Background Patients with rheumatoid arthritis (RA) are at increased risk of heart failure and vascular events. Small increases in circulating N-terminal pro-brain natriuretic peptide (NT-proBNP) are associated with an increased risk of a cardiovascular event, and high levels signal left ventricular dysfunction. Data on the effects of tumour necrosis factor α(TNFα) blocking agents on circulating NT-proBNP levels in patients with active RA are lacking but may be informative.
Methods 171 consecutive patients with RA (28-joint disease activity score >3.2) without congestive heart failure (NYHA class III or IV) were scheduled to receive adalimumab once every 2 weeks. Serum NT-proBNP concentrations were measured simultaneously on stored baseline and 16-week samples. Paired sample t tests were used to observe differences in biomarkers before and after adalimumab administration. Correlations between the biomarkers and changes in circulating log NT-proBNP levels were evaluated with the Pearson test and multivariable linear regression analyses of correlates were performed (forward selection procedure).
Results Circulating levels of NT-proBNP decreased significantly after 16 weeks of adalimumab administration (median NT-proBNP 83.0 pg/ml vs 69.5 pg/ml, p=0.004). Changes in NT-proBNP levels were associated with changes in pulse pressure (r=0.18, p=0.02), systolic blood pressure (r=0.16, p=0.04) and erythrocyte sedimentation rate (r=0.18, p=0.02). On multivariable analysis, changes in pulse pressure and erythrocyte sedimentation rate remained independently associated with changes in circulating NT-proBNP levels.
Conclusions These observations show that blocking TNFα in patients with RA without evident heart failure decreases NT-proBNP levels by about 18%. This suggests no treatment-induced deterioration in cardiac function and a potential cardiovascular risk benefit.
Statistics from Altmetric.com
Rheumatoid arthritis (RA) is characterised by excess cardiovascular (CV) mortality and morbidity.1 CV involvement in RA covers a wide spectrum from myocardial ischaemia to clinically evident heart failure (HF).2,–,4 Inflammation is a key feature of RA which may in part lead to the pathogenesis of atherosclerotic disease and HF.5 6 Observational studies demonstrating that tumour necrosis factor α (TNFα) blockade or methotrexate therapy is associated with a lower risk of atherosclerotic disease in patients with RA are commensurate with the inflammatory hypothesis of cardiovascular disease (CVD).7,–,9 However, CV end-point trials of such agents in RA are lacking, and the interpretation of the effect of the highly effective anti-inflammatory TNFα blocking agents on the risk of CV in patients with RA is complex. Furthermore, trials with TNFα blocking agents against HF in non-RA patients were halted because of a higher combined risk of death from any cause or hospitalisation for HF in patients with moderate to advanced HF.10 11 Further data are therefore required to extend our understanding of the effect of TNFα blockade on cardiac function and CV risk in patients with RA without HF.
N-terminal pro-brain natriuretic peptide (NT-proBNP) is a cardiac neurohormone predominantly released from cardiomyocytes in response to left ventricular volume expansion and pressure overload.12 Much interest has recently focused on NT-proBNP because of its function as a clinically relevant biomarker for HF.13 14 Moreover, NT-proBNP is independently associated with increased CV risk in patients with and without pre-existing CVD,15 although the potential use of NT-proBNP in risk scores remains uncertain.16 It has recently been shown that NT-proBNP levels are increased in patients with RA, and that they are associated with inflammation and disease activity.17,–,19 Moreover, there is evidence that RA itself is associated with an increased risk for CVD and HF.2,–,4 Based on these observations, one may expect that dampening of inflammation in patients with RA would lower circulating NT-proBNP levels but, to date, no studies have addressed this intriguing possibility. We therefore investigated the short-term effects of TNFα blocking therapy (ie, adalimumab) on the NT-proBNP levels in a large group of patients with RA before and after 16 weeks of treatment. From these data we aimed to identify (1) whether there was any evidence of deterioration in ventricular function (increased NT-proBNP) upon adalimumab administration in patients with RA without moderate to advanced HF, and (2) whether dampening inflammation with adalimumab was associated with reductions in NT-proBNP levels (and thus in keeping with suggestions of a lower CVD risk with such agents) in patients with RA.
