The ‘real world’ is relative—biased

This editorial refers to ‘Hyperglycaemic disorders associated with PCSK9 inhibitors: a real-world, pharmacovigilance study’, by A. Goldman et al.https://doi.org/10.1093/eurjpc/zwab209. More than three decades of statin application in the real world have confirmed the cardiovascular mortality benefits of lipid-lowering (Figure 1), have led to a shift in the phenotype of atherosclerotic plaque morphology, probably demanding new, more individualized treatment approaches after acute coronary syndrome and have—not least!—forced us to think about how we communicate relative risks and adverse drug effects to our patients.1–5 Graphic depiction of cardiovascular benefit, main mechanism of action as well as glucose metabolism-related (red) and other (grey) main side effects of statins and PCSK9 inhibitors. *PCSK9 inhibitors were given on top of statin and compared to a control group receiving guideline lipid lowering (incl. statin). The development of inhibitors of the proprotein convertase subtilisin/kexin type 9 (PCSK9) has been met with excitement about likely even greater anti-atherogenic effects, due to a stronger reduction of plasma low-density lipoproteins together with the lowering of further pro-atherogenic lipid species—triglycerides and lipoprotein(a) (Figure 1).6,7 Regarding off-target effects, we have as of yet a limited number of high-quality randomized controlled trials and only a few years of follow-up data in the real world. The study by Goldman et al.8 in this issue of the European Journal of Preventive Medicine aimed to better estimate the risk of PCSK9 inhibitor-induced dysregulation of glucose metabolism by querying the adverse event reporting system of the federal drug administration (FAERS) for reported adverse events in patients using PCSK9 inhibitors as compared with statins and ezetimibe. They retrospectively analysed over 7 million cases from 2015 to 2020, of which 0.98% reported on evolocumab and 0.22% on alirocumab. In order to compare both drugs, we would need to know the numbers of patients treated with each drug. Unfortunately, this information is not available from FAERS, but we know from economic reports that prescriptions of alirocumab were approximately half of those of evolocumab in April 2019 and continued to diverge even more. Both PCSK9 inhibitors together, as well as evolocumab alone, but not alirocumab alone, were associated with a higher reporting rate of hyperglycaemic events as compared with the whole database, which was higher than for ezetimibe but lower than reported for statins. PCSK9 inhibitor-related reports were more often about mild hyperglycaemia than about diabetes and concerned diabetic patients more frequently than non-diabetic patients. In a small subgroup of patients, where this information was available, hyperglycaemic events disappeared upon discontinuation of the drug. Can we reconcile those observations with our current biochemical understanding of both drugs? Muscle pain, new-onset diabetes mellitus (NODM) and liver damage are among the most often discussed side effects of statins and differ with the statin taken and dosage, as well as patient’s characteristics and co-medication (Figure 1).9 They can occur since mevalonate, the product of the statin target 3-hydroxy-3-methylglutarylcoenzyme A (HMG-CoA) reductase, is not only needed for the synthesis of cholesterol and subsequent products but also in non-sterol isoprenoid synthesis. Importantly, HMG-CoA synthase is not specific to the liver but is also expressed in other tissues, including the pancreas, adipose tissue, skeletal muscle, and others. Isoprenoids are widely used in our biochemical makeup. They are integral components of the coenzyme Q10—crucial for mitochondrial respiration and adenosine triphosphate (ATP) synthesis. Isoprenoids are also needed to anchor GTPases to the cell membrane, an important prerequisite for many signal transduction processes, including insulin-mediated signalling. A number of further mechanisms, including the maturation of preadipocytes,10 have been described to be impaired by statins, thereby providing explanations for the increased risk for NODM seen with some statins—mainly the lipophilic ones (reviewed in11). Data regarding adverse effects of PCSK9 inhibition on glucose metabolism are less clear. Mechanistic studies in cell culture and animal models, as well as observations in human patients carrying the loss-of-function PCSK9 p. R46L mutation, show heterogeneous results.12 PCSK9 facilitates the internalization and degradation of the LDL receptor—in the liver as well as in extrahepatic tissues (Figure 1). Maintained cellular LDL receptor