https://orcid.org/0000-0002-8721-0515
Abstract
Dual antiplatelet therapy (DAPT) with a P2Y12 inhibitor on top of aspirin is the cornerstone of therapy after acute coronary syndromes (ACS). Nonetheless, the safest and most efficacious P2Y12 for older patients who are both at high ischaemic and bleeding risk remains uncertain. We aimed to examine the effect of available P2Y12 inhibitors on ischaemic and bleeding endpoints in older adults with ACS.
Randomized clinical trials that reported separately the results of adults older >70 years for at least the primary endpoint [composite of death, myocardial infarction (MI), and stroke]. Seven studies (14 485 patients-years) were included. Network meta-analysis showed that prasugrel was associated with similar occurrence of the primary endpoint and of a secondary ischaemic endpoint (composite of MI and stroke) and was most likely the best treatment [Surface Under the Cumulative Ranking curve Analysis (SUCRA) 54.5 and 59.8, respectively]. With regards to major bleedings, clopidogrel showed the highest likelihood of event reduction (SUCRA 70.1%), while ticagrelor of stent thrombosis (SUCRA 55.6%). Our meta-regression with a fixed proportion of patients managed invasively of 100% confirmed these trends with increasing SUCRA.
Among older subjects with ACS, DAPT should be balanced upon ischaemic and bleeding risks as prasugrel is associated with the highest probability of reduction of ischaemic events and clopidogrel of bleedings. Ticagrelor had highest SUCRA for stent thrombosis reduction but seems suboptimal in older adults.
Introduction
Treating older adults with acute coronary syndromes (ACS) poses a daily clinical challenge particularly concerning the choice of dual antiplatelet therapy (DAPT) as recommended by international guidelines.1,2 While the potent P2Y12 inhibitors prasugrel and ticagrelor, on top of aspirin, are superior to clopidogrel in terms of ischaemic events,3 they pose a higher hazard of bleedings. This is an important concern in older adults who often present with high bleeding and ischaemic burden,4,5 and a recent randomized clinical trial (RCT) dedicated to patients older than 70 years suggested that clopidogrel might be a better choice from an overall clinical perspective.6 The present systematic review and network meta-analysis with pooled data from all available RCTs aimed at defining the best P2Y12 inhibitor for older adults when ischaemic or bleeding endpoints are considered.
Methods
This article has been reported in accordance with the Preferred Reporting Items for System Reviews and Meta-Analysis (PRISMA-NMA).7 The study protocol is available in Supplementary material online, Methods S1 and was registered in PROSPERO (ID: 178406).8
Data sources and searches
A systematic search of MEDLINE, EMBASE, Google Scholar, and the Cochrane Central Register of Controlled Trials (CENTRAL) from database inception through April 2020 (Supplementary material online, Methods S2). The reference lists of included studies were searched for additional studies. Systematic reviews were identified and screened for additional trials (Figure 1).
Summary of evidence search and selection.
After removal of duplicates, the title and abstracts of the search results were screened for relevance by a single author (C.M. or A.M.). The full text of the remaining results was independently assessed in duplicates by two authors (C.M. and A.M.) for inclusion, based on predetermined criteria. Any disagreement was decided upon by a senior author (G.C.). The final list of included studies was decided upon discussion between authors with full agreement required prior to inclusion.
Study selection
Papers were considered eligible if they: (i) were an RCT; (ii) enrolled only older patients OR had a pre-specified subgroup of older subjects after ACS; (iii) compared ticagrelor or prasugrel, on top of aspirin, with each other or with the former standard of therapy (clopidogrel); (iv) had a follow-up of at least 1 year; (v) provided information on any of the pre-specified primary, secondary and safety endpoints; and (vi) were published in English language. Older adults were included if they were ≥70 years old.