Patients and methods
For the present study we used a biobank accrued from a prospective observational study cohort. The cohort comprised 171 consecutive patients with RA, all of whom weretreated with adalimumab at the rheumatology outpatient clinic of the Jan van Breemen Institute, Amsterdam. Patients were either treated with adalimumab and concomitant disease-modifying antirheumatic drugs (DMARDs) or with adalimumab alone. All patients used adalimumab 40 mg subcutaneously every 2 weeks. In patients with an inadequate response as judged by the treating rheumatologist, the dosing frequency of adalimumab could be increased to 40 mg per week. Serum samples were collected before the first injection with adalimumab at baseline and after 16 weeks of treatment.
All patients fulfilled the American College of Rheumatology 1987 revised criteria for RA and had active disease, indicated by a disease activity score in 28 joints (DAS28) of more than 3.2 despite earlier treatment with two DMARDs including methotrexate at a dose of 25 mg per week or at the maximal tolerable dose, according to the Dutch consensus statement on the initiation and continuation of TNFα blocking therapy in RA.20 In this respect, it should be emphasised that patients with RA with known severe congestive heart failure (NYHA class III or IV) were withdrawn.
All patients with RA were seen by a research physician and completed a questionnaire recording demographic data, medical and medication history. A physical examination was performed also including the DAS28.21 Clinical response was assessed by the European League Against Rheumatism (EULAR) criteria.22 Blood samples were taken to measure creatinine, inflammatory parameters (ie, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP)) and the serological markers IgM-rheumatoid factor (IgM-RF) and anti-cyclic citrullinated peptide antibodies as described previously.23
Cardiovascular risk factors
Cardiovascular risk factors assessed during physical examination were blood pressure and body mass index. Height and weight were measured barefoot wearing light clothes only. The body mass index (BMI) was calculated as the ratio of weight to squared height. Systolic and diastolic blood pressure was obtained on the right arm with the subject in sitting position. Renal function was assessed by using the Cockcroft–Gault formula in ml/min: ((140−age) × body weight/creatinine × 71) × 0.85 for women). The formula is given in traditional units; to convert it to ISU (mg/dl), creatinine is multiplied by 88.4. Patients were classified as smokers or non-smokers. Information about diabetes and CVD was extracted from medical records. In brief, coronary diseases were defined as having a history of myocardial infarction or a percutaneous transluminal coronary angioplasty or surgery for ischaemic heart disease, and cerebral arterial disease as having a history of a stroke or a transient ischaemic attack or a carotid endarterectomy.
Serum NT-proBNP concentrations were determined in serum stored at −20°C for a median of 3.6 years using the Elecsys 2010 electrochemiluminescence method (Roche Diagnostics, Burgess Hill, UK). Samples were snap-thawed at 37°C and assayed on the analyser, which was calibrated using the manufacturer's reagents. Manufacturer's controls were used to monitor assay drift, using both a high and low control with limits of acceptability defined by the manufacturer. Low control coefficient of variation (CV) was 6.7% and high control CV was 4.9%. Collected serum samples were analysed in a blind fashion on the same day in a single run with all samples from each individual grouped together to overcome interassay variations.
Data are expressed as mean (SD) or median (interquartile range) as appropriate. As NT-proBNP, CRP and ESR levels were not normally distributed, data were analysed with and presented as the (natural) logarithms of these values in all analyses. Pearson correlation coefficients were calculated to evaluate correlations between NT-proBNP, RA and CV variables at baseline. Paired sample t tests were used to observe differences before and after adalimumab administration. Multivariable linear regression analyses were performed according to a forward selection procedure, introducing those variables that showed a statistically significant association with the outcome measure log NT-proBNP (p≤0.10). The following variables were introduced: age, gender, DAS28, CRP, Cockroft–Gault, BMI, pulse pressure, smoking, antihypertensives, statins and aspirin. Similar analyses were performed evaluating the association of changes in circulating log levels of NT-proBNP with changes in RA and CV variables. Calculations were made using SPSS 16.0 software and the threshold for significance was set at p<0.05.
Baseline characteristics of patients with RA
Table 1 shows the baseline clinical characteristics of patients evaluated in this prospective cohort study. Approximately three-quarters (79%) of the patients were women. Most patients were IgM-RF positive (74%) and already had erosions on the radiographs (74%). Median RA disease duration was 8 years and disease activity was high with a mean DAS28 of >5. Nodules were present in approximately 30% of the study population. With regard to antirheumatic treatment, methotrexate was most frequently used (78%), followed by prednisone (38%), sulfasalazine (15%) and hydroxycholoroquine (7%). Patients had a mean BMI of about 25 kg/m2, a quarter of all the patients smoked, 7% had diabetes and 7% reported a previous CV event. Finally, 27% of all patients used antihypertensives, 11% statins and 7% aspirin.
