Stopping Statins

• Editorials
• statins
• coronary disease

Within every difficulty lies an opportunity. — —Albert Einstein

How can you safely enroll subjects on statin therapy with coronary heart disease (CHD) for a clinical trial that randomizes subjects to this therapy only after a suitable dietary lead-in/drug-washout period? This vexing problem confronted the investigators of the Treating to New Targets (TNT) study.¹ Their concern was prompted by the alarming report of Heeschen and colleagues,² who found a nearly 3-fold increase (2.93 [95% CI, 1.64 to 6.27]; P=0.005) in the risk of death and nonfatal myocardial infarction (MI) in those in whom statin drugs were not continued after admission with an acute coronary syndrome.² A follow-up letter to the editor revealed that this was a statistical rather than a clinical phenomenon. In approximately one third of the study population, missing data for troponin T at baseline were excluded from the multivariate model used to adjust for possible confounding factors. A reanalysis showed only a nonsignificant trend toward greater cardiac risk with abrupt statin discontinuation.³ The authors concluded that there was insufficient evidence to support their original conclusions. Nonetheless, this raised questions about potential deleterious effects of stopping statins in both acute and chronic CHD patients.

See p 2333

Fortunately, the authors of the TNT trial seized on the opportunity afforded to them by their study design to see if there were significant adverse effects from statin withdrawal in their study of subjects with stable CHD.⁴ TNT is an ongoing, large-scale trial with a double-blind parallel group design comparing 2 doses of atorvastatin (10 and 80 mg once daily) to determine clinical end point differences. This large effort enrolled 16 619 in a dietary lead-in/drug-washout period and, of these, 15 432 eligible subjects began treatment with atorvastatin 10 mg/d on an open-label basis. Of the participants who entered the dietary lead-in/drug-washout period, 57% were currently receiving statin therapy and had their statin therapy discontinued. When event rates for CHD were carefully examined, it was determined that statins could be safely withdrawn, as the 30-day Kaplan-Meier event rate of 0.20% for the combined end point of CHD death, nonfatal MI, resuscitated cardiac arrest, and fatal or nonfatal stroke during the washout phase was not significantly different from the 0.26% rate seen in the open-label phase with the participants on atorvastatin 10 mg/d.

Was this concern biologically plausible? Statins are the popular name given to a class of cholesterol-lowering drugs known as HMG-CoA reductase inhibitors.⁵ They inhibit the enzyme HMG-CoA reductase, which is critically involved in the rate-limiting step of cholesterol biosynthesis in the liver. Within weeks of starting treatment with a statin, there is a decreased rate of cholesterol synthesis that leads to an upregulation of low-density lipoprotein (LDL) receptors and a resultant fall in blood cholesterol levels. LDL cholesterol (LDL-c) and intermediate-density lipoprotein levels fall in a dose-dependent fashion (available statins vary in potency); triglycerides fall in proportion to LDL-c lowering; and there are small but variable increases in high-density lipoprotein cholesterol (HDL-c). Available clinical trial data point to the significant LDL-c lowering produced by statins as the major reason for their remarkable efficacy in improving total mortality^6–8 and reducing CHD events.⁹

There continues to be great interest in the pleiotropic effects of statins.¹⁰ These include effects on endothelial dysfunction, the inflammatory response, the stability of the atherosclerotic plaque(s), nitric oxide availability, and even platelet function and/or thrombus formation. Davignon¹⁰ has recently reviewed these effects in this journal, noting that these pleiotropic effects were characterized by both rapid onset and offset. It can be argued that in the acute coronary syndrome, in which inflammation promotes rupture of atherosclerotic plaques and inflammatory markers are present in high concentration, the pleiotropic effects of statin therapy could be more important in determining outcomes than in the chronic state.¹¹ Indeed, when Brigham investigators examined samples for CD40 ligand, a proinflammatory mediator, from subjects enrolled in the Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) Study, they demonstrated that not only were particularly high levels of CD40 ligand associated with recurrent events, but, notably, the excess risk attributed to them was eliminated by the high-dose (80-mg/d) atorvastatin therapy.¹² Furthermore, in patients with acute or impending stroke, data from mouse models of cerebral ischemia show that acute termination of statin treatment results in a rapid loss of protection independent of lipid lowering.¹³ Thus, further prospective studies of those who are unable to continue with statin therapy in acute settings for both coronary and cerebral ischemia are needed.