Data extraction and quality assessment
Data were extracted using a dedicated electronic database, independently and in duplicate, by two authors (C.M. and A.M.). The data extracted from each paper included baseline participant characteristics, inclusion criteria, study drug and control treatments, follow-up duration and endpoint data. A ‘patient-years’ computation was used to address different follow-up periods across the studies.
Risk of bias assessment was conducted by two authors in duplicate (C.M. and A.M.) using the Cochrane Collaboration risk of bias tool (ROB2) across five domains: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. We used the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach to assess confidence in estimates of effect (quality of evidence) associated with specific comparisons.9 The Egger’s test was used to identify asymmetry of funnel plots for publication bias.
Data synthesis and analysis
The primary outcome was the composite of all-cause death, myocardial infarction (MI), and stroke. Secondary endpoints were the composite of MI and stroke and the occurrence of definite or probable stent thrombosis according to Academic Research Consortium.10 Safety endpoint was cumulative of major bleedings. Definitions of endpoints in individual studies are in Supplementary material online, Table S1.
A Bayesian hierarchical network meta-analysis (Figure 2) was performed using the BUGSnet package11 on R (version 3.6.1). Fixed and random-effect models were selected for each outcome based on the smallest deviance information criteria. When unavailable, a random-effect model was used (Supplementary material online, Figure S1 and Table S2). Analysis was performed using the Markov-chain Monte Carlo methods, based on 100 000 iterations with a burn-in of 10 000. Convergence was assessed with the Gelman–Rubin convergence diagnostic test12 (Supplementary material online, Figure S2). We used a random seed and vague priors. Results are presented as odds ratios (ORs) with 95% credible intervals (CrIns). Transitivity (similarity between sets of trials with respect to important effect modifiers) was assessed by constructing summary to qualitatively assess baseline clinical similarities of trial populations (Supplementary material online, Table S3).
Network plot for primary endpoint. The nodes represent P2Y12 inhibitors to be compared and the edges represent the observed direct comparisons in the included trials. The size of the nodes is proportional to the number of patients assigned to each drug and the thickness of edges is proportional to the sample size of each direct comparison. In blue, the number of studies for each direct comparison.
The probability that each treatment class ranked in a given position (from best to worst) was estimated and presented in ranking plots.13
Network consistency was analysed by analysis of tracepolts, leverage plots, and posterior mean deviance comparison plots (Supplementary material online, Figure S3).
Between trial heterogeneity was assessed with using the I2 statistics.
Number needed to treat/harm were calculated based on the maximum outcome estimated. number needed to treat for net effect (NNTnet) was calculated14 to weight the benefit of prasugrel and ticagrelor vs. clopidogrel. According to available evidence and consensus among authors, a weight = 2 was attributed to the primary endpoint and a weight = 1 to major bleedings.15
A frequentist random-effect network meta-analysis was also performed using the packages netmeta and dmetar on R. Results are presented as OR with 95% confidence intervals. Two-tailed P-values of 0.05 were used for statistical significance. The P-score statistic was used to assess the mean probability of superiority of each drug class to alternative treatments for a given outcome.16
Summary of direct (vs. indirect) evidence is in Supplementary material online, Figure S4.
Sensitivity analysis and meta-regression
A sensitivity analysis excluding studies that randomized patients to prasugrel 10 mg was performed. Moreover, a meta-regression was fit with the proportion of patients invasively managed in each study. Additional statistical methods are in Supplementary material online, Methods S3.
Results
Study search and study characteristics
Our systematic research identified 314 articles, of which seven were RCT dedicated to or with a pre-specified subgroup of older patients6,17–22 (Figure 1). In total, 14 485 patients-year were included in the network meta-analysis. Of them, 6697 (46.2%) patients received clopidogrel, 5289 (36.5%) prasugrel, and 2499 (17.3%) ticagrelor. For direct comparison, three studies compared prasugrel to clopidogrel, two ticagrelor to clopidogrel, and two prasugrel to ticagrelor. The baseline characteristics of studies were deemed sufficiently similar based on sex, age, diabetes mellitus, and clinical presentation to permit network comparison (Supplementary material online, Table S3).
