Low adherence to cardiovascular risk assessment guidelines in patients with rheumatoid arthritis: a retrospective chart review of routine clinical practice

An observational cohort study was conducted by undertaking a retrospective chart review. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement checklist for cohort studies (Online Resource 1) was used to guide the reporting of this study [23].

Medical records were sought for all patients with a new diagnosis of RA, attending the rheumatology service of a large university hospital and tertiary referral centre located in the south of Ireland. The rheumatology centre serves a catchment population of 584,156 [24] and operates under the governance structure of a model 4 university teaching hospital. In Ireland model 4 hospitals are tertiary referral centres that provide a wide range of sub-specialties delivering ambulatory and inpatient care with an Emergency Department on site [25]. Ethical approval was granted by the Clinical Research Ethics Committee of the Cork Teaching Hospitals (protocol reference ECM 4 (L) dated July 30th, 2024) with wavier of informed consent.

As European recommendations were published in 2017 [17], we included all patients newly diagnosed with RA in the following year (between 1st January and 31st December 2018) regardless of disease severity. Given that these recommendations suggest CVD risk assessment should be undertaken for all patients with RA at least once every 5 years, the decision was made to examine patients’ routine care over a five-year period, from 2018 to 2023. To reduce selection bias, a non-probability purposive sample of medical records of all adult patients aged 18 years and above, seen for their first time by a rheumatology HCP in 2018 with a documented diagnosis of serology positive or serology negative RA, were included. The process of making a clinical diagnosis of RA is the prerogative of an experienced rheumatologist and is often distinct from that used in disease classification criteria [26]. Therefore, in line with work published in this area [27‐29], we included medical records of patients who had a diagnosis of RA made solely from the clinical assessment of a rheumatologist, as opposed to meeting international RA classification criteria [30].

We excluded medical records of patients diagnosed with juvenile arthritis or any other form of pediatric inflammatory joint disease and patients younger than 18 years of age. Patients with a known history of CVD at the time of diagnosis were excluded, as the presence of CVD requires immediate clinical management. CVD status was confirmed by reviewing patients’ medical records, including referral letters, medical and nursing notes, and correspondence from GPs.

A review code book including variable definitions was designed to aid data extraction (Online Resource 2). The data extraction tool was piloted by (LM) and (GM) independently on four patients with a new diagnosis of RA in 2017. Data from the pilot were not included in the data analysis.

Data analysis was undertaken using descriptive statistics. The Shapiro–Wilk test was used to check normality of all metric data. Continuous variables were reported using means and standard deviations for data with a normal distribution. Medians and interquartile ranges were used to report non-normally distributed data. Categorical data were reported using percentages and frequency counts, with percentages rounded to one decimal place. Categories with zero values were omitted from tables.

In 2018, the rheumatology outpatient department saw a total of 2,834 individual adult patients. A review of all 2,834 GP correspondence letters for eligibility showed 829 patients seen in 2018 had a diagnosis of RA, of which, 33 were new to the service. Of those, 21 were deemed eligible for inclusion in this study (Fig. 1).

Bild vergrößern

Patient characteristics, clinical healthcare utilisation, and clinical data representing baseline patient group and subsection analysis of those at follow-up who completed five years of routine care are shown in Table 1. The mean (SD) age of patients in 2018 at baseline was 53.5 (12.7) years. Over half of participants were female (57.1%, n = 12). The median symptom duration at baseline was seven months (IQR: 3.5, 15 months). Of the 21 patients included, five (23.8%) did not complete five full years of routine care due to discharge from service (9.5%, n = 2), patient death (9.5%, n = 2), and lost to follow-up (4.8%, n = 1).

The number of face-to-face and virtual consultations varied each year, reflecting the impact of COVID-19. In 2020, virtual consultations increased, coinciding with a decline in face-to-face visits (Fig. 2). The trend demonstrating the rise in virtual and decline in face-to-face consultations for the 21 patients included at baseline was mirrored in the subsection analysis of yearly consultation type reflecting the 16 patients who completed five years of routine care. The difference between both groups was mostly seen in the years 2018 and 2019 indicating that the five patients lost to follow-up contributed most to the face-to-face consultation numbers for those years (Fig. 2).

Bild vergrößern

Despite European guidelines recommending CVD risk assessment every five years and after major rheumatic treatment changes, our findings indicate that, in a single large rheumatology centre, these assessments are not routinely documented as part of routine care. CVD risk factor screening was consistently insufficient, resulting in a lack of documented clinical data needed for risk assessment in over half of patients at study entry and a quarter of patients at the five-year follow-up.

