Objectives There is a lipid paradox in rheumatoid arthritis describing that despite low lipids related to systemic inflammation, there is an increased cardiovascular (CV) risk. Our aim was to evaluate if baseline lipid levels or baseline systemic inflammation were associated with the statin dose sufficient to achieve lipid targets in patients with inflammatory joint diseases.
Methods In this longitudinal, short-term follow-up observational report, we evaluated 197 patients who did and 36 patients who did not reach the recommended low density lipoprotein cholesterol (LDL-c) target. The patients were, after CV risk evaluation, classified to either primary or secondary CV prevention with lipid lowering treatment (LLT). LLT was initiated with statins and adjusted until at least two lipid targets were achieved. Intensive LLT was defined as rosuvastatin ≥20 mg, atorvastatin and simvastatin at the highest dose (80 mg), and conventional LLT were defined as all lower doses.
Results In an independent sample t test, systemic inflammation or lipid levels at baseline were not associated with the statin dose (intensive or conventional) needed to achieve recommended LDL-c target (C reactive protein/erythrocyte sedimentation rate: p=0.10 and p=0.11, and LDL-c/total cholesterol: p=0.17 and p=0.34, respectively). The baseline inflammatory status and lipid levels in patients who did and did not obtain LDL-c goal were comparable (C reactive protein/erythrocyte sedimentation rate: p=0.32 and p=0.64, and LDL-c/total cholesterol: p=0.20 and p=0.83, respectively).
Conclusions Systemic inflammation or lipid levels did not influence the intensity of statin treatment needed to obtain guideline recommended lipid targets in CV prevention. Whether the background inflammation in patients with inflammatory joint diseases over time influences the CV risk reduction related to statins is yet unknown.
Statistics from Altmetric.com
The increased risk of cardiovascular (CV) disease in patients with inflammatory joint diseases (IJDs) (rheumatoid arthritis (RA), ankylosing spondylitis (AS) and psoriatic arthritis (PsA)) is well documented.1–3 Guidelines for CV prevention are available for the general population,4 and recommendations for CV risk management for patients with IJD have been published.5 These suggestions advice that among the preventive measures initiated in patients with IJD, lipid lowering (LL) treatment should follow national guidelines for the general population.
Despite the documented increased CV disease risk and the recent recommendations, there exists limited data on CV disease prevention in patients with IJD. A significant underuse of statins for primary prevention in RA has been reported,6 ,7 and patients with RA seem to receive less preventive treatment after suffering a myocardial infarction than patients without RA.8 However, it has also been shown that patients with and without IJD have comparable effect of statins on CV risk reduction.9 ,10
The experience of primary CV prevention in patients with IJD is also sparse. From a preventive cardio-rheuma clinic we have recently reported that approximately 90% of patients with IJD—those with and without established CV disease—reached lipid goals in less than three consultations.11
In the general population there is a linear relationship between cholesterol levels and risk of CV disease. Patients with RA and AS have lower lipid levels compared with persons without joint disease,9 ,12 ,13 which may be related to inflammation. The impact of systemic inflammation on lipid levels and ratios in patients with RA has been studied in detail and the complex interactions between systemic inflammation and lipids have been reviewed.14–17 A confounding variable is that systemic inflammation (which patients with IJD have) may increase the CV disease risk.18 In relation to these issues, there are several unanswered questions connected to LL treatment in patients with IJD. Furthermore, patients with IJD commonly use several medications for their rheumatic joint disease, with the risk of drug to drug interactions. All these factors may also influence statin treatment in patients with IJD, regarding lipid reduction and the potential high risk of adverse events related to polypharmacy.
The aims of this report were to describe the dose (intensive vs conventional) and type of statin needed to obtain guideline recommended lipid goals for CV prevention in patients with IJD. Second, to explore if baseline inflammation or lipid levels were associated with the statin dose needed to obtain recommended lipid targets. Finally, to describe adverse events related to statins in this high-risk patient group.
Patients and methods
Patients with IJD were referred from a rheumatology outpatient clinic or from general practitioners to the Preventive Cardio-Rheuma clinic at the Department of Rheumatology, Diakonhjemmet Hospital, Norway, for CV risk evaluation during the period from March 2009 to July 2012.
Guideline recommended low density lipoprotein cholesterol (LDL-c) goals are for primary prevention ≤2.5 mmol/L and ≤1.8 mmol/L for secondary prevention.4 Due to the aims of this report, all the patients included in the main analyses obtained recommended LDL-c target. We evaluated whether baseline inflammation or lipid levels were different in patients who did or did not obtain recommended LDL-c targets, to reveal if these parameters had influenced the lipid goal attainment. Baseline describes the time point when the patient first attended the Preventive Cardio-Rheuma clinic.