Risk of bias and publication bias
Risk of bias assessment is shown in Supplementary material online, Figure S5 and was low for all studies except one, for all domains explored. Quality of evidence according to GRADE assessment is in Supplementary material online, Table S4. Funnel plots and Egger’s test results are in Supplementary material online, Figure S6 and no significant bias was observed.
Primary outcome: all-cause death, myocardial infarction, and stroke
For this endpoint, all studies reported at least one event in each group (Supplementary material online, Table S5). In total, the primary endpoint occurred in 2139 (14.7%) patients: 1006 (15.0%) of the 6697 treated with clopidogrel, 644 (12.2%) of the 5289 with prasugrel, and 489 (19.5%) of the 2499 with ticagrelor. In a standard meta-analytic approach, prasugrel (OR 0.96; 95% CrIn 0.85–1.08; high confidence) and ticagrelor (OR 0.98; 95% CrIn 0.86–1.13; high confidence) similar occurrence of primary endpoint compared to clopidogrel. Forest plots of relative risk are in Supplementary material online, Figure S7. Nonetheless, our Surface Under the Cumulative Ranking curve (SUCRA) Bayesian analysis showed that prasugrel is most likely the best treatment with regards to the primary endpoint of all-cause death, MI, and stroke (54.5%) when compared with ticagrelor and clopidogrel (SUCRA 32.9 and 12.6, respectively) (Table 1).
SUCRA values for each P2Y12 inhibitor and outcomes
| Main analysis | Meta-regression analysis (invasive management = 100%) | Sensitivity analysis (prasugrel 5 mg only) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SUCRA for primary endpoint | |||||||||||
| Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor |
| 1 | 12.6 | 54.5 | 32.9 | 1 | 11 | 61.3 | 27.7 | 1 | 21.1 | 42.5 | 36.4 |
| 2 | 41.6 | 28.2 | 30.1 | 2 | 33.6 | 27.2 | 39.2 | 2 | 42.5 | 27.3 | 30.2 |
| 3 | 45.7 | 17.3 | 37.0 | 3 | 55.4 | 11.5 | 33.1 | 3 | 36.4 | 30.2 | 33.4 |
| SUCRA for secondary ischaemic endpoint | |||||||||||
| 1 | 27.8 | 59.8 | 12.3 | 1 | 10.8 | 73.8 | 15.4 | 1 | 33.2 | 52.3 | 14.4 |
| 2 | 57.9 | 26.5 | 15.6 | 2 | 53.2 | 19.0 | 27.8 | 2 | 51.8 | 26.2 | 22 |
| 3 | 14.2 | 13.7 | 72.1 | 3 | 35.9 | 7.25 | 56.8 | 3 | 15 | 21.5 | 63.5 |
