Myopathy in older people receiving statin therapy: a systematic review and meta-analysis

Abstract

Objective

The aim of the present study was to determine the risk of myopathy in older people receiving statin therapy.

Methods

Eligible studies were identified searching Ovid Medline, EMBASE, Scopus, CINAHL, Cochrane and PSYCHINFO databases (1987 to July 2014). The selection criteria comprised randomized controlled studies that compared the effects of statin monotherapy and placebo on muscle adverse events in the older adult (65+ years). Data were extracted and assessed for validity by the authors. Odds ratios and 95% confidence intervals (CIs) were used to calculate binary outcomes. Evidence from included studies were pooled in a meta-analysis using Revman 5.3.

Results

The trials assessed in the systematic review showed little or no evidence of a difference in risks between treatment and placebo groups, with myalgia [odds ratio (OR) 1.03, 95% CI 0.90, 1.17; I² = 0%; P = 0.66] and combined muscle adverse events (OR 1.03, 95% CI 0.91, 1.18; I² = 0%; P = 0.61) (myopathy). No evidence was found for an increased risk of rhabdomyolysis (OR 2.93, 95% CI 0.30, 28.18; I² = 0%; P = 0.35) in the seven trials that reported this. No trials reported mortality due to a muscle-related event. Discontinuations due to an adverse effect were reduced in the treatment group compared with placebo (OR 0.74, 95% CI 0.50, 1.09; I² = 0%; P = 0.13).

Conclusion

The results obtained from the present review suggest that statins are relatively safe, even in older people. There was no evidence to suggest an increased risk of myopathy in older adults receiving statin therapy. There is slightly increased seen with rhabdomyolysis when compared with the general population, although the event is relatively rare. Statins should be prescribed to elderly people who need it, and not withheld, as its myopathy safety profile is tolerable.

Introduction

Statins are generally safe and effective but, as with most medications, they are associated with adverse effects. Statin-associated myopathy is one of the most common adverse effects seen [1]. This constitutes a spectrum of symptoms, ranging from an asymptomatic rise in serum creatine kinase, to mild-to-moderate pain or discomfort (myalgia), myositis (an inflammation of the muscle) and, ultimately, rhabdomyolysis [2]. Myotoxic symptoms are usually reversible on discontinuation or dose reduction [3] and occur within weeks to months after initiation of statin therapy [4,5]. The incidence of the most common symptom, myalgia, ranges from about 1% to 5% in clinical trials as compared to up to 25% in observational studies [2]. A recent survey [Understanding Statin use in America and Gaps in Education (USAGE)] found that up to 60% and 25% of former and current statin users, respectively, reported muscle-related side effects [4]. The risk of myopathy is higher in patients aged over 80 years, with poly-pharmacy, presence of co-morbidities, reduced muscle mass, and impaired renal and liver function [6]; these are seen frequently in older adults, leading to the effects of drug intensification experienced in this age group [7,8].

Historically, older adults were excluded from large randomized controlled trials (RCTs) involving statins [9] until publications from the Heart Protection Study Collaborative group [10], Study Assessing Goals in the Elderly (SAGE) [11] and Prospective Study of Pravastatin the Elderly at Risk (PROSPER [12]) provided data on the efficacy of statins from RCTs in patients older than 65 years of age. These trials showed that, even though older patients had the highest risk of cardiovascular disease and were most likely to derive the most benefits from lipid-lowering therapy, they are the least likely group to receive a statin [13].

Statins remain underused, with only about half of those who should benefit from a statin receiving one [14]. Patients >65 years are less likely to receive a statin prescription compared with their younger counterparts [15,16]. In a study by Allen Maycock [9], 19.8% of patients aged over 80 years, 21.2% aged 65–79 years and 28% <65 years were taking a statin (P < 0.001). Several factors account for this age differential, including concerns about safety and confusing evidence concerning risk factors in this group. However, the major safety issue associated with statin therapy is the fear of myopathy, especially in the older population [14,17].

The present study aimed to determine the risks of muscle-related adverse effects in older patients receiving statin therapy.

Methods

Trials selected

We selected RCTs which compared statin therapy with placebo or usual care, including both primary and secondary prevention trials. There was no limit to study duration. All analysed studies were based on the intent-to-treat (ITT) principle (see Table 1).

Table 1.

