Adherence to statin therapy favours survival of patients with symptomatic peripheral artery disease

Abstract

Aims 

We hypothesized that adherence to statin therapy determines survival in patients with peripheral artery disease (PAD).

Methods and results 

Single-centre longitudinal observational study with 691 symptomatic PAD patients. Mortality was evaluated over a mean follow-up of 50 ± 26 months. We related statin adherence and low-density lipoprotein cholesterol (LDL-C) target attainment to all-cause mortality. Initially, 73% of our PAD patients were on statins. At follow-up, we observed an increase to 81% (P < 0.0001). Statin dosage, normalized to simvastatin 40 mg, increased from 50 to 58 mg/day (P < 0.0001), and was paralleled by a mean decrease of LDL-C from 97 to 82 mg/dL (P < 0.0001). The proportion of patients receiving a high-intensity statin increased over time from 38% to 62% (P < 0.0001). Patients never receiving statins had a significant higher mortality rate (31%) than patients continuously on statins (13%) or having newly received a statin (8%; P < 0.0001). Moreover, patients on intensified statin medication had a low mortality of 9%. Those who terminated statin medication or reduced statin dosage had a higher mortality (34% and 20%, respectively; P < 0.0001). Multivariate analysis showed that adherence to or an increase of the statin dosage (both P = 0.001), as well as a newly prescribed statin therapy (P = 0.004) independently predicted reduced mortality.

Conclusion 

Our data suggest that adherence to statin therapy is associated with reduced mortality in symptomatic PAD patients. A strategy of intensive and sustained statin therapy is recommended.

Statin adherence, Peripheral artery disease, All-cause mortality

Introduction

Statins are the best-documented cardiovascular drugs. There is convincing evidence that they consistently reduce all-cause and cardiovascular mortality as well as morbidity.^1–3

Surprisingly, adherence to guideline-recommended statin therapy is suboptimal,^4–7 it is generally low^8,9 and depends on different risk strata and other patient characteristics.¹⁰ For example, adherence is particularly low in patients with peripheral artery disease (PAD),⁴ even lower than in patients with coronary artery disease (CAD). This is a major concern in the face of the extremely high risk of PAD patients.¹¹

Few systematic data exist on the prognostic value of statin adherence,¹² and the correlation between adherence and cardiovascular outcome has been addressed only in a few studies.^6,13–16 Peripheral artery disease is an ideal population to study the outcome in relation to the adherence to statin therapy, but here data are not available at all. In particular, a dose-response relationship between adherence and outcome has not been reported from real -world data.

We tested the hypothesis, that adherence to statin therapy is associated to survival in patients with PAD. We analysed baseline and follow-up data of statin dosage, of LDL-C levels achieved, and of survival in a cohort of 691 patients with symptomatic PAD.

Methods

Study design

In this longitudinal, observational, single -centre study, we investigated lipid levels at presentation as well as adherence to and intensity of statin therapy and lipid changes over time. These parameters were related to survival in patients with symptomatic PAD. Recruitment of 691 consecutive patients was performed in the outpatient clinic of our hospital. These patients were referred to the Swiss Cardiovascular Center between January 2010 and December 2017 for endovascular treatment of their symptomatic PAD. All baseline information and initial laboratory results were collected. All this data was then entered in the institutions data bank for patients undergoing endovascular treatment. A follow-up data was equally collected and entered in the databank.

The database and participant informed consent form on research involving humans were approved by the Swiss Ethics Committee. All authors had full access to all the data in the study and take responsibility for its integrity and the data analysis.

Cardiovascular risk factors and comorbidities were recorded at baseline. Hypertension was defined according to the 2013 ESC/ESH guidelines,^17,18 Type 2 diabetes was diagnosed according to current American Diabetes Association (ADA) clinical practice recommendations,¹⁸ and hyperlipidaemia was defined according to the 2016 ESC/EAS guidelines.¹ Smoking status was determined in regard to being active smoker, having stopped >6 months before or having never smoked.

