Discussion
In this registry study, slightly more than 50% of older patients with ACS did not receive PCI, even though this procedure was the dominant therapeutic approach to ACS in the area, and in spite of a substantially more frequent application, as compared to the first wave of the same registry study conducted a few years before. The extent of PCI underutilisation increased with participants' clinical complexity and overall risk status. On the other hand, the net long-term survival benefit from PCI application did not blunt, yet progressively increased, with greater background risk, as represented by the SC score, and controlling for conditions, such as severe anaemia, low glomerular filtration rate and low platelet count, which may represent contraindications to invasive procedures.
Several previous investigations in different health care systems reported a similar age-related and comorbidity-related decline in the delivery of efficacious therapies in patients with ACS. In nearly 12 000 ACS cases from 55 hospitals in Switzerland, elderly patients were less likely to receive aspirin, β-blockers, PCI, or thrombolysis, whereas primary PCI was performed in 75% of STEMI and 60% of NSTEMI patients aged ≤50 years, and only in 25% and 12% of those aged 80+ respectively. Restrictions to the application of PCI solely justified by an advanced age appear unsupported by current evidence-based guidelines. A meta-analysis of 6763 participants in 22 randomised clinical trials, usually enrolling patients with STEMI, suggested that age, in five strata from ≤50 to >80 years, did not limit per se the efficacy of PCI: compared to fibrinolysis, PCI reduced 30-day mortality, reinfarction and stroke rates in all strata above the age of 50 years, including the 1427 (21%) participants aged 70–80 years (in whom OR ranged from 0.36 to 0.55) and the 410 (6%) aged >80 years (with OR ranging from 0.53 to 0.72). An interventional approach has been explicitly recommended in older patients for the first time in 1999 and 2000 in the American College of Cardiology/American Heart Association guidelines for STEMI and NSTEMI and age is still not mentioned in current ACS guidelines as a contraindication to PCI.
However, the reported evidence is only partially satisfactory because of systematic exclusion of more complex patients from randomised clinical trials, where older participants are less, and definitively healthier, than in clinical practice. Nevertheless, well-conducted observational studies help fill the gap between randomised trials and clinical practice. In a sample of 1876 ACS patients younger and 683 older than 75 years, the ACACIA registry study from Australia reported that long-term survival advantage from PCI increased with aging, independent of other factors. The first wave of the AMI-Florence registry showed a prognostic advantage from PCI that increased with the severity of comorbidity. Compared to that study, the present one provides some novel findings that should be highlighted. First, cases of NSTEMI were not excluded and we specifically focused on the oldest (75+ years) participants, whereas the previous study included STEMI subjects aged 65+ years. Second, risk stratification was not limited to comorbidities, but included the SC, a tool whose validity has been proven in different populations and recognised in a recent systematic review. Finally, we adjusted our survival analyses for several conditions that represent possible contraindications for PCI, as well as for enzymatic estimates of infarct size, variables that had not been considered in the previous study. Thus, our findings give stronger support, compared to the first AMI-Florence wave and the ACACIA study, to more extensive application of PCI to older frail ACS patients.
This study provides further evidence that the SC, developed for patients admitted to hospital for a variety of medical reasons, is specifically valid also in ACS. One-year mortality, indeed, increased by 10% per each point increase in the SC score, adjusting for disease-specific variables. Thus, the SC confirms its validity as a simple, low-cost prognostic tool, which does not require direct patient examination, as it is purely based on pre-existing administrative data.
Lack of random allocation to PCI prevents drawing firm conclusions on causation and represents a limitation of the present study. Although we have tried to adjust our analyses for several covariates, residual confounding cannot be totally excluded. A more detailed data collection on timing and type (ie, single-vessel vs multivessel) of PCI application would have been useful to a deeper understanding of how PCI is applied in the real world. However, these details should have had no relevant impact on our main findings shown in figure 1. Finally, this study may not accurately reflect the situation in other countries and even in other Italian regions, because accessibility to PCI is particularly high in the Florence area; therefore, our prevalence of PCI application might represent a best-case scenario, at least in Italy.
When strengths and weaknesses of this study are honestly balanced, we believe that our results challenge the common attitude, seen in clinical practice, to a 'therapeutic nihilism' towards older patients with ACS. Because of increased in-hospital mortality and complication rates, frail, complex older patients represent a high-risk cohort, whose short-term and long-term outcomes after PCI do not rely purely on cardiologic variables. For these reasons, their treatment requires a combination of the highest possible technical skills and of balanced, expert clinical judgment. Whereas new randomised studies of adequate size and scope would be welcome to gather more robust evidence, our findings suggest that a priori discrimination of old, frail patients from the benefit of PCI appears unjustified, and might be responsible for a substantial share of avoidable ACS mortality.