| SUCRA for secondary endpoint of stent thrombosis | |||||||||||
| 1 | 2.01 | 42.4 | 55.6 | 1 | 3.88 | 35.6 | 60.6 | 1 | 2.74 | 45 | 52.3 |
| 2 | 24.0 | 41.5 | 34.5 | 2 | 31.5 | 50.5 | 18 | 2 | 28.9 | 33.1 | 38 |
| 3 | 74.0 | 16.1 | 9.91 | 3 | 64.7 | 13.9 | 21.5 | 3 | 68.4 | 21.9 | 9.67 |
| SUCRA for safety endpoint of bleeding | |||||||||||
| 1 | 70.1 | 9.54 | 20.4 | 1 | 66.7 | 8.13 | 25.2 | 1 | 57 | 23.1 | 19.8 |
| 2 | 25.7 | 26.6 | 47.7 | 2 | 26.7 | 30.6 | 42.6 | 2 | 33.4 | 27.1 | 39.6 |
| 3 | 4.22 | 63.9 | 31.9 | 3 | 6.57 | 61.2 | 32.2 | 3 | 9.62 | 49.8 | 40.6 |
| Main analysis | Meta-regression analysis (invasive management = 100%) | Sensitivity analysis (prasugrel 5 mg only) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SUCRA for primary endpoint | |||||||||||
| Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor |
| 1 | 12.6 | 54.5 | 32.9 | 1 | 11 | 61.3 | 27.7 | 1 | 21.1 | 42.5 | 36.4 |
| 2 | 41.6 | 28.2 | 30.1 | 2 | 33.6 | 27.2 | 39.2 | 2 | 42.5 | 27.3 | 30.2 |
| 3 | 45.7 | 17.3 | 37.0 | 3 | 55.4 | 11.5 | 33.1 | 3 | 36.4 | 30.2 | 33.4 |
| SUCRA for secondary ischaemic endpoint | |||||||||||
| 1 | 27.8 | 59.8 | 12.3 | 1 | 10.8 | 73.8 | 15.4 | 1 | 33.2 | 52.3 | 14.4 |
| 2 | 57.9 | 26.5 | 15.6 | 2 | 53.2 | 19.0 | 27.8 | 2 | 51.8 | 26.2 | 22 |
| 3 | 14.2 | 13.7 | 72.1 | 3 | 35.9 | 7.25 | 56.8 | 3 | 15 | 21.5 | 63.5 |
| SUCRA for secondary endpoint of stent thrombosis | |||||||||||
| 1 | 2.01 | 42.4 | 55.6 | 1 | 3.88 | 35.6 | 60.6 | 1 | 2.74 | 45 | 52.3 |
| 2 | 24.0 | 41.5 | 34.5 | 2 | 31.5 | 50.5 | 18 | 2 | 28.9 | 33.1 | 38 |
| 3 | 74.0 | 16.1 | 9.91 | 3 | 64.7 | 13.9 | 21.5 | 3 | 68.4 | 21.9 | 9.67 |
| SUCRA for safety endpoint of bleeding | |||||||||||
| 1 | 70.1 | 9.54 | 20.4 | 1 | 66.7 | 8.13 | 25.2 | 1 | 57 | 23.1 | 19.8 |
| 2 | 25.7 | 26.6 | 47.7 | 2 | 26.7 | 30.6 | 42.6 | 2 | 33.4 | 27.1 | 39.6 |
| 3 | 4.22 | 63.9 | 31.9 | 3 | 6.57 | 61.2 | 32.2 | 3 | 9.62 | 49.8 | 40.6 |
SUCRA values are presented as percentage of area under the cumulative rank probability. The larger of the SUCRA value, the better the treatment regimen performance with respect to the outcome. The highest SUCRA values for each endpoint are highlighted in green.