Lipid target levels were defined according to the ESC and NLA guidelines with a recommended goal for LDL-C of <70 mg/dL.^1,2 Non-high-density lipoprotein cholesterol (HDL-C) was calculated as total cholesterol minus HDL-C. The recommended goal for non-HDL-C was <100 mg/dL.^1,2

Intensity of statin treatment was categorized as high, moderate, and low. Based on the ACC guidelines,¹⁹ treatment with atorvastatin 40–80 mg and rosuvastatin 20–40 mg was categorized as high intensity; atorvastatin 10–20 mg, fluvastatin 40 mg twice daily, pravastatin 40–80 mg, rosuvastatin 5–10 mg and simvastatin 20–40 mg as moderate intensity, and fluvastatin 20–40 mg, pravastatin 10–20 mg and simvastatin 10 mg as low intensity. Statin potency was normalized to simvastatin to allow comparison between compounds.¹⁰

Inclusion criteria

Consecutive patients with chronic, symptomatic PAD (Fontaine Stage II, III, or IV) due to atherosclerotic disease referred for primary endovascular lower limb revascularization. Written informed consent was obtained.

Exclusion criteria

Patients with non-atherosclerotic arterial obstructive disease, endovascular redo procedures (restenosis) at the time of inclusion, acute or embolic PAD, known vasculitis or other than atherosclerotic PAD. We excluded patients below the age of 40 with high probability of non-atherosclerotic PAD; and those with documented statin intolerance, i.e. myopathy, or clear contraindications for a statin therapy, as well as patients with thyroid disorders.

Data collection

Consecutive database

The serum lipid profiles of the patients were obtained after a 12-h overnight fast after hospital admission, and for a second lipid reading at an interval of 50 ± 26 months in an outpatient setting.

Statin type and dosage were recorded and data were entered into the department’s database on PAD patients. Information on all-cause mortality was obtained from the hospital’s medical information system (iPDOS).

Statistical analysis

All analyses were performed using the GraphPad Prism© statistical software package, version 7.0c (GraphPad^®, San Diego, CA, USA) and SPSS, Version 25 (SPSS Inc., Chicago, IL, USA). Categorical data are presented as absolute numbers and percentages. The distributions of metrical variables are presented as means with standard deviations (SDs). The frequencies of categorical variables for two and three comparison groups were compared by the Fisher’s exact test or the χ² test, as appropriate. Continuous variables were compared by Mann–Whitney–Wilcoxon test for two comparison groups and by Kruskal–Wallis test for three-group comparison. Kaplan–Meier survival analysis was used to estimate all-cause mortality rates using the log-rank test, being reported with the 95% confidence interval . P-values <0.05 were considered significant. All P-values are results of two-sided tests.

Univariate analysis was performed including patient-related variables (Table 1). To investigate the effect of these covariates (age, gender, smoking history, diabetes mellitus, hypertension, hyperlipoproteinaemia, family history for premature atherosclerotic diseases, and renal function), multivariate logistic regression analysis was conducted to test for influence of statin adherence on all-cause mortality. The significance level for entry of independent variables from the univariate model into the multivariate model was 0.1. The explanatory covariates included age, diabetes, family history, adherence, or change in statin medication (newly prescribed statin, increase or decrease of statin dosage, and discontinuation of statins) as well as statin intensity category (high, moderate, or low).

Table 1

Patient characteristics at baseline

Demographics	Patients (n = 691)
Age (years), mean ± SD	72.53 ± 12.05
<55, n (%)	62 (9)
55–75, n (%)	349 (51)
>75, n (%)	280 (40)
Female gender, n (%)	233 (34)
Body mass index (BMI) (kg/m2)	26.4 ± 4.5
CVD family history, n (%)	104 (15)
Current smoker, n (%)	440 (64)
Diabetes, n (%)	214 (31)
Hypertension, n (%)	588 (85)
Hyperlipoproteinaemia, n (%)	522 (76)
Chronic kidney disease (CKD)	478 (69)
Stage 2	254 (53)
Stage 3a	116 (24)
Stage 3b	74 (15)
Stage 4	27 (6)
Stage 5	7 (2)
Creatinin (µmol/L)	97.51 ± 68.10
Glomerular filtration rate (eGFR) (mL/min)	69.78 ± 22.27
CAD, n (%)	297 (43)
CVD, n (%)	97 (14)
Polyvascular (+CAD + CVD), n (%)	59 (9)
Severity of PAD (Fontaine stage)
Intermittent claudication
II, n (%)	401 (58)
Critical limb ischaemia
III, n (%)	87 (13)
IV, n (%)	203 (29)

We used the classic multivariate logistic regression model. Although time of death was known, the exact time of therapy change was assessed only at the end of the study, which did not allow for exact Cox regression analysis. We sought to reveal the influence of covariates on death itself but not on time to death. Furthermore, the covariates themselves were also determined only once and not on a time axis.