Baseline characteristics of included studies

Study	Country	Statin	Duration (years)	Age range (years)	Mean age years(SD)	Participants (N)	Statin (N)	Placebo (N)	Gender (% male)	Race (% white)
AFCAPS/TexCAPS 1998 [18]	USA	L†	5.2	65–75	NR	1416	715	701	75	NR
CARDS 2004 [19]	UK/Ireland	A†	3.9*	65–76	69 (NR)	1129	572	557	69	96
GISSI-HF 2008 [20]	Italy	R‡	3.9*	≥70	NR	2014	1002	1012	NR	NR
JUPITER 2008 [21]	26 countries across North/South America, Europe and Africa	R‡	1.9	70–97	74 (72–77)*	5695	2878	2817	49	70
MIRACL 2001 [22]	Europe, North America, South Africa, Australia	A†	0.3	≥65	74 (6.0)	1672	849	823	59	90
Pedersen (4S) 1994 [23]	Nordic countries	S†	5.4*	≥65	67 (1.4)	1021	518	503	76	NR
PLAC I 1995 [24]	USA	P‡	3.0	NR	NR	94	42	52	NR	NR
PROSPER 2002 [12]	Scotland, Ireland, the Netherlands	P‡	3.2	70–82	75 (3.3)	5804	2891	2913	48	NR

Abbreviations are as follows: NR, not reported; N, number of participants; L, lovastatin; A, atorvastatin; R, rosuvastatin; S, simvastatin; SD, standard deviation; P, pravastatin;

^*median (interquartile range);

^†lipophilic statin;

^‡hydrophilic statin.

Types of participant

Men and women aged 65 years and older, regardless of ethnic background were eligible. No restrictions were applied on cholesterol levels.

Types of intervention

Statins of any type and at any dose were eligible for inclusion, excluding cerivastatin, which was withdrawn from the market in 2001 [25]. Only statins given as monotherapy, and not in combination with any other cholesterol-lowering drug, were included.

Primary outcomes

The definitions used in the present review were taken from the American College of Cardiology/American Heart Association/National Heart Lung and Blood Institute (ACC/AHA/NHLBI) clinical advisory [26].

• Myopathy: any statin-associated muscle disease;

• Myalgia: muscle ache or weakness without increased creatine kinase (CK) levels;

• Rhabdomyolysis: muscle symptoms associated with marked elevation of CK levels greater than 10 times the upper limit of normal, with or without myoglobinuria. This can be life-threatening without intervention.

Secondary outcomes

• Withdrawal from trial or discontinuation due to adverse effects;

• Mortality from myopathy-related cause.

Search strategy

In July 2014, eligible studies were identified searching Ovid Medline, EMBASE, and the Cochrane, PSYCHINFO, Scopus and Cumulative Index of Nursing and Allied Health Literature (CINAHL). The National Institutes of Health Clinical Trials website (http://www.clinicaltrials.gov) was searched. Grey literature was accessed and a manual review of reference list carried out. Authors of unpublished or ongoing research that was relevant to the present study were contacted.

A combination of Medical Subject Headings (MeSH) terms was used. ‘hydroxymethylglutaryl coenzyme A reductase inhibitors’ or ‘anticholesterolaemic agents’ or ‘statins’ or ‘pravastatin’ or ‘fluvastatin’ or ‘simvastatin’ or ‘atorvastatin’ or ‘rosuvastatin’ or ‘lovastatin’ and ‘randomized controlled trials’ and ‘myopathy’ or ‘muscle disease’ and ‘aged’ or ‘elderly’ or ‘65 years and over’ and ‘adverse effects’ or ‘side effects’. Full details of the search strategy are included in the supplementary files. Publications were limited to those in the English language.

Data collation and analysis

Selection of studies Retrieved articles were exported to a predefined extraction form and duplicate records deleted. Articles were screened and irrelevant articles excluded. Both authors then conducted a detailed review of abstracts. The full texts of potentially relevant trials were obtained and those which did not meet the eligibility criteria were excluded.

Assessment of risk of bias in included studies The methodological quality of included studies was assessed using the criteria outlined in the Cochrane Handbook for Systematic Reviews [27], using the following domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective outcome reporting bias and other biases.

Each domain was assessed and graded as high, low or unclear risk and a risk of bias graph was generated.

Data synthesis Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using binary data, using RevMan 5.3 software (Cochrane). Quantitative analysis was carried out based on the ITT principle. An I² value was generated and >50% was taken to indicate moderate (or higher) heterogeneity between trials. A fixed-effects model was used, except in cases were heterogeneous results were obtained, in which case the random effects model was employed [28]. Attempts were made to explain differences when significant heterogeneity occurred. We planned to assess publication bias using funnel plots if a sufficient number of trials were obtained [29].