CV risk stratification was performed and the patient was further classified to either primary or secondary prevention with LL treatment, or to have low risk with no indication for medical intervention. LL treatment was initiated with statins and adjusted until at least two lipid targets were achieved (or at the nearest possible level of the lipid targets), before the patients were discharged for further follow-up by the general practitioner. The type and dose of statin was chosen according to the drug efficacy across doses obtained by the various statins in the STELLAR (the Statin Therapies for Elevated Lipid Levels compared Across doses to Rosuvastatin) trial, which revealed that doubling of any doses of any statins only increases the lipid reduction with 6% (the 6% rule).19 Change of type of statin was done in cases of adverse events or if lipid reduction was inadequate to obtain lipid target. Details concerning CV risk evaluation at the Preventive Cardio-Rheuma clinic have been published recently.11
Statin dose needed to obtain guideline recommended LDL-c targets was grouped into intensive, moderate and low dose defined as follows: Intensive: rosuvastatin 20 mg and 40 mg, atorvastatin 80 mg and simvastatin 80 mg, Moderate: rosuvastatin 5 mg and 10 mg, atorvastatin 20 mg and 40 mg and simvastatin 40 mg, Low: atorvastatin 10 mg, simvastatin 10 mg and 20 mg and all doses of pravastatin. Low dose statin treatment was adequate to obtain lipid targets only in a few patients (n=7). Therefore, to increase the statistical power in the analysis evaluating the impact of baseline lipid levels and inflammation levels on what dose of statin needed to obtain recommended LDL-c target, the groups using low and moderate doses were clustered into a conventinal dose group. Information about adverse event(s) was collected systematically in a standardised way in addition to any change of medications during the observation time.
This is a longitudinal, short-term follow-up observational report. Ethical approval and informed patient consent was therefore not required. The data collection/publication has been recommended and approved by the Oslo University Hospitals’ Office of Privacy and Data Protection (2011/7318).
Recording of CV risk factors
The following risk factors were recorded: Smoking status, diabetes, medications, family history of premature CV disease in first degree relatives (male/female <55 years/<65 years) and presence of established CV disease (myocardial infarction, percutaneous coronary intervention and coronary artery bypass grafting) as well as history of stroke, transient ischaemic attack or peripheral vascular disease.
Blood samples were analysed for lipid moieties (total cholesterol (TC), high density lipoprotein cholesterol (HDL-c) and triglycerides), liver enzymes (alanine transaminase and aspartate transaminase), creatine kinase (CK) and C reactive protein (CRP). All blood variables were analysed routinely at the Diakonhjemmet Hospital laboratory (European standard Accredited 2009) in COBAS 600,20 except LDL-c which was calculated by Friedewald’s formula.21 Erythrocyte sedimentation rate (ESR) was analysed by the Westergren method.
Brachial blood pressure was measured using an Omron M7 after 5 min rest in a supine position. If the blood pressure exceeded 140/90 mm Hg, three consecutive measurements were done and a mean was calculated.
A 12-lead ECG was recorded digitally. B-mode ultrasound of both carotid arteries was performed to evaluate if atherosclerotic plaques were present as previously described.22
A quarter (23.9%) of the patients was already using statins when referred, and the referral reason was in general intolerance or adverse event(s) to one or more statins. Therefore, those who were statin naïve when referred were compared with statin users. The three joint disease groups were also compared to evaluate if there were any differences regarding baseline characteristics and adverse events.
The data are presented as crude data and the results are expressed as number (%) for dichotomised variables, and mean±SD and median and IQR for normally and non-normally distributed characteristics, respectively. In table 1, independent samples t test, analysis of variance and χ2 test were used to compare the data as appropriate. Non-normally distributed variables were log transformed before the analysis of variance was conducted.
To reveal if there was any difference in the dose of statin required to obtain lipid goals across the three IJD groups, χ2 test was used. To evaluate if there were inequalities regarding number of statin changes and statin dose adjustments, in statin naïve patients and patients already using a statin when referred, independent samples t test was used.
Mann-Whitney U test and the independent sample t test were used to compare normally distributed and non-normally distributed variables in the following analyses: (1) To compare the baseline inflammatory status and lipid levels in patients who did and did not obtain LDL-c goal, and (2) For evaluation of the association between baseline lipid levels or systemic inflammation levels and statin dose (conventional vs intensive) needed to obtain recommended LDL-c goals. Data analyses were performed using IBM SPSS V.20.