Results

Patients’ characteristics

Demographic data of patients are summarized in Table 1. Prevalence of cardiovascular risk factors in the study population was high, with nearly two-thirds being active smokers; and around one-third had diabetes mellitus. Hypertension was the most common risk factor in the population with 588 individuals (85%). The majority (58%) of the patients had intermittend claudication (Fontaine Stage II), and 42% suffered from critical limb ischaemia (Fontaine Stages III and IV) with one-third of the later already having peripheral ulcerations at the time of referral.

Statin therapy

Table 2 summarizes the statin therapy. The baseline data reflect outside providers' prescribing habits of statins. The most frequently prescribed statins were atorvastatin (46.4%) or rosuvastatin (26%). Initially, 73% of PAD patients were on statins. At follow-up, we observed a significant increase in statin use to 81% (P < 0.01). Statin dosage, normalized to simvastatin 40 mg, increased from a mean (± SD) of 50 ± 31 to 58 ± 34 mg/day (P < 0.0001).

Table 2

Distribution of statins and lipid levels at baseline and follow-up

Statin	Baseline	Follow-up	P-value
	N = 511 (74.0)	N = 560 (81)	<0.01
Atorvastatin, n (%)	237 (46.4)	278 (49.6)	0.39
Rosuvastatin, n (%)	133 (26.0)	175 (31.3)	0.08
Simvastatin, n (%)	91 (17.8)	70 (12.5)	0.01
Pravastatin, n (%)	44 (8.6)	34 (6.0)	0.13
Fluvastatin, n (%)	6 (1.2)	2 (0.4)	0.16
Pitavastatin, n (%)	—	1 (0.2)	>0.99
Statins prescribe in
PAD only, n (%)	218 (61)	277 (78)	<0.0001
+ CAD, n (%)	209 (88)	200 (84)	0.29
+ CVD, n (%)	29 (76)	34 (89)	0.22
+ Polyvascular, n (%)	49 (83)	45 (76)	0.49
Lipid levels (mean ± SD)	Baseline (N = 691)	Follow-up (N = 691)	P-value
Cholesterol (mg/dL)	171.8 ± 51.4	157.9 ± 44.0	0.01
HDL-cholesterol (mg/dL)	48.7 ± 16.6	49.0 ± 17.4	0.58
Non-HDL-cholesterol (mg/dL)	120.5 ± 40.9	106.2 ± 39.0	<0.0001
LDL-cholesterol (mg/dL)	97.3 ± 44.0	82.2 ± 36.7	<0.0001
Triglycerides (mg/dL)	169.0 ± 138.1	160.2 ± 123.9	0.18

All bold values are significant.

Concerning the atherosclerotic extent, we observed a higher non-prescription rate in patients with PAD alone (39%) compared to patients with a concomitant atherosclerotic disease [CAD (12%), cardiovascular disease (CVD ; 24%), and polyvascular disease (17 %)]. Over the follow-up period, a significant drop in non-prescription of statins was documented only for isolated PAD patients (22%; P < 0.0001; see Table 2).

Among the patients who discontinued their statin medication, 57% had been on Atorvastatin, 34% on Rosuvastatin, 8% on Simvastatin, and 9% on Pravastatin, before stopping medication. This pattern is almost indistinguishable from one of adherent patients. Thus, the compound used is apparently not related to discontinuation.

The proportion of patients receiving a high-intensity statin medication increased similarly over time from 23% to 38% (P < 0.0001), whereas lower intensities decreased, as well as the proportion of patients not on statins (P < 0.01) (Figure 1A). Notably, statin dosage in the high-intensity segment increased from 40 ± 16 mg (normalized to simvastatin 40 mg) to 87 ± 25 mg.