SUCRA values for each P2Y12 inhibitor and outcomes
| Main analysis | Meta-regression analysis (invasive management = 100%) | Sensitivity analysis (prasugrel 5 mg only) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SUCRA for primary endpoint | |||||||||||
| Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor |
| 1 | 12.6 | 54.5 | 32.9 | 1 | 11 | 61.3 | 27.7 | 1 | 21.1 | 42.5 | 36.4 |
| 2 | 41.6 | 28.2 | 30.1 | 2 | 33.6 | 27.2 | 39.2 | 2 | 42.5 | 27.3 | 30.2 |
| 3 | 45.7 | 17.3 | 37.0 | 3 | 55.4 | 11.5 | 33.1 | 3 | 36.4 | 30.2 | 33.4 |
| SUCRA for secondary ischaemic endpoint | |||||||||||
| 1 | 27.8 | 59.8 | 12.3 | 1 | 10.8 | 73.8 | 15.4 | 1 | 33.2 | 52.3 | 14.4 |
| 2 | 57.9 | 26.5 | 15.6 | 2 | 53.2 | 19.0 | 27.8 | 2 | 51.8 | 26.2 | 22 |
| 3 | 14.2 | 13.7 | 72.1 | 3 | 35.9 | 7.25 | 56.8 | 3 | 15 | 21.5 | 63.5 |
| SUCRA for secondary endpoint of stent thrombosis | |||||||||||
| 1 | 2.01 | 42.4 | 55.6 | 1 | 3.88 | 35.6 | 60.6 | 1 | 2.74 | 45 | 52.3 |
| 2 | 24.0 | 41.5 | 34.5 | 2 | 31.5 | 50.5 | 18 | 2 | 28.9 | 33.1 | 38 |
| 3 | 74.0 | 16.1 | 9.91 | 3 | 64.7 | 13.9 | 21.5 | 3 | 68.4 | 21.9 | 9.67 |
| SUCRA for safety endpoint of bleeding | |||||||||||
| 1 | 70.1 | 9.54 | 20.4 | 1 | 66.7 | 8.13 | 25.2 | 1 | 57 | 23.1 | 19.8 |
| 2 | 25.7 | 26.6 | 47.7 | 2 | 26.7 | 30.6 | 42.6 | 2 | 33.4 | 27.1 | 39.6 |
| 3 | 4.22 | 63.9 | 31.9 | 3 | 6.57 | 61.2 | 32.2 | 3 | 9.62 | 49.8 | 40.6 |
| Main analysis | Meta-regression analysis (invasive management = 100%) | Sensitivity analysis (prasugrel 5 mg only) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SUCRA for primary endpoint | |||||||||||
| Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor | Rank | Clopidogrel | Prasugrel | Ticagrelor |
| 1 | 12.6 | 54.5 | 32.9 | 1 | 11 | 61.3 | 27.7 | 1 | 21.1 | 42.5 | 36.4 |
| 2 | 41.6 | 28.2 | 30.1 | 2 | 33.6 | 27.2 | 39.2 | 2 | 42.5 | 27.3 | 30.2 |
| 3 | 45.7 | 17.3 | 37.0 | 3 | 55.4 | 11.5 | 33.1 | 3 | 36.4 | 30.2 | 33.4 |
| SUCRA for secondary ischaemic endpoint | |||||||||||
| 1 | 27.8 | 59.8 | 12.3 | 1 | 10.8 | 73.8 | 15.4 | 1 | 33.2 | 52.3 | 14.4 |
| 2 | 57.9 | 26.5 | 15.6 | 2 | 53.2 | 19.0 | 27.8 | 2 | 51.8 | 26.2 | 22 |
| 3 | 14.2 | 13.7 | 72.1 | 3 | 35.9 | 7.25 | 56.8 | 3 | 15 | 21.5 | 63.5 |
| SUCRA for secondary endpoint of stent thrombosis | |||||||||||
| 1 | 2.01 | 42.4 | 55.6 | 1 | 3.88 | 35.6 | 60.6 | 1 | 2.74 | 45 | 52.3 |
| 2 | 24.0 | 41.5 | 34.5 | 2 | 31.5 | 50.5 | 18 | 2 | 28.9 | 33.1 | 38 |
| 3 | 74.0 | 16.1 | 9.91 | 3 | 64.7 | 13.9 | 21.5 | 3 | 68.4 | 21.9 | 9.67 |
| SUCRA for safety endpoint of bleeding | |||||||||||
| 1 | 70.1 | 9.54 | 20.4 | 1 | 66.7 | 8.13 | 25.2 | 1 | 57 | 23.1 | 19.8 |
| 2 | 25.7 | 26.6 | 47.7 | 2 | 26.7 | 30.6 | 42.6 | 2 | 33.4 | 27.1 | 39.6 |
| 3 | 4.22 | 63.9 | 31.9 | 3 | 6.57 | 61.2 | 32.2 | 3 | 9.62 | 49.8 | 40.6 |
SUCRA values are presented as percentage of area under the cumulative rank probability. The larger of the SUCRA value, the better the treatment regimen performance with respect to the outcome. The highest SUCRA values for each endpoint are highlighted in green.
