Meta-analysis questions strength of ties between statin-induced LDL lowering, CV outcomes

John T. Wilkins, MD, MS

It's not clear to me what this paper adds to our current understanding.

Absolute benefits are always lower than the relative risk reductions reported in clinical trials. This is a consequence of basic math. Relative risks are ratios of the event rates in the intervention and nonintervention groups (typically expressed as a percentage), whereas absolute benefits are differences in the event rates between groups. For statins, the relative risk is consistent across different groups of baseline risk, so that the absolute benefits can be derived by multiplying the relative risk reduction by the baseline absolute risk of the treatment group. The only way an absolute risk would ever be equal to a relative risk would be if the baseline risk for an outcome is 100%, which is never the case.

It's important to remember that risk reductions are observed over different time horizons. This analysis seems to ignore follow-up time. Most of these trials were 2 to 4 years in length, yet we prescribe statins for much longer time horizons, which results in greater cumulative absolute benefit to their use.   
The details of which trials are included in analyses like this have a strong effect upon the conclusions. This study includes fewer studies than have been reported in other meta-analyses, which may affect the results as well.  

The researchers’ inconclusive meta-regression that fails to determine if degree of LDL lowering predicts ASCVD events is an artifact of their methods. If the authors used individual-level data (as was done previously in studies that demonstrated the association between lowering of LDL and risk reduction), they likely would have had power to detect the association.  

Clinical recommendations as well as clinical decision making should incorporate the balance between the expected absolute benefit and absolute risks for any given intervention. Consistent with this, current and past guideline recommendations for statin therapy in primary prevention have been based upon expected absolute risks and benefits of statin therapy in different patient subgroups.    

It is very well understood that use of statins in low-risk primary prevention patients likely has low absolute near-term benefits in terms of ASCVD and mortality risk reduction.  

At best, this study could serve as a reminder to consider absolute risks and benefits when considering any clinical intervention (lifestyle, drug, surgical, etc) for a patient.  

There are significant limitations to the methods employed in this analysis and there is nothing being presented here that has not been carefully considered by guideline committees in the past.

Udho Thadani, MD, FACC, FAHA, FCCS

Statins, by lowering total and LDL cholesterol, reduce the risk for major vascular events (coronary deaths or nonfatal MIs, strokes and coronary revascularization procedures). Current guidelines recommend the use of statins to improve CV outcomes for both secondary and primary prevention. Physicians often discuss with their patients the relative risk (RR) reduction rather than the absolute risk (AR) reduction of statins, as it is often difficult to predict AR reduction in an individual patient.

Byrne and colleagues in their systematic review and meta-analysis conclude that statin-related AR reductions are only modest compared to RR reductions for all-cause mortality (0.8% vs. 9%), MI (1.3% vs. 29%) and stroke (0.4% vs. 14%). They have appropriately excluded from their analysis the softer endpoints of coronary revascularization procedures. They suggest that physicians should discuss with their patients AR reductions rather than RR reductions when starting statin therapy. They conclude that 77 participants would need to be treated with a statin for 4.4% years, on average, to prevent 1 MI. Furthermore, in their discussion, they mention the study by Rory Collins, FRS, and colleagues which reported that 5-year treatment of 10,000 patients with an adequate dose of a potent statin would result in 5 cases of myopathy, 50 to 100 cases of diabetes and 5 to 10 cases of hemorrhagic stroke (Collins R, et al. Lancet. 2016;doi:10.1016/S0140-6736(16)31357-5).

Although I agree with Byrne and colleagues’ suggestion that the physician should discuss AR reduction and adverse effects of statin therapy with their patients, it is equally important to discuss the beneficial effects of statin therapy as reported by Collins and colleagues. Based on individual patient data meta-analysis they reported that by treating 10,000 patients with a statin, one would prevent 1,000 (10% AR reduction) major CV events. Furthermore, this reported beneficial effect of statins had already taken into consideration the reported adverse events quoted by Byrne and colleagues.

In my opinion, it  is unlikely that current guidelines will be changed based on the review by Byrne and colleagues, as the guidelines have already factored individual data-based meta-analysis in their recommendations.