Figure 1

(A) Intensity of statin treatment at baseline and follow-up. (B) Absolute decrease of low-density lipoprotein cholesterol (in m/dL) from baseline to follow-up (Δ) according to statin intensity.

We observed that the intensification of statin therapy was paralleled by a decrease of total cholesterol, LDL-C, and non-HDL-C (Table 2). As expected, absolute LDL-C decrease over time (Δ) was highest in patients with high-intensity statin treatment, compared to moderate or low intensity or non-statin treatment (P < 0.05) (Figure 1B).

Low-density lipoprotein cholesterol level attainment and survival data

Low-density lipoprotein cholesterol was reduced significantly in all statin treatment groups, except in the group with the decreased dosage. Being never on statins or discontinuation of statins did not change LDL-C levels similarly (Figure 2A).

Figure 2

(A) Achieved mean low-density lipoprotein cholesterol reductions depending on treatment incentives. (B) Kaplan–Meier survival curves of patients on and off statins. (C) Kaplan–Meier survival curves of patients with a switch in statin dosage. New statin: significance level (P <0.0001); no change dosage: significance level (P <0.0001); increase dosage: significance level (P<0.01).

Regarding the outcome, we observed that patients being never on statins had a higher mortality rate (34%) compared to patients having been continuously on statins at the same dosage (13%) or having newly received a statin (8%) (Figure 2B; P < 0.0001).

Additionally, we found that patients with an intensified statin medication had a favourable outcome (9% mortality) in contrast to patients who terminated statin medication (34% mortality) or reduced the statin dosage (20% mortality) (Figure 2C; P < 0.05).

Patients’ characteristic regarding statin adherence

Next, we analysed the patterns of baseline risk factors separately for patients on statin therapy, those with statin discontinuation and those never on statins. As can be seen from Table 3, patients on statins were significantly younger, but classical risk factors, i.e. smoking, diabetes, hypertension, and hyperlipoproteinaemia, were significantly more common.

Table 3

Patient characteristics regarding statin adherence from baseline to follow-up

Demographics	Statin therapy (n = 556)	Statin discontinuation (n = 65)	Never statins (n = 70)	P-value
Age (years), mean ± SD	70.1 ± 10.8	72.5 ± 12.0	73.7 ± 14.5	<0.01
<55, n (%)	45 (8)	7 (11)	10 (14)	0.20
55–75, n (%)	305 (55)	25 (38)	19 (27)	<0.0001
>75, n (%)	206 (37)	33 (51)	41 (59)	<0.001
Body mass index (BMI) (kg/m2)	26.5 ± 4.5	26.8 ± 4.7	25.4 ± 4.3	0.29
Female gender, n (%)	178 (32)	23 (35)	32 (46)	0.14
CVD family history, n (%)	91 (16)	6 (9)	7 (10)	0.14
Current smoker, n (%)	282 (51)	30 (46)	24 (34)	0.03
Diabetes, n (%)	25 (34)	16 (25)	11 (16)	<0.01
Hypertension, n (%)	483 (87)	51 (78)	54 (77)	0.03
Hyperlipoprotaeinaemia, n (%)	435 (78)	49 (75)	38 (54)	<0.001
Chronic kidney disease (CKD), n (%)	386 (69)	41 (63)	51 (73)	0.45
Stage 2, n (%)	210 (54)	23 (45)	21 (51)	0.44
Stage 3a, n (%)	91 (24)	10 (24)	15 (29)	0.66
Stage 3b, n (%)	60 (16)	7 (17)	7 (14)	0.90
Stage 4, n (%)	20 (5)	3 (7)	4 (8)	0.66
Stage 5, n (%)	5 (2)	0 (0)	2 (2)	0.24
Creatinin (µmol/ L)	96.8 ± 63.4	88.1 ± 36.9	114.3 ± 133.7	0.32
Glomerular filtration rate (eGFR) (mL/min)	70.3 ± 22.1	71.3 ± 21.7	65.3 ± 24.6	0.34
CAD, n (%)	200 (36)	23 (35)	15 (21)	0.05
CVD, n (%)	35 (6)	1 (2)	3 (4)	0.25
Polyvascular (+CAD + CVD), n (%)	45 (8)	8 (12)	6 (9)	0.49
Severity of PAD (Fontaine stage)
Intermittent claudication
II, n (%)	334 (60)	33 (51)	34 (49)	0.08
Critical limb ischaemia
III, n (%)	69 (12)	7 (11)	11 (16)	0.66
IV, n (%)	153 (28)	25 (38)	25 (36)	0.09

All bold values are significant.