expression through PCSK9 inhibition therefore enhances hepatic cholesterol import. The cardiovascular benefit—reduced plasma cholesterol levels—is thought to be mediated mainly through the stabilization of liver LDL receptors, but the mechanism is active also in other LDL receptor expressing tissues, including the pancreas. PCSK9 inhibitors may promote intracellular increase of cholesterol in pancreatic beta cells, thereby impairing insulin secretion.13PCSK9 knockout mice also have lower pancreatic insulin levels, glucose intolerance and show altered beta-cell histology, albeit beta-cell specific knockout did not confirm those results.14,15 Hence, in vitro and in vivo data might indeed help to explain the generally lower rates of hyperglycaemic adverse events with PCSK9 inhibitors as compared with statins and stronger association of PCSK9 inhibition to hyperglycaemia, but not NODM, while statins show higher rates of NODM. The interpretation of data from adverse event reporting systems is fraught with limitations and caveats, including under-reporting, partial, erroneous, and multiple reporting of individual cases by caregiver, patient, and lawyer, and is often lacking data relevant for better interpretation of the observations, such as co-therapy. Information on co-medication is often limited in reporting systems,16 but might give valuable information on drug combination associated with higher vs. lower numbers of events, especially in patients with type 2 diabetes mellitus or with metabolic syndrome. In addition, geographic differences in drug prescription and adherence as well as reporting of adverse events and differences regarding the person reporting—caregiver, the patient themselves or a lawyer—differ, as is evident for the reporting of evolocumab vs. alirocumab in the current study.8,16 High costs currently limit PCSK9 inhibitor prescription to selected patient groups in some European countries due to reimbursement policies and thereby provide a socio-economic as well as co-morbidity associated bias.17 Finally, reporting is heavily influenced by public opinion, such as prominent discussion of adverse effects in the media and company advertising. Often, adverse events are observed more frequently in the real world.18 On the other hand, adverse drug effects might not always be really attributable to the drug suspected, as in the case of statins.19 ‘Public opinion’ effects are mirrored by fluctuations of reporting over time, as well as the ‘Weber effect’, a peak in the 2nd year after a drug enters the market.16,20 Nevertheless, analyses of reporting systems databases shed light on a corner of drug therapy that is not covered by clinical trials with their carefully selected patient cohorts and well supervised treatment, thus exerting a greater psychological pressure to stay adherent, in addition to free distribution of the study medication. In real life, no previous cardiovascular event is necessary to qualify for prescription of PCSK9 inhibitors, as has been defined for the ODYSSEY OUTCOMES study.7 Moreover, clinical trials on PCSK9 inhibition have been conducted in the presence of a high-dose statin, while in the real world, a considerable number of patients might receive PCSK9 inhibitors without concomitant other lipid-lowering drugs, often because of intolerance or non-adherence to the initial medication.17 The presence of the statin in both control and treatment group may have masked any potential PCSK9i effect on glucose metabolism in clinical trials, where similar molecular mechanisms are involved. Compared with statins, real-world experience with PCSK9i is very limited yet and the manuscript by Goldman et al.8 adds to our knowledge. Ongoing surveillance and interpretation of data will require a close look at biases in the reporting systems, including comorbidities and socio-economic aspects of patients, as well as regional guidelines on reporting, prescription and reimbursement. Strengthening and expansion of registries and surveys such as the EUROASPIRE programme will provide an additional valuable set of information from a separate viewpoint. Cardiovascular benefit, albeit expected to be high, and a favourable profile of adverse effects are not the only relevant factor deciding on the success of PCSK9i in clinical practice. Especially, affordability and reimbursement criteria can represent a reason for discontinuation or non-prescription of PCSK9i.17 Conflict of interest: The author has no conflicts of interest pertinent to this article. The opinions expressed in this article are not necessarily those of the Editors of the European Journal of Preventive Cardiology or of the European Society of Cardiology.