Baseline patient characteristics
The baseline characteristics are presented in table 1, and the data are compared as follows:
First, we analysed the differences across the three diagnosis groups. There was an age difference across RA/AS/PsA within statin naïve patients and within statin users (p=0.01 and p=0.004, respectively). As expected, the majority of patients with RA was female and the patients with AS were more often male compared with patients with PsA, where the gender distribution was more equal. The patients with AS had significantly longer disease duration, within statin naïve patients and in statin users (p=0.001 and 0.04, respectively). Combined dyslipidaemia (triglycerides ≥2.0 mmol/L and HDL-c ≤1.0 mmol/L) was more often present in patients with PsA compared with patients with RA and AS in those who were statin naïve and in statin users (p=0.01 and p=0.03, respectively). This was reflected in lower HDL-c levels in patients with PsA in both cohorts (p=0.04 and p=0.04, respectively).
Second, we evaluated the differences at baseline between patients who already used statins when referred and the statin naïve patients. Statin users were significantly older than statin naïve patients (p=0.01). The disease duration was also different between the two cohorts (p=0.001). As expected, statin users had lower lipids and consequently less hyperlipidaemia than statin naïve patients. Diabetes, hypertension and established CV disease were more often present in statin users. A higher proportion of statin naïve patients used synthetic and/or biological disease modifying antirheumatic drugs (bDMARDs) compared with statin users (p<0.001, for both).
Use of LL medication in statin naïve patients and statin users
Table 2 describes the course of statin medication until guideline recommended lipid targets were obtained in patients with IJD (RA, AS and PsA) for primary and secondary prevention, in statin naïve patients (n=150) and statin users (n=47). There was no difference regarding the quantity of patients in need of primary and secondary prevention among the three diagnoses (p=0.77, data not shown), and we therefore pooled the IJD patients.
As expected, patients in need of secondary prevention, having a recommended lower LDL-c goal, were in need of higher doses of statins to obtain LDL-c target. In statin naïve patients rosuvastatin was more frequently used in secondary prevention (p<0.001), while the same pattern was seen for atorvastatin in primary prevention (p<0.001). For statin users, the type of LL treatment was more equally distributed.
In patients categorised to be in need of primary preventive treatment, there were no significant differences in type or dose of statin to obtain lipid goals. Of note is that more statin naïve patients were in need of rosuvastatin compared with statin users (p<0.001) and simvastatin was more often applied by patients who were statin users when referred compared with statin naïve patients (p<0.001) in the secondary prevention group. There was a higher need for statin change, number of changes and dose adjustments in statin users compared with statin naïve patients in primary (p=0.03/0.02/0.002, respectively) and secondary prevention (p<0.001 for all). This was, as anticipated, probably related to the referral reason.
Associations of systemic inflammation and lipid levels at baseline with dose of statin needed to obtain lipid goals
Thirty-six patients did not achieve LDL-c target. The baseline inflammatory status and lipid levels in patients who did and did not obtain LDL-c goal were comparable (CRP/ESR: p=0.32 and p=0.64 and LDL-c/TC: p=0.20 and p=0.83, respectively). Separate analyses were performed for patients who did and did not reach the LDL-c goal and who were categorised to be in need of primary (CRP/ESR: p=0.80 and p=0.20) or secondary prevention (CRP/ESR: p=0.26 and p=0.84), since secondary prevention is associated with a more stringent LDL-c goal for which intensive therapy is more often needed regardless of the burden of inflammation. Therefore, only patients who obtained lipid goals (n=197) were included into further analyses in this report.
Systemic inflammation or lipid levels at baseline were not associated with the statin dose (intensive or conventional) needed to achieve recommended LDL-c targets (CRP/ESR: p=0.10 and p=0.11 and LDL-c/TC: p=0.17 and p=0.34, respectively) (figure 1). As a sensitivity analysis, we compared the groups using low, moderate and intensive statin doses, in stead of conventional vs intensive dose, and the main findings were reproduced in that baseline systemic inflammation was not associated with statin dose needed to obtain LDL-c targets (CRP/ESR: p=0.25 and 0.19). Separate analyses stratifying patients on bDMARDs, synthetic DMARDs (sDMARDs), prednisolone and non-steroid anti-inflammatory drugs (NSAIDs) (see online supplementary figure S1), illuminated that there was no significant impact of antirheumatic medication (bDMARDs, sDMARDs, prednisolone and NSAIDs) on the relation between baseline lipid levels and systemic inflammation on statin dose needed to obtain lipid targets.