As can be seen in Figure 3, in the age category > 75 years, patients on statins showed a lower all-cause mortality in contrast to those discontinuing their statin medication or being never on statins (P < 0.02).

Figure 3

All-cause mortality regarding patients on statins, with statin discontinuation and for patients never on statins in the age category >75 years.

Multivariate logistic regression analysis of statin adherence

To test whether the observed effects on all-cause mortality were independent of confounders, we performed a multivariate logistic regression analysis. From a univariate analysis with classical risk factors, we identified age, diabetes mellitus, and family history of CVD as potential confounders (all P < 0.01), but not gender, hypertension, or smoking. The multivariate analysis then revealed that diabetes and family history (P < 0.001 and P < 0.01, respectively), but not age or kidney function, had an influence on mortality. Next to that, the analysis confirmed newly prescribed statins (P = 0.004), as well as long-term adherence to statins (P = 0.001), including an increase of dosage (P = 0.001), to be independent predictors for a lower all-cause mortality.

Cardiovascular vs. non-cardiovascular mortality

Of the patients not on statins (n = 135) during the observation period [either having never received a statin (n = 70) or having discontinued their statin medication (n = 65)], 43 (32%) patients died during the observational period: 18 (13%) due to cardiovascular and 25 (19%) due to non-cardiovascular causes. Among patients on statins [either continuously on the same dosage or with an increased or decreased dosage (n = 556)], 66 (12%) died during the observational period: 37 (7%) due to cardiovascular and 29 (5%) from non-cardiovascular causes. Thus, relatively more patients without statins died, both from cardiovascular and from non-cardiovascular causes. However, baseline demographics for all three subgroups (Table 3) showed that patients on statins had a higher baseline risk profile with regard to classic risk factors. In contrast, patients not on statins or discontinuing their treatment were older but not necessarily sicker.

Discussion

The main finding of this longitudinal observational study is that statin adherence improves survival in PAD. Our data close an important gap in knowledge in preventive cardiovascular therapy. We present the first European study linking statin adherence to all-cause mortality in PAD patients. Most importantly, there is a significant dose-response relationship between parameters of adherence and survival.

Some of our data deserve a more detailed discussion. High-intensity statin use along our observation period has increased from 23% to 38%, a rate becoming almost equal to that of moderate-intensity treatment. In patients switching from moderate-intensity to high-intensity statin treatment, LDL-C was reduced by 21 mg/dL from baseline. As derived from the Cholesterol Treatment Trialists formula,²⁰ a decrease in endpoints of around 11% could be expected. Although not directly comparable, survival in our patients never on statins was about 18% lower (69%) than in patients on statins (87%). Moreover, statin discontinuation decreased survival from 87% to 66%. From these findings, we can conclude that there is a robust relation between statin intensity, LDL-C reduction, and all-cause mortality.

An interesting question arises regarding the contribution of cardiovascular and non-cardiovascular mortality to the all-cause mortality described above. Indeed, both contributed to the reduction of all-cause mortality in the statin-treated population. For a long time, it was not clear, that lipid -lowering reduces non-cardiovascular mortality. Very recently, however, strong evidence for a reduction of non-cardiovascular mortality was established by the ODYSSEY outcomes trial.²¹

Our data are supported by a recent population-based study in US Veterans,²² where Arya et al.²² found, that especially in PAD patients under a high-intensity statin regime at the time of diagnosis, mortality and limb amputation were significantly reduced. Similarly, to the findings by Arya et al., more than a quarter of our patients were not on statins (29%). Likewise, in our cohort, patients with PAD alone presented with a similar low prescription rate (61%) compared to the study by Arya (58%). The overall mortality rate in our population was lower (16%). However, our data on intensified statin therapy revealed a similar outcome with a mortality reduction of 22–25% compared to patients never on statins or discontinuing their statin medication. Over and beyond the study by Arya, our focus on adherence and intensity in statin therapy thus adds further detailed longitudinal information on target level attainment and outcome in Europe.

