impaired renal function, and alert to infrautilization of this strategy among the latter type of patients.

SEMICYUC.

None.