The side effects were comparable across the different diagnoses, and between statin naïve patients and statin users. The patients with PsA in the statin naïve group experienced somewhat more side effects regarding any adverse event (p=0.07), especially considering side effects classified as others (p=0.01) (table 3). No cases of rhabdomyolysis were observed. Permanent discontinuation of statins was only necessary in two patients. One patient had an increase in the CK and another patient had a rise in the liver enzymes. The increase in liver enzymes and CK was not ≥three times upper normal limit.
Some of the patients were generally intolerant to statins. They had used four to five different statins with intolerable adverse events, which were not related to an increase in liver enzymes or CK, but to gastrointestinal discomfort (nausea, diarrhoea, constipation and gases), muscle ache, loss of memory and fatigue (table 3). In some cases, a ‘start low and go slow’ introduction of statins was adapted. For example 5 mg rosuvastatin twice a week was used as a starting dose, with dose increments after 4 weeks to 5 mg 3 days/week for 4 weeks, and then 5 days/week for 4 weeks, etc, until the max tolerable and required dose was reached. A similar dose-titration regimen was used with atorvastatin, starting on 20 mg twice a week. These patients finally reached recommended lipid goals.
Compliance problems were often related to patients’ firm conviction that statins would give serious adverse events and/or the reluctance to appreciate the importance of LL treatment as a part of CV disease prevention. Despite low lipids, some of these patients had asymptomatic atherosclerosis in the carotid artery, and the cardiologist was unable to convince the patient about the benefit of risk reduction related to a further decrease in lipid levels and therefore the necessity of statin treatment.
Studies and reports concerning statin medication in patients with IJD are scarce. We have a broad clinical experience with CV risk evaluation and LL treatment with statins in the Preventive Cardio-Rheuma clinic at Diakonhjemmet Hospital.11 To our knowledge, this is the first report illuminating that systemic inflammation or lipid levels, before statins are initiated, does not influence the intensity of LL treatment needed to obtain guideline recommended LDL-c goal. The reason(s) why systemic inflammation does not influence the efficacy of statins regarding LL is not clear. Most interestingly, in a population-based incident cohort, patients with RA were less likely to achieve therapeutic goals for LDL-c than the non-RA subjects after 90 days of statin use, which was related to increased ESR at baseline.23 Possible reasons for the divergent results may be the fact that the Olmsted cohort is population-based and that our patients are strictly controlled concerning statin dose increments to achieve targets, and that doses of LL medication needed to obtain lipid goals were systematically recorded in all patients. The patients in the Olmsted cohort had high inflammation, which reflects high disease activity, and thus may result in compliance problems with non-RA medications. In our cohort, the inflammatory levels in patients who did and did not achieve lipid goals were comparable. Hence, the mechanisms behind the response to statins regarding LL in the general population may also be of importance for patients with IJD.24
Statin related adverse events have been reported to be comparable among patients with and without IJD in post hoc analyses of two large statin trials.9 ,10 The results from the present analysis confirm these findings. Statins have an anti-inflammatory effect, which can be related to CRP reduction.25–27 Atorvastatin reduced swollen joint count and significantly improved the Disease Activity Score using 28 joint counts in patients with RA in the Trial of Atorvastatin in Rheumatoid Arthritis study.28 Thus, statins may have an additive anti-inflammatory effect in addition to antirheumatic medication. To reveal if this effect is clinically meaningful, larger studies are needed. In another perspective it is shown that statin induced reduction in CRP, alongside the LDL-c reduction, in the primary prevention JUPITER (Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin trial) study, resulting in a relative hazard reduction >40% of the first major CV event.29 A serious concern related to statins in patients with IJD has been the risk of myopathy due to muscle and joint pain, which is part of their immunological disorder. Statins are in general associated with a limited risk of myopathy. However, the potential drug to drug interactions are known to be increased in high CV risk populations due to the need of multiple medications that share common metabolic pathways.30 Of particular concern is the interaction of statins with other LL drugs, when used in patients with mixed dyslipidaemia such as immunosuppressed patients after organ transplantation, HIV infected patients or patients with rheumatic joint disease using ciclosporin. In the general population the incidence of hospitalised rhabdomyolysis has been reported to be similar and low for monotherapy with atorvastatin, simvastatin and pravastatin.31 Combined statin-fibrate therapy increased risk for rhabdomyolysis, especially in elderly patients and patients with diabetes, and particularly the cerivastatin-fibrate combination which inferred a risk of approximately a tenth of the treated patients. None of our patients used a statin-fibrate combination. Further, we did not observe any sign of rhabdomyolysis and to our knowledge there are no reports on rhabdomyolysis in patients with IJD. Since rhabdomyolysis is extremely rare,32 ,33 the possibility of observing a case with rhabdomyolysis was supposedly low in our rather small population. We therefore recommend clinicians to be alert to the potential drug to drug interactions and adverse events in patients with IJD during long-term statin therapy.