Changes in CV- and RA-related risk factors
During the study period eight of the included patients commenced treatment with antihypertensives, one with aspirin and none with statins or glucose-lowering therapy. In addition, antirheumatic treatment remained unchanged during 16 weeks of adalimumab administration. Treatment with adalimumab led to a significantly improved DAS28 score (and its individual components) at 16 weeks compared with baseline (table 2). Blood pressure and renal function did not change with adalimumab treatment.
NT-proBNP and other CV risk factors in patients with RA at baseline
NT-proBNP correlated with age (r=0.38, p<0.01), female gender (r=0.23, p<0.01), systolic blood pressure (r=0.20, p=0.01), pulse pressure (r=0.23, p<0.01), Cockroft–Gault (r=−0.28, p<0.01), CRP (r=0.22, p<0.01) and ESR (r=0.22, p<0.01). Circulating levels of NT-proBNP levels were higher in patients with RA treated with cardioprotective treatment, including antihypertensives (p=0.04) and aspirin (p=0.03), and in patients with RA with previous CVD, although the latter did not attain formal statistical significance. At multivariate analyses, only age, gender (higher in women) and CRP remained independently associated with circulating levels of NT-proBNP.
Change in NT-proBNP levels
Circulating levels of NT-proBNP declined (p=0.004) in patients with RA after 16 weeks of treatment (figure 1). This translated to a relative change in median circulating NT-proBNP levels of −17.6% (IQR −35.7% to 26.0%) pg/ml. Overall, NT-proBNP decreased in 64% of patients and increased in 36%. The change in circulating NT-proBNP levels upon adalimumab administration did not differ between EULAR responders (n=161; change in NT-proBNP levels −16% (IQR −35% to 29%)) and EULAR non-responders (n=10; change in NT-pro-BNP levels −20% (IQR −41% to 3%)) (p for difference=0.93), although this may be due to small numbers of non-responders. NT-proBNP levels decreased more in patients with NT-proBNP levels >100 pg/ml at baseline (n=71) than in those below this level (p for difference <0.01).
Associations of changes in NT-proBNP with CV risk factors
Changes in circulating levels of NT-proBNP after 16 weeks of treatment with adalimumab were significantly associated with changes in pulse pressure (r=0.18, p=0.02; figure 2A), systolic blood pressure (r=0.16, p=0.04) and ESR (r=0.18, p=0.02; figure 2B). There was some evidence for a weak association between changes in circulating levels of NT-proBNP and CRP (r=0.14, p=0.07). Following multivariable linear regression analysis, only pulse pressure (p=0.018) and ESR (p=0.036) remained independently associated with changes in circulating levels of NT-proBNP. Finally, all relevant associations remained when patients newly prescribed antihypertensives (n=8) or aspirin (n=1) after adalimumab administration were excluded from the analyses.
To our knowledge, this is the first study to evaluate the effect of TNFα blocking agents on NT-proBNP in patients with active RA. We observed a reduction of approximately 18% in NT-proBNP levels after 16 weeks of treatment with adalimumab (p=0.004). Interestingly, this decrease may be more pronounced in those with baseline NT-proBNP levels >100 pg/ml, although this observation requires future corroboration. Given that NT-proBNP is strongly linked to cardiac function, our findings go against the possibility that TNFα blockade worsens ventricular function in patients with RA without prevalent HF, a clinically important observation given the adverse history of TNFα blocking studies in patients with HF. Moreover, small differences in NT-proBNP levels in healthy people, very similar to the low circulating concentrations reported here, have been shown to predict CV events,24 so our findings support epidemiological findings indicating that TNFα blockade may lessen CV risk in patients with RA.7 Importantly, the present study does not provide any evidence that TNFα blockade is advisable in patients with RA with moderate to advanced HF. Our data only suggest that blocking of excess TNFα (such as that seen in patients with RA) in the absence of HF does not increase ventricular stress and may even be beneficial in terms of CV risk.