Ischaemic and bleeding endpoints
Reporting of secondary outcomes in the pre-specified subgroup was variable, with not all trials presenting these data. Definition of major bleeding events was heterogeneous among RCT included (Supplementary material online, Table S1) but was considered acceptable for the purpose of this analysis.23 The networks and diagnostics for secondary outcomes are in Supplementary material online, Figure S8 and Table S5. Again, with a standard approach prasugrel, when compared with clopidogrel, was associated with similar rates of ischaemic events (OR 0.95; 95% CrIn 0.65–1.26; high confidence) and with a parallel, non-significant increase in major bleedings (OR 1.28; 95% CrIn 0.81–2.07; high confidence). Ticagrelor was associated with a non-significant increase in ischaemic events (OR 1.13; 95% CrIn 0.77–1.64; high confidence) and major bleedings (OR 1.14; 95% CrIn 0.71–2.00; high confidence). Instead, our Bayesian SUCRA analysis revealed that prasugrel is the most likely best treatment with regards to the secondary endpoint of cumulative MI and stroke (59.8%) (Figures 3 and 4), while clopidogrel with regards to major bleedings reduction (70.1%). Ticagrelor showed a suboptimal performance with regards to ischaemic event reduction with a SUCRA of 12.3% vs. 27.8% for clopidogrel (Table 1).
Surface under the cumulative ranking curve analysis (SUCRA) for ischaemic and bleeding endpoints. In blue, the area with the highest probability of reduction of ischaemic events (myocardial infarction and stroke); in red, the area with the highest probability of reduction of bleeding events. Arrows indicate SUCRA change if older patients managed invasively are considered.
Ranking plots for primary endpoint (A); secondary ischaemic endpoint (B); bleeding endpoint (C); and endpoint of stent thrombosis (D).
Stent thrombosis
Evidence suggested that both prasugrel (OR 0.42; 95% CrIn 0.04–4.05; high confidence) and ticagrelor (OR 0.32; 95% CrIn 0.01–2.36; high confidence) are associated with a lower occurrence of definite or probable stent thrombosis vs. clopidogrel that is not significant with a standard approach. In a Bayesian framework, ticagrelor was shown to be the best drug to reduce stent thrombosis (SUCRA 55.6%), while clopidogrel the worst with a SUCRA of 2% (Figure 4).
Meta-regression and sensitivity analysis
Our meta-regression confirmed our main results: with a proportion of patients managed invasively fixed at 100%, prasugrel was the most likely best treatment with regards to primary endpoint and of secondary ischaemic endpoint, with SUCRA increased to 61.3% and 73.8%, respectively (Table 1). Clopidogrel and ticagrelor were the best treatment with regards to major bleedings and stent thrombosis, respectively, with similar SUCRA to the main analysis (Supplementary material online, Figure S9).
Moreover, when compared with clopidogrel, the benefit of potent P2Y12 increased linearly with the proportion of patients managed invasively, and this effect was more pronounced with prasugrel when primary and secondary ischaemic endpoints were considered, and with ticagrelor when stent thrombosis was considered (Supplementary material online, Figure S10).
Our sensitivity analysis of 12 676 patients-years excluded a single that randomized patients to prasugrel 10 mg17 and further confirmed our main results. In particular, prasugrel was associated with a higher reduction of primary endpoint (OR 0.98; 95% CrIn 0.85–1.14) and of secondary ischaemic endpoints (OR 0.97; 95% CrIn 0.55–1.43) vs. clopidogrel and remained the best treatment with regards to these endpoint (SUCRA 42.5% and 52.3%, respectively). Clopidogrel was still the most likely best treatment with regards to major bleedings and ticagrelor to stent thrombosis, with similar SUCRA to the main analysis (57% and 52.3%, respectively) (Supplementary material online, Figure S11). Number needed to treat and harm are presented in Supplementary material online, Table S6. NNTnet revealed that 25 patients have to be treated with prasugrel or 33 with ticagrelor for the benefit in terms of primary endpoint to exceed the harm of major bleedings vs. clopidogrel.
