Primary angioplasty vs. thrombolysis: the end of the controversy?

The scientific ‘battle’ (for the optimal reperfusion therapy in acute myocardial infarction) between pharmaco-oriented and balloon-oriented cardiologists has already lasted 16 years. Tarantini et al.,[1] in a meta-analysis, have suggested a sophisticated mathematical equation to calculate the acceptable primary percutaneous coronary intervention (p-PCI)-related delay [(compared with thrombolysis (TL)] for individual patients presenting with acute ST-segment elevation myocardial infarction (STEMI). This newly proposed equation is based on the three main independent predictors of 30-day survival: (i) baseline mortality risk; (ii) presentation delay risk; and (iii) p-PCI (vs. TL) delay.

Tarantini's calculations showed that when the baseline mortality risk is 4%, the acceptable p-PCI (vs. TL) delay is only 35 min. Patients with higher baseline risk do benefit from a p-PCI strategy even when the transport times are extremely long: with 10% baseline risk the acceptable p-PCI delay is 153 min; with a baseline risk of 18% the acceptable p-PCI delay is (theoretically calculated) even >5 h (!).

This result supports the widespread use of p-PCI and almost total abandonment of TL. For >80% of the European population, a cath-lab with a PCI programme is available within 30 min transport distance. It is ‘just’ a matter of appropriate organization of the regional networks: >80% of STEMI patients can be treated by p-PCI—as has been clearly demonstrated in several European countries.[2] This study showed that over half of the European countries are already able to provide p-PCI services to the vast majority of STEMI patients. TL should be used only in sparsely populated regions with extremely long transfer distances (e.g. remote Norwegian fjords and mountains, Greek islands, Alaska, etc.). However, even in these remote regions, TL should be used ‘en route’ to a p-PCI centre, as was recently shown by the NORDISTEMI study.[3]

Tarantini et al.[1] concluded that ‘patients with a mortality risk <4.5% are unlikely to obtain a survival benefit by p-PCI compared with TL’. This is, however, an oversimplification. It was repeatedly shown that even patients with low mortality risk (e.g. young patients with inferior STEMI) have better outcomes with p-PCI compared with TL. The apparent lack of benefit (from p-PCI) is caused only by the lack of statistical power in low risk subgroups. Concerning the baseline risk, there is no single subgroup of STEMI patients showing benefit from TL compared with p-PCI. Patients in all baseline risk subgroups do benefit from p-PCI; the statistical power of this benefit is weak in low risk subgroups due to mathematical (not medical) reasons. While the absolute mortality difference (p-PCI vs. TL) decreases with decreasing baseline risk, the relative mortality benefit from p-PCI remains similar across all baseline risk subgroups.

Many discussions about p-PCI and TL omit four important problems: Age and Killip class are two main baseline risk factors predicting the 30 day outcome in STEMI.[12] Thus, if we want to implement the suggestions of Tarantini practically, we end up with the following recommendation: all patients >65 years of age[13] and all patients presenting in Killip class >I should be treated by p-PCI. Patients <65 years presenting in Killip class I should also be treated by p-PCI unless the p-PCI-related delay is substantially longer than 35 min. For this subgroup (young patients without signs of acute heart failure), TL remains a very good option in remote regions with long transfer distances or in places with suboptimal organization of STEMI patient care.

The time mismatch. The time delay of p-PCI (compared with TL) is usually calculated as the difference between door-to-needle time (TL) and door-to-balloon time (p-PCI). This is wrong in principle, because it is comparison of time to the start of treatment (TL) vs. time to the effect (reperfusion) of a treatment (p-PCI). This is a comparison of apples vs. oranges. If the realistic time to reperfusion were to be analysed, one should add at least 30 min (probably 60 min would be even more appropriate) on top of the door-to-needle time in thrombolysed patients. Also, this 30 min delay is similar to the usual sheath-to-balloon time in a p-PCI setting. In other words, because we do not know the exact time of reperfusion in TL, a comparison of door-to-needle time (TL) vs. door-to-sheath time (p-PCI) would be a much fairer method to compare the reperfusion therapies.

Different effectiveness. TL is effective in 40–60% of patients, while p-PCI is effective in ∼90%. Thus, comparing TL vs. p-PCI is comparing a semi-effective therapy vs. a fully effective therapy. The mortality benefit derived from the p-PCI strategy is not related to the fact that the underlying stenosis is removed (Figure 1). The mortality benefit of p-PCI is caused by the simple fact that p-PCI is twice as effective as TL in opening the artery.

Figure 1

Left anterior descending coronary artery in a patient with STEMI. Upper left: proximal occlusion. Upper right: reperfusion after thrombolysis with significant residual stenosis. Lower: reperfusion after primary PCI without any residual stenosis.

Uncertain thrombosis onset. The presentation delay (frequently used to stratify patients between p-PCI and TL therapy[4,5]) is based on the patient's subjective medical history. It was nicely shown by Rittersma[6] that coronary thrombi are substantially older (from several hours up to a few weeks) than the subjective history indicates. Thus, the presentation delay is the weakest predictor of the ultimate outcome—as was well demonstrated in the study of Tarantini et al.[1]

Facilitated PCI was abandoned due to the results of several randomized trials.[7-10] However, the proportion of patients with really long transport times was very low in these trials. Two other trials[3,11] suggest that for remote regions with very long transport distances, the pharmacomechanic approach may be valid.

The widespread (everywhere) use of Tarantini's equation for individual patients would add unnecessary complexity on the pre-hospital STEMI care in most regions, where p-PCI is available and should be applied to all STEMI patients. The suggested calculation is an ideal solution for sparsely populated regions with long transfer distances or for regions with suboptimal patient care organization.

Conflict of interest: none declared.