Rosuvastatin was more frequently used in statin naïve patients compared with those who were already using statins when referred. Simvastatin was more often used by patients when referred compared with statin naïve patients. The reason for this may be that general practitioners and cardiologists initiate LL medication with different statins, including that until recently it has been compulsory by the Norwegian Health authorities to initiate LL treatment with simvastatin.
Statin naïve patients more often used sDMARDs and bDMARDs. Patients with high disease activity do not often consult healthcare personnel other than their rheumatologist34 and hence this reflects the low use of statins. Furthermore, the pleiotropic effect of statins, through anti-inflammation, may decrease disease activity and further reduce the need of DMARDs.28 Further studies are warranted to elucidate these pathophysiological mechanisms.
A limitation to our report is that we have a relatively short follow-up time (approximately 3–12 months) and thus whether the background inflammation in patients with IJD over time will influence the CV risk reduction related to statins is yet unknown. This is a single-centre report from a specialist clinic which may have an impact on the adherence to the treating physician's recommendations. With respect to generalisability of the results some weaknesses may be the relatively small number of patients and a diverse population with significant differences in terms of age and sex.
Due to a somewhat heterogenic population in our report, a weak association between systemic inflammation and the efficacy of statins may still be possible, but a large and clinical important difference is unlikely. This is an observational report and obviously not as well controlled as an experimental study. Concerning compliance to LL medication, it was most likely high, bearing in mind that approximately 90% of our patients reached recommended lipid targets. Another limitation is that the LL in our patients with IJD cannot be related to causal effect on future CV events. Information about whether the patients were on stable antirheumatic therapy before referral to the Preventive Cardio-Rheuma clinic is not available, but the patients had overall low CRP values. Even if the patients had changed medication within the time frame of the study, we have shown that there was no significant impact of antirheumatic medication (bDMARDs, sDMARDs, prednisolone and NSAIDs) on the relation between baseline lipid levels and systemic inflammation on statin dose needed to obtain lipid targets.
Controlled clinical trials in the general population have proven that patients at CV risk in the general population benefit from effective LL with statins.29 ,35–37 Meta-analyses of clinical trials have shown that patients benefit from LL treatment regardless of baseline lipid levels.38 Most interestingly, contrary to that in the general population, where low lipids levels infer low CV Risk, lower lipid levels (LDL-c and TC) related to CPR and ESR give a higher CV disease risk in RA.18 Prospective studies are warranted to evaluate if patients with and without IJD have comparable risk reduction of CV disease after treatment with statins over time.
In conclusion, the baseline inflammatory status and cholesterol levels in patients with IJD who did and did not obtain LDL-c goal were comparable. The intensity of statin treatment needed to obtain lipid goals was not related to either the baseline systemic inflammation or lipid levels, suggesting that these factors may be of limited value when developing individual LL CV preventive strategies.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online figure
Handling editor Hans WJ Bijlsma
Contributors All authors have: (1) given substantial contribution to the conception and design and/or analysis and interpretation of the data, (2) drafted and/or revised the manuscript critically for important intellectual content, (3) given final approval of the version to be submitted for publication.
Funding This study has received funding from the South-Eastern Regional Health Authority of Norway.
Competing interests TKK has received speaker's and/or consulting honoraria and/or research grants from AbbVie, BMS, Merck/Schering Plough, Pfizer/Wyeth, Roche, Schering-Plough and UCB. TRP has received speaker’s honoraria and consulting fees from Pfizer and Merck/Schering Plough and speaker's honoraria from AstraZeneca. IH has received speaker's honoraria and consulting fees from Pfizer and Merck/Schering Plough and consulting fees from AstraZeneca and Roche. AGS has received speaker’s honoraria form Merck/Schering Plough, AbbVie, BMS and Wyeth, and speaker’s honoraria and consulting fee from AbbVie and Hoffmann LaRoche/Genentech.
Ethics approval This is a longitudinal, short-term follow-up observational report. Ethical approval and informed patient consent was therefore not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data collection/publication has been recommended and approved by the Oslo University Hospitals’ Office of Privacy and Data Protection (2011/7318).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.