Subgroup and sensitivity analysis Subgroup analyses of solubility (lipophilic vs. hydrophilic statins) and dose (high vs. low) were performed. There was insufficient information to analyse other a priori subgroups. Sensitivity analyses of trial quality, those finishing early, and primary and secondary prevention were also conducted.

Results

Search results

The initial database search yielded 434 citations. An additional four articles were obtained from manual review of identified articles. A total of 438 articles were screened; 397 articles were excluded after a title and abstract screen and 19 were excluded for the following reasons: eight duplicates, six review articles and two observational studies were found, and the remaining three either compared varying doses of statins or co-administered statins with other therapies.

Full-text articles were obtained for the remaining 22 articles and were assessed for eligibility. Fourteen [10,11,30–41] were excluded for varying reasons, shown in Supplementary Table 1. Ultimately, eight RCTs were included in the present review. Published and unpublished data were sought from authors and included in the review (see Figure1).

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart

Characteristics of included studies

The studies were published between 1994 [23] and 2008 [20,21]. The total number of elderly participants was 18 845, contributing to a mean weighted follow-up period of 2.9 years (54 650 patient-years).

All studies had a multicentre design. Study locations were variable: Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) spanned 26 countries, covering North and South America, Africa and Europe. All the studies, except Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI-HF) and Pravastatin Limitation of Artherosclerosis in Coronary Arteries (PLAC-1) [20,24], recruited patients from Europe. Study population varied in size, from 94 (0.005%) patients in PLAC-I [24] to 5804 (31%) in PROSPER [12]. Study duration ranged from 16 weeks Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) Trial [22] to about 5.4 years in Scandinavian Simvastatin Survival Study (4S) [23].

Only one trial, Prospective Study of Pravastatin in Elderly at Risk (PROSPER) [12] solely randomized patients older than 65 years; the other studies either did a subgroup or post hoc analysis of older people. Three studies [18,19,21] were primary prevention studies, contributing to 43.7% of the study population. Two studies did not report gender proportions [20,24]; of those that did, the male population ranged from 48% to 76% of the total. Participants were predominantly Caucasian (70–96%). The mean age of participants ranged from 67 (±1.4) years to 75 (±3.3) years.

All trials compared statin with placebo; there were two studies with each of rosuvastatin, pravastatin and atorvastatin, with one for lovastatin and simvastatin. No studies were included for fluvastatin.

Three trials [18,19,21] were stopped early (1.4–3.0 years) on the grounds of efficacy. Data on outcomes were extracted from the safety and adverse reports for each study. Reports on myalgia were provided in four trials; rhabdomyolysis was reported in seven studies. Myopathy was reported in all included studies. Discontinuation or withdrawal from the study due to adverse effects was reported in three trials. None of the studies reported mortality due to a muscle adverse effect.

Risk of bias in included studies

Randomization sequence generation was deemed adequate in half of the studies [12,19,20]. Four studies [18,22–24] provided a limited description of how the randomization sequence was generated and were deemed unclear. Four studies did not describe the allocation concealment process [18,19,22,24]. Three studies [12,20,21] had the lowest risk in allocation concealment, and the 4S study [23] had an unclear risk. All studies were blinded to a high standard and none showed any selective reporting. As we identified fewer than 10 studies, we were not able to generate funnel plots. Six studies [12,18–21,23] actively followed participants through routine visits where adverse effects were evaluated. Of these, two studies [18,21] employed the use of questionnaires and a structured interview to obtain adverse events information, whereas the others did not describe the methods or tools of assessment of events. The methods of collection and evaluation were not described at all in two studies [22,24]. Outcomes were defined in all studies and no groups of patients were excluded from adverse effect analysis in any study. All trials gave a clear summary of the statistical methods used and provided a priori sample size and power calculations, which were both met in all cases. A detailed description of the recruitment process was provided in all studies. All trials included were analysed based on the ITT principle. Four studies [12,19,22,23] provided information on dropouts or withdrawals due to adverse effects of intervention. All of the studies were funded by pharmaceutical companies, of which three declared no involvement other than funding. The risks are summarized in Figure2.

Figure 2.

Risk of bias table

Effect of intervention

Forest plots were employed to display graphically the results from each trial and the resultant pooled estimates [28]. Figure3 summarizes the main findings for our primary and secondary end points.

Figure 3.