Although our study in 171 patients could be considered somewhat simplistic in design in terms of relating treatment to CV risk, it is worthwhile contextualising our findings, which we feel are at least as important as the often smaller TNFα blocking studies addressing effects on vascular function. In these studies, TNFα blockers have been shown to improve endothelial function in patients with RA with severe disease refractory to conventional therapy.25 The benefits of TNF blockade on NT-proBNP and vascular function, in turn, broadly concur with studies showing reduction in progression of subclinical atherosclerosis in patients with RA receiving anti-TNFα therapy.26 27 Both vascular function and NT-proBNP are considered intermediate markers of vascular disease but, in contrast to the limited prospective data on the range of physiological vascular function measures, circulating NT-proBNP is a well investigated and established predictor of CV events in multiple studies and across a range of populations.15 Moreover, NT-proBNP measurement is automated, standardised and clinically used in the diagnosis of HF whereas physiological vascular function measures have not yet entered mainstream clinical practice.
The prevalence of CVD and HF are increased in RA,2,–,4 and increased prevalence of diastolic and perhaps systolic dysfunction have also been observed in patients with RA.28,–,30 Circulating levels of NT-proBNP are higher in patients with RA than in control subjects, which may be a reflection of the presence of atherosclerotic disease and/or subclinical cardiac stress and mechanical stretch which may progress to overt HF.18 31 Traditional CV risk factors do not explain the excess CV burden in the RA population, and RA-specific risk factors such as systemic inflammation32,–,34 have been implicated in the excess CV risk in this population. In the present study, circulating NT-proBNP levels correlated with the inflammatory markers CRP and ESR, which is in agreement with the findings of others,17 18 and this observation—together with the finding that NT-proBNP was higher in women (as widely reported) and at older age—lends external validity to our findings. Our observations perhaps fit with the hypothesis that inflammation contributes to arterial stiffening and subsequently to an increased ventricular load.35 36
Patients with RA, particularly those with raised levels of inflammation, have increased arterial stiffness relative to the general population which is reflected by a significantly higher carotid to brachial pulse wave velocity and particularly to a higher aortic (carotid to femoral) pulse wave velocity, an effect which TNFα blocking agents may normalise.37 We now extend these findings to show that TNFα blocking in healthy patients with RA also leads to a reduction in NT-proBNP and the magnitude of change is associated with change in pulse pressure, although the latter change in itself was not significant. Based on these observations, one may argue that TNFα blocking agents reduce arterial stiffness (and hence pulse pressure) and subsequently left ventricular overload, resulting in lower levels of NT-proBNP.38 This may be one explanation as to why TNFα blocking agents are associated with a lower risk of CVD in patients with RA, although several other mechanisms may of course come into play.
The strengths and limitations of the present study merit careful consideration. Major strengths are its prospective study design and large sample size, as well as the robust and blinded NT-proBNP measurement protocol. It should be noted that the magnitude of the change in NT pro-BNP with TNFα blockade (~18%) substantially exceeded the CVs for this assay, providing further reassurance of a genuine effect. Moreover, important confounders such as antirheumatic treatment and, in the majority (>94%) of participants, cardioprotective treatment remained unchanged during the follow-up period. A final strength is that, unlike most potential novel predictors of vascular events, NT-proBNP is in clinical use and has a well-validated assay. Weaknesses include lack of a placebo controlled group and lack of a healthy control group. However, withholding treatment with proven efficacy from clinically eligible patients or, indeed, giving medication with considerable side effects (such as TNFα blockers) to healthy controls poses ethical difficulties. The lack of direct clinical imaging techniques (eg, echocardiography) or CVD endpoints included in this study can also be considered as a limitation, although we reiterate that NT-proBNP is a well-established predictor of future CV risk in general populations with and without CVD.15
In conclusion, this is the first study to report on the effect of TNFα blockers on circulating NT-proBNP levels in patients with active RA. Our novel observations suggest that blocking TNFα in patients with RA without HF does not increase but rather decreases circulating NT-proBNP levels by around 18%, going against any increase in ventricular wall stretch in these patients. Such observations are reassuring and support a potential beneficial effect of TNFα blockers on vascular risk and ventricular function in patients with RA without HF.
The authors thank Margret de Koning and Ingrid Visman for their contribution in collecting the data.
MTN and NS contributed equally.
Funding The study was facilitated by the Clinical Research Bureau of the JBI which receives financial support from the Dutch Arthritis Association. MJP received a EULAR bursary and this research was conducted while he was an ARTICULUM Fellow.
Competing interests NS and PW have had reagents for the measurement of NT-proBNP donated for unrelated research projects by Roche International. There are no other conflicts of interest.
Ethics approval The study was approved by the local medical ethics committee and all patients gave written informed consent.
Provenance and peer review Not commissioned; externally peer reviewed.
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.