Given the extraordinary record of the success of statins in reducing coronary events in patients with moderate to very high risk, what difficulties await as we begin to prescribe statin therapy to a growing number of patients who require it for optimal risk reduction? Stopping statins may become a greater issue as physicians, for optimal CHD risk reduction, will prescribe, as recommended, either higher doses of statins or more complex regimens.¹⁴ Indeed, it is now suggested that a certain percentage lowering of LDL-c is needed for optimal benefit. Thus, the new Adult Treatment Panel III implications article¹⁴ recommends minimal statin doses that result in at least a 30% to 40% reduction in LDL-c. In addition, the Adult Treatment Panel III panel suggested that for those at “very high” CHD risk, a new, lower goal for LDL-c lowering of <70 mg/dL was a reasonable strategy on the basis of available data (pending studies like TNT that are already in progress). They suggested that this optional LDL-c goal should be considered in those with:

• Acute coronary syndrome,

• CHD and other severe risk factors such as diabetes,

• CHD and multiple risk factors such as seen in those with metabolic syndrome, and

• CHD and uncontrolled risk factors such as cigarette smoking.

Moreover, when a patient with CHD has a high triglyceride level and/or a low HDL-c level as well as an increased LDL-c level, the addition of a fibrate or nicotinic acid to an LDL-lowering drug can be considered. This can be done more safely now because the fenofibrate-statin combinations do not result in the raised statin concentrations that were seen with gemfibrozil-statin combinations, as a result of gemfibrozil’s interference with statin glucuronidation in the liver.¹⁵ Moreover, once daily, extended-release niacin seems to be better tolerated than short-acting niacin, is associated with less liver toxicity than long-acting niacin, and can be safely combined with statins.¹⁶

Although stopping statins shouldn’t occur frequently, side effects can occur. A clinical advisory on the use and safety of statins pointed out that these are generally well-tolerated drugs.¹⁷ Elevated liver transaminases (defined as >3 times the upper limit of normal in most studies) occur in 0.5% to 2.0% of clinical trial subjects and are dose dependent. Indeed, the American College of Physicians’ background paper on pharmacological lipid lowering in people with type 2 diabetes suggested that routine monitoring of liver or muscle enzymes probably is not warranted unless patients have symptoms or baseline liver enzyme abnormalities or are taking drugs that interact adversely with statins.¹⁸ In the PRavastatin Or atorVastatin Evaluation and Infection Therapy (PROVE-IT) trial, 80-mg (high-dose) atorvastatin therapy resulted in elevated liver transaminases in 3.3% as contrasted with only 1.1% in the pravastatin group (P<0.001), indicating that liver testing is worth the effort in those given high-dose statin therapy for acute coronary syndrome.¹⁹

Myopathy and muscle-related symptoms and/or creatine kinase (CK) elevations are also concerns for clinicians. The statin clinical advisory pointed out that to reduce the risk of myopathy with statin drugs, physicians should obtain baseline liver and CK measurements and consider them before prescribing statins to those who are at higher risk for statin-associated myopathy. I suggest 4 general categories to consider to circumvent problems that require stopping statins:

• Personal characteristics: advanced age, especially in women and in patients >80 years; elderly women; frailty; small body frame; alcohol abuse; muscle disorders; and grapefruit juice intake (those at risk consume large quantities and take a statin that uses the 3A4 P450 system for its metabolism)

• Disease burden: multisystem disease, especially chronic renal insufficiency with diabetes; unrecognized hypothyroidism; and antecedent liver or muscle disease (although fatty liver, if confirmed, may improve with lipid-lowering therapy)

• Medication burden: use of multiple medications (eg, by an HIV patient), particularly medications that interact with statins via a variety of mechanisms (including fibrates [especially gemfibrozil], niacin [rarely], cyclosporine, and warfarin) or medications that inhibit aspects of the P450 cytochrome system, which is used by some of the statins for their metabolism (eg, azole antifungals, itraconazole and ketoconazole, erythromycin and clarithromycin, HIV protease inhibitors, verapamil, and amiodarone elevate concentrations of statins that use the 3A4 P450 system, and fluconazole raises levels of statins that use the 2C9 P450 system)

• High-risk clinical situations: perioperative periods for major surgery in which unanticipated hypotension, heart or renal failure, or need for potent intravenous medications could increase the risk of drug toxicity

Careful readers, note that in the TNT trial, for example, subjects were not enrolled if the liver transaminases exceeded 1.5× the upper limit of normal or unexplained CK levels >6× the upper limit of normal were found. This suggests a useful clinical strategy to avoid stopping statins. Approach each patient to whom you prescribe a statin as if he or she were in a clinical trial. Use the above list to help in statin selection. Also, avoid stopping statins if only minor elevations of liver transaminases are seen. Take a careful history if intermittent vague aches and pains occur; most do not require stopping statins. Do listen for and examine for muscle weakness; causes of weakness must be determined, whether they be statin induced (can occur with a normal CK) or not.²⁰

The TNT investigators have shown that in stable CHD patients, stopping statins for a short time is safe. We do not know if it is safe to permanently stop statins or discontinue them for an extended period of time, especially in the acute setting. If they have to be stopped, strategies should be considered to restart either at a lower dose or with another statin that is better tolerated because therapy with statins has been shown to be so effective in both primary and secondary CHD prevention.