Frequentist network meta-analysis
The result of our frequentist meta-analysis further confirmed those of the Bayesian approach. In particular, P-score analysis confirmed that prasugrel have the highest probability of being the best treatment for reduction of primary and secondary ischaemic outcomes and of stent thrombosis, while clopidogrel have the highest likelihood of being the best treatment for major bleeding reductions (Supplementary material online, Table S7).
Discussion
Our pooled network meta-analysis of 14 485 older patients-years suggests that after ACS both prasugrel and ticagrelor confer an advantage vs. clopidogrel in terms of a composite endpoint of all-cause death, MI, and stroke, with prasugrel being the likely best choice. Our analysis also suggests that prasugrel is associated with the lowest rates of MI and ischaemic stroke and clopidogrel of major bleedings.
After an ACS, DAPT with potent P2Y12 is a therapeutic milestone that confer a lower risk of ischaemic events and deaths at the price of a higher hazard of bleeding. Nonetheless, such an advantage was not observed in subgroups of older patients in the cornerstone RCTs17,18 and in recent dedicated trials6,21 which included bleeding events in their composite primary endpoint. Furthermore, since older adults present with high bleeding risk,4 in everyday clinical practice clopidogrel is often preferred to either ticagrelor or prasugrel despite the residual ischaemic risk.5,24 In summary, only few and contrasting data are available to choose the best P2Y12 inhibitor treatment in older patients.25 Despite the absence of statistical significance with traditional meta-analytic approach, the framework of a Bayesian meta-analysis with SUCRA results offer aid in this scenario.
First, the results of our network meta-analysis suggest that also older patients can benefit from potent P2Y12 inhibitors regarding ischaemic event reduction, which has a clear biological rationale: novel drugs are fast-acting and have a reliable metabolism, which might play an even bigger role in older adults with age-related organ changes26 leading to altered drug absorption, bioavailability, distribution and elimination.25 Jointly, potent P2Y12 inhibitors have a SUCRA of 87.4% of being the best treatment with regards to a bleeding-free primary endpoint vs. clopidogrel, which further increase in patients undergoing invasive management (89%). Since such a strategy was demonstrated to be superior to conservative management in older adults,27 this is a clinically relevant result. Prasugrel has the highest chance of being the best treatment after an ACS with regards to this primary endpoint (54.5%).
Second, a similar pattern was observed for our purely ischaemic secondary endpoint: prasugrel had the highest chance of being the best treatment for MI and stroke reduction (59.8%) which further increased in patients managed invasively (73.8%). On the contrary, clopidogrel had lower chance of being the best treatment for ischaemic endpoint (27.8%) and even lower with invasive management (10.8%) (Figure 3). Worse performance of clopidogrel were observed also with regards to stent thrombosis, with very low likelihood of reduction of this endpoint (2%).
Third, both prasugrel and ticagrelor presented a higher hazard of major bleedings, which are associated with early mortality in ACS.28 If reduction of major bleedings is a pivotal concern, clopidogrel is likely the best treatment (70.1%) even in our sensitivity analysis of patients managed invasively (66.7%).
Fourth, the results of our main analysis were confirmed in a sensitivity analysis which excluded the TRITON-TIMI 38,17 the only RCT that used the prasugrel 10 mg dose which is now contra-indicated in subjects older ≥75 years. Similar results were observed in terms of ischaemic risk reduction along with an expected better profile in terms of major bleedings (Table 1). Therefore, our results should be interpreted in the current prescription criteria, prasugrel 10 mg in patients with age >70 and <75 years and prasugrel 5 mg in patients ≥ 75 years.