Primary and secondary analysis. AFCAPS/TexCAPS, Air Force/Texas Coronary Atherosclerosis Prevention Study; CARDS, Collaborative Atorvastatin Diabetes Study; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell'Insufficienza Cardiaca-Heart Failure; JUPITER, Justification for the Use of Statins in Primary Prevention; MIRACL, Myocardial Ischaemia Reduction with Aggressive Cholesterol Lowering; 4S, Scandinavian Simvastatin Survival Study (Pedersen); PLAC-I, Pravastatin limitation of atherosclerosis in the coronary arteries; PROSPER, the Prospective Study of Pravastatin in the Elderly at Risk

Primary outcomes

Myalgia Four trials, with 14 300 participants, provided reports on myalgia. A total of 551/7190 (0.08%) developed myalgia, compared with 527/7110 (0.07%) in the placebo group. When the data were pooled, there was no difference in risk between statin and placebo (OR 1.03, 95% CI 0.90, 1.17; P = 0.66). No heterogeneity was observed; I² = 0% (see Figure3 and Supplementary Table 2).

Rhabdomyolysis Rhabdomyolysis was rare, affecting 2/8618 participants, with no events (0/8555) occurring in the placebo group. Seven trials reported this outcome. Pooled estimates yielded no increased risk in the statin group (OR 2.93, 95% CI 0.30, 28.18; I² = 0%; P = 0.35) (see Figure3 and Supplementary Table 2).

Myopathy Muscle events were compiled in all eight studies, with a total of 18 845 participants. There were no differences in the proportion of myopathy cases in the treatment group [553/9467 (0.06%)] compared with the placebo group [527/9378 (0.06%)]. Pooling of results revealed no difference in the risk of myopathy between statin therapy and placebo in the elderly (OR 1.03, 95% CI 0.91, 1.18; I² = 0%; P = 0.61) (see Figure3 and Supplementary Table 2).

Secondary outcomes

Mortality due to myopathy-related events No study reported this outcome.

Discontinuation due to adverse effects Three studies (3822 participants) reported this outcome. A total of 92/1939 (0.05%) assigned to the treatment group withdrew owing to an adverse effect compared with 82/1883 (0.04%) in the placebo group. When results were pooled, the OR was 1.08 (95% CI 0.80, 1.46; I² = 79%; P = 0.62). There was evidence of significant heterogeneity, which remained the same (I² = 79%) on reanalysis using the random-effects model. An attempt to explain heterogeneity was carried out by excluding the MIRACL study, which reported a markedly increased risk of myopathy in the statin group, compared with the placebo group; this led to attenuation (OR 0.74, 95% CI 0.50, 1.09; I² = 0%; P = 0.13) of overall effect estimates. None of our preplanned subgroup and sensitivity analyses altered our primary or secondary outcomes (see Figure3 and Supplementary Table 2).

Discussion

The main objective of the present review was to provide reliable estimates of the risks of myopathy in the older person. Only one trial, PROSPER, specifically studied the elderly population; the remaining trials considered the older person as part of their subgroup analyses. The trials assessed in the systematic review showed little or no evidence of a difference in risks between treatment and placebo groups, with regard to myalgia, combined muscle adverse events (myopathy) or rhabdomyolysis. No trial reported an increased rate of death in older people taking statins. We did not show any increased rate of stopping statin medications in older people.

The study had several limitations. Analysis from subgroups should be interpreted with caution. The results from all of the included studies, except PROSPER, were obtained from subgroup data and not all studies specified a priori subgroup analysis. Power is also likely to be limited in a subgroup analysis [42,43] and the risks of obtaining spurious results are higher due to the inability of subgroups to account for multiple hypotheses assumed in the overall trial [42]. Power was further reduced because we did not include every trial that involved older people – for example the Heart Protection Study [44], which included people aged up to 80 years, because the age-specific data were not available to us.

Although RCTs are the ‘gold standard’ and provide the highest evidence of internal validity, controlling for potential confounders and effect modifiers, they are also most unlikely to miss events owing to close follow-up [25]. Despite this, they have the disadvantage of relatively small sizes, so might not be adequately powered to detect a true effect, especially when events are rare, such as rhabdomyolysis [25]. This, coupled with the careful selection (and exclusion) of patients randomized in clinical trials (patients with co-morbidities such as renal and hepatic insufficiencies, and those receiving drugs with potential interactions with statins [45] were excluded), may have led to an underestimation of the effects we observed. Additionally, there is little consensus on muscle adverse effect terminology, potentially resulting in underdiagnosis. This makes it difficult to compare directly the risk of statin myopathy in clinical practice with that in clinical trials [2,46]. It has therefore been suggested that observational studies give a better broad representation of the incidence of adverse effects in patients on statin therapy [47].

