Author + information
- M. Bilal Iqbal, MD, PhD∗ (, )
- Imad J. Nadra, MD, PhD,
- Lillian Ding, MSc,
- Anthony Della Siega, MD and
- Simon D. Robinson, MD
- ↵∗Victoria Heart Institute Foundation, 1900 Richmond Road, Victoria, BC, V8R 4R2, Canada
We thank Dr. Giri and colleagues for interest in our recent paper (1). They raise an important issue of survivor bias, which as correctly pointed out, is a limitation to all observational analyses that have addressed multivessel disease in ST-elevation myocardial infarction. To address survivor bias, we have now performed conditional Kaplan-Meier and landmark analyses to explore the influence of early mortality on outcomes (2,3), using crude landmark times of 10 and 30 days. In these analyses patients who survive beyond the landmark times are included, potentially removing the effect of early mortality on long-term outcomes.
Figure 1 shows conditional Kaplan-Meier landmark analyses, with statistically significant log-rank tests. Applying the landmark method at 10 days, Cox regression analyses demonstrate that culprit vessel intervention (CVI) at index procedure followed by staged intervention (CVI-S) is associated with lower 2-year mortality when compared with multivessel intervention at index procedure (MVI) (hazard ratio [HR]: 0.58; 95% confidence interval [CI]: 0.38 to 0.89; p = 0.013) and culprit-vessel-only intervention at index procedure (CVI-O) (HR: 0.68; 95% CI: 0.46 to 0.99; p = 0.049). Similarly, when applying the landmark method at 30 days, Cox regression analyses demonstrate that CVI-S is associated with lower 2-year mortality when compared with MVI (HR: 0.54; 95% CI: 0.33 to 0.89; p = 0.017) and CVI-O (HR: 0.60; 95% CI: 0.46 to 0.94; p = 0.026). It is important to note that landmark analyses can reduce the statistical power to detect differences. Nevertheless, these analyses demonstrate that even when accounting for early mortality, CVI-S continues to be associated with lower mortality when compared with either MVI or CVI-O.
Giri et al. have pointed out that our findings are similar to previously reported observational data, which had led to inappropriate consensus guideline recommendations that have recently been altered with the emergence of small-scale prospective randomized data. We would like to clarify that this is not completely true. Although our study suggests that culprit-vessel should be treated at the index presentation, it also suggests better outcomes with staged revascularization of nonculprit disease before discharge. This supports a strategy of staged complete in-hospital revascularization consistent with recent CVLPRIT (4) and DANAMI-PRIMULTI (5) trials.
We acknowledge that the propensity-matched and inverse probability of treatment-weighted analyses demonstrate a greater than anticipated mortality difference between MVI and CVI-S. This is likely to reflect hidden bias, naturally inherent to observational analyses. Given that 20% of patients underwent MVI, there is certainly a strong selection process, and patients undergoing MVI may represent a sicker cohort or those with hemodynamic instability. For example, a greater portion in the MVI group received intra-aortic balloon pump, implying a sicker patient population. This raises the issue of unmeasured confounding. We would like to highlight that propensity-score methods only address measured confounders and residual confounding due to unmeasured factors cannot be excluded.
We acknowledge the limitations of observational studies, and regardless of the statistical methodologies, there will always be some residual bias. Thus, our results should only be considered as hypothesis generating. Although we have used landmark methods to partially address the issue of survivor bias, we agree with Giri et al. that survivor bias can only truly be overcome through prospective assignment to treatment strategies. We certainly look forward to the results of the COMPLETE trial (6).
Please note: All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Iqbal M.B.,
- Nadra I.J.,
- Ding L.,
- et al.
- Anderson J.R.,
- Cain K.C.,
- Gelber R.D.
- Dafni U.
- Engstrom T.,
- Kelbaek H.,
- Helqvist S.,
- et al.
- ↵Complete vs Culprit-only Revascularization to Treat Multi-vessel Disease After Primary PCI for STEMI (COMPLETE). ClinicalTrials.gov 2013. Available at: http://clinicaltrials.gov/ct2/show/NCT01740479. Accessed February 2014.