Fifth, ticagrelor showed suboptimal performance in terms of ischaemic events reductions with a SUCRA lower than clopidogrel in the main analysis but also the lowest rates of stent thrombosis, which corroborates previous findings.26 Also prasugrel presented a large reduction of stent thrombosis, while clopidogrel had the lowest likelihood of reduction of this event (SUCRA 2%). These findings might encourage the use of low-dose prasugrel as the standard potent P2Y12 in older adults and of ticagrelor in selected subjects at higher risk for stent thrombosis and low bleeding risk.
Our analysis included the results of the ISAR-REACT 5 trial22 that demonstrated a superiority of prasugrel vs. ticagrelor in terms of primary endpoint of death, MI, and stroke, which is the same definition of our primary endpoint. This difference might be enhanced in older patients for several reasons including the availability of a lower dose (5 mg) prescribed in subjects ≥75 years, which maintains pharmacodynamic advantages (including faster and reliable platelet inhibition) with a lower price in terms of early bleedings and related mortality. Moreover, increased adherence to therapy with prasugrel (one pill daily) might play an important role in older adults with long, complicated therapeutic regimens.29 Furthermore, as prasugrel is administered only after coronary anatomy is known, the choice of this particular drug might avoid over-prescription in patients with an aborted diagnosis ACS,30 which might be particularly detrimental in older subjects at high bleeding risk.
In conclusion, several parameters should be taken into account when choosing DAPT regimens in older adults and the design of our meta-analysis give some insights (Figure 3). First of all, neither clopidogrel nor potent P2Y12 inhibitors should be precluded to older adults based solely on age. Prasugrel might be a wise choice as it had the best performance in reducing the primary endpoint of all-cause death, MI, and stroke; nonetheless, ischaemic and bleeding burden should always be taken into account. If high ischaemic burden is present, prasugrel offers the highest likelihood of being the best drug to reduce risk of MI and ischaemic stroke, even compared to ticagrelor, and especially if invasive management is planned. On the other hand, if high bleeding risk is present, the prescription of potent P2Y12 inhibitors should be cautious as both were associated with an increased risk of major bleedings compared to clopidogrel, which offers the best reduction of major bleedings. In clinical practice, bleeding and ischaemic risk factors coexist in the same individual. It was previously demonstrated that in the older adults population thrombotic risk spikes in the first months after drug-eluting stent implantation, while bleeding events increase more linearly with DAPT exposure.24 In such scenario, optimal DAPT is a moving target as bleeding and ischaemic risks are not stationary over time, and a short course (up to 1 month with new generation stents)31 of DAPT with prasugrel followed by clopidogrel or only aspirin (de-escalation therapy) might be a reasonable strategy.
Limitations
Some limitations should be considered. Despite the high quality of RCTs included gives our results a high confidence in a Bayesian framework, no formal statistical significance was observed and CrINs were wide: therefore our results should be intended to tailor the best therapy on the individual patient and not as a surrogate of RCT evidence. The POPular-Age trial6 defined older subjects as >70 years, while all other patients included were ≥75 years. Moreover, the same trial randomized patients to ticagrelor or prasugrel, at physician discretion; as 99.6% received ticagrelor, the whole group was considered as randomized to this drug.
Conclusions
Ischaemic and bleeding burden should always be balanced when tailoring DAPT for older adults after an ACS as no standard therapy appears optimal. Potent P2Y12 inhibitors minimize the likelihood of ischaemic events also in the elderlies and low-dose prasugrel appears the wisest choice. Clopidogrel may be sub-optimal, especially if invasive management is planned, but is the safest choice with regards to bleeding events and might be the drug of choice if bleeding risk is prominent.
Supplementary material
Supplementary material is available at European Heart Journal – Cardiovascular Pharmacotherapy online.
Conflict of interest: none declared.
Acknowledgements
In memory of Alberto Castellaneta (Bari, 1935-2020)