In the recent years, there was an increasing interest in statin adherence and its impact on patients’ outcome. Most data where obtained from patients with CAD. It is known that adherence to statins after not > 6 months has come down to only 70%.²³ Paradoxically, adherence is lower in patients at higher risk, e.g. diabetes.^24,25 In PAD, a similarly low adherence rate of 69% after 3 years was reported.²⁶ Indeed, higher adherence to statin among other cardioprotective therapy was associated with decreased all-cause and cardiovascular mortality as well as revascularization procedures and associated hospitalization in CAD patients.²⁷ Moreover, patients with statin intolerance, a major risk factor for non-adherence,²⁸ are more likely to encounter recurrent cardiovascular events, like myocardial infarction (MI), and show a higher all-cause mortality, than patients with high statin adherence.²⁹ In a study on critical limb ischaemia, patients discharged on guideline-recommended statin intensity had a lower mortality and lower major adverse limb event rates than those on moderate statin intensity regime. Yet, no information on LDL-C achievement, dosage range, or dose-response relation was given.³⁰

A program in Canada investigated a multifactorial risk intervention in PAD including, among others, antiplatelet agents, statins, angiotensin-converting enzyme inhibitors, blood pressure control, and smoking cessation. This combined intervention improved outcome, but no information on the specific effect of statins per se and of adherence was reported.³¹ Similarly, Armstrong et al.³² reported beneficial effects of combined risk factor interventions in patients with claudication or critical limb ischaemia. Again, to the best of our knowledge, no data on statin therapy or adherence in this setting were available. Taken together, our results close the gap in knowledge of the interaction of statin adherence, LDL-C lowering, and survival in symptomatic PAD patients. This is also supported by data from primary prevention studies pointing to a lower incidence of cardiovascular events associated with better adherence.^33,34

Ascertainment of endpoints is an important issue in prospective observational studies. Because all-cause mortality—and not amputation—is the major consequence of PAD, we chose all-cause mortality as our single combined endpoint. The black-or-white recording of survival obviates the confounders in cause-specific assessment of death in patients with PAD. Only recently, the ODYSSEY OUTCOMES^21,35 and EMPA-REG Outcome³⁶ trials have underlined that all-cause mortality is of overwhelming importance in a risk intervention, it is considered the preferred primary endpoint.³⁷ A priori, in a PAD population like ours, fatal stroke and MI are relatively rare events in comparison to total mortality. We have, therefore, looked for cardiovascular mortality including MI and stroke, but decided for all-cause mortality as the primary endpoint for the aim of statistical power as well as an unequivocal endpoint (see review in the EHJ-CVP³⁷).

The practical consequences of our data are evident. A proper guideline-recommended statin treatment saves lives of patients with symptomatic PAD. These data should encourage the entire medical community to intensify efforts for sustained high-intensity statin therapy.

Strength and limitations

The strengths of our investigation are a large representative study cohort of a clinically important population with extremely high cardiovascular risk as well as a thorough documentation of therapeutic doses, LDL-C levels, and consequent adaption of treatment over a long observational period.

The limitations are evident. Our study reports a single -centre data set and is of observational nature and open for bias. Our study protocol did not systematically cover reasons of non-adherence. Furthermore, we did not include an adherent and non-adherent control group without PAD. Nevertheless, we found chronic kidney disease not to be associated with non-prescription or non-adherence, whereas age was only significantly associated with discontinuation.

Conclusion

Adherence to statins is crucial for survival in high-risk PAD patients. It is worthwhile to install an intensified statin treatment but intermediate adherence is better than none, and even de novo statin therapy is beneficial (‘it is never too late ’). Decrease of dosage is as deleterious as discontinuation or being never on statins while suffering from symptomatic PAD. Systematic programmes are necessary to improve statin adherence in patients with PAD.

Supplementary material

Supplementary material is available at European Heart Journal Cardiovascular Pharmacotherapy online.

Conflict of interest: none declared.