It is therefore important that we contrast our findings with those seen in observational studies. In the case of myopathy, a study conducted by Roberts et al. [48], in which 33 550 participants aged >60 years receiving statin therapy were studied, the OR of the risk of developing myopathy was 1.07 (95% CI 1.02, 1.12), again, similar to our estimate (OR 1.03, 95% CI 0.91, 1.18). Their more precise confidence intervals compared to ours is likely due to their larger sample size.

Myalgia was the most commonly reported muscle adverse effect, occurring in 766 per 10 000 persons receiving statins compared with 741 per 10 000 persons receiving placebo. This is in contrast with results obtained from observational studies, which had higher incidences, ranging from about 5% to 15% [46,49,50]. Despite this, the rate of discontinuation due to adverse effects seen in the present review did not differ between treated and placebo groups. However, our results did show heterogeneity, which was attenuated by the removal of the MIRACL study. The MIRACL study showed an exaggerated risk of discontinuation with statins; this study employed a maximum dose (80 mg daily) of atorvastatin in comparison with the CARD [19] and 4S [23] studies, which could account for the trend seen. A meta-analysis by Silva et al. [51], which evaluated the adverse effects of statins and carried out a side-by-side comparison of different statins, found that atorvastatin had the greatest risk of adverse effects when compared with other statins, a finding that is likely to hold true for older people.

When we compared estimates from our review assessing the risk of rhabdomyolysis (OR 2.93, 95% CI 0.30 28.18) against a large (n = 2 873 188) systematic review of observational studies [47], we found our results to be consistent with theirs (OR 2.63, 95% CI 1.50, 1.61). Based on our findings, we would not challenge the well-reported association between statins and rhabdomyolysis but our study demonstrated that, in absolute terms, this remains a very rare event in older people.

Lipophilic statins have a higher propensity than hydrophilic statins to cause myopathy side effects because of their greater solubility in the tissues. We found no evidence of this effect in our population. This contradicts the results found in the literature, which show that hydrophilic statins are associated with a reduced risk of myopathy compared with lipophillic statins [52,53]. The different doses employed in the intervention could possibly account for this difference as some studies used maximal doses, and studies have shown that myopathy is dose dependent [13]. Comparing atorvastatin doses in the CARDS and MIRACL studies (10 mg vs. 80 mg), the risks of myopathy were increased almost threefold in the latter. It is interesting to note that simvastatin, the most lipophilic statin, had the same risks, even when used in low doses, in the 4S trial as high-dose atorvastatin in MIRACL.

In this present study comparing the effect of low/moderate-intensity vs. high-intensity statins on muscle adverse effects in older people showed an increased risk of about 16% with high-intensity statins.

In general, the present review consisted of well-conducted, high-quality trials. However, they were conducted in predominantly male Caucasian patients and were limited to European and American countries. Owing to insufficient data, it was not possible to compare risks in terms of race and gender in the older population, so the generalizability of our findings, in terms of gender and ethnicity, is unclear.

Conclusion

The results obtained from the present review suggest that statins are relatively safe in the older population, in terms of muscle adverse effects. Nothing suggested an increased risk of myopathy in the older adult receiving statin therapy. There was a slightly increased risk of rhabdomyolysis (OR 2.93, 95% CI 0.30, 28.18; I² = 0%; P = 0.33) in this review compared with effects estimates (OR 2.63, 95% CI (1.50, 1.61; I² = 98%) from a recent systematic review of observational studies carried out in the general population [47]), although the event is relatively rare. There is an increased risk of muscle–related adverse events with higher doses of statin therapy in the older population. The present systematic review found the myopathy safety profile of statins to be tolerable in older people. Studies comparing the effect of dose ranges in this population are needed to inform clinical practice. Most of the results obtained in the present review did not provide strong supporting evidence because of the relatively sparse data. Larger reviews are required in the future in order to get more precise estimates of the risk of myopathy in older people

Competing Interests

The authors declare no conflict of interests. All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf and declare no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

Supporting Information

Additional Supporting Information may be found in the online version of this article at the publisher's web-site:

Table S1 Characteristics of excluded studies

Table S2 Summary of Primary, Secondary, Subgroup and Sensitivity analysis

Supporting info item