Of those who had clinical data documented, retrospective calculation of CVD risk demonstrated 80% of patients had an undetected moderate to very high CVD risk at time of diagnosis. At five-year follow-up, of those that had clinical data documented, 83.3% had an undetected moderate to very high CVD risk. The most consistently screened CVD risk factor was hypertension, followed by smoking status at baseline. Patients weight measurements were frequently documented at face-to-face clinic visits; however, height (necessary to calculate BMI) was only documented on one occasion for two patients. Communication to patients’ GPs about CVD risk factor assessment was inadequate, with only three documented recommendations for assessment in primary care over five years of routine care.

It has been reported that, despite being aware of RA related CVD risk, rheumatologists do not systematically screen for risk factors in practice [45]; this finding is consistent with our study. Barriers to CVD risk factor screening and subsequent assessment in rheumatology clinics (e.g., time, human resources, a lack of standardised protocols, and the asymptomatic nature of risk factors) have been identified [45‐48]. Another reason may be that rheumatology HCPs prioritise control of RA disease activity over targeting a reduction in overall CVD risk. It is well understood that elevations in acute phase reactants, pro-inflammatory cytokines, and circulating RA specific autoantibodies in states of high RA disease activity increases endothelial dysfunction and promotes atherosclerotic plaque formation [49‐51]. Therefore, it is not surprising that a reduction in systemic inflammation mitigates CVD risk in patients with RA [49, 52, 53]. While targeting RA disease control to reduce systemic inflammation and subsequent CVD risk is crucial, this approach alone is insufficient, as it overlooks the impact of traditional risk factors such as hypertension, smoking, and dyslipidaemia on CVD risk. Adequate identification of these traditional CVD risk factors is necessary so HCPs can tailor preventative care strategies and advise patients on lifestyle modifications to reduce future cardiovascular risk. Medications used to treat RA such as corticosteroids and NSAIDs, and in certain cohorts JAK inhibitors, can also increase CVD risk resulting in hypertension, thrombus formation and an increase in insulin resistance [54‐57]. For this reason, rheumatology HCPs must balance disease control with cardiovascular safety in patients with RA, which includes the recognition of preventative care and the importance of a formal CVD risk assessment in practice.

In our review, 75% (n = 12) of 16 patients had sufficient CVD risk factor screening data documented to facilitate retrospective calculation of CVD risk at five-year follow-up. This rate was higher than previously published chart reviews of only 0.68% (n = 3) [42] and 10% of 107 medical records [20]. Improvement in CVD risk screening documentation seen at five-year follow up in our study may be because of an improved awareness of the importance of CVD risk factor screening and documentation amongst rheumatology HCPs in more recent years compared to 2018. It may also be due to a perception from rheumatology HCPs that CVD risk increases the longer one is diagnosed, perhaps as disease duration was a criterion included in the preceding guidelines [58]. However, RA disease duration does not independently affect CVD risk [59]. For this reason, patients who are newly diagnosed (even those pre-diagnosis) who have high levels of inflammation can be at higher CVD risk [60, 61]. Although an improvement in documented CVD risk screening was observed in this review it is important to note that elements of patient education and/or lifestyle advice that might have been given to patients during the clinic consultation, may not have been documented in the medical notes. As this review relied on documented evidence to report outcomes, we were unable to verify undocumented interventions.

Several CVD risk factor screening and assessment clinic protocols have been reported in the literature with varying degrees of success. Rheumatology nurse-led CVD risk factor screening clinics in the USA reported successful outcomes targeting improved GP follow-up for rheumatic patients with hypertension and those who smoke [62, 63]. A more recent Danish exploratory study reported patients with RA found CVD risk factor screening consultations delivered by trained rheumatology nurses improved their awareness of CVD risk and highlighted the importance of maintaining a healthy lifestyle [64]. A Dutch study reporting results of a dedicated CVD risk screening programme noted that despite targeted advice and systematic referral to patients GPs, only one third of patients who had a very high CVD risk had adjustment of cardio preventative medication [65]. They went on to report that undergoing the CVD risk screening programme prompted a reasonable number of patients to increase their exercise, improve their diet, and lose weight [65]. In comparison a Danish cohort study reported GP follow-up rates of only 50% in patients with RA after receiving CVD risk screening consultation from a trained rheumatology nurse [66]. The authors suggested this may be due to lack of patient awareness regarding the implications of high-risk stratification. This highlights the importance of improving patient education strategies irrespective of the HCP tasked to deliver the intervention and demonstrates that cardiovascular preventative care for patients with RA remains a challenge in routine care.

We identified several missed opportunities for CVD risk assessment. Every patient consultation is an opportunity for intervention. Missed assessments delay preventative care for high-risk patients. Raising rheumatology HCP awareness of their roles in CVD prevention aligns with the “Making Every Contact Count” initiative [67]. This programme offers HCPs support in helping patients to make better lifestyle choices and lower their risk of chronic disease as is endorsed by the Health Service Executive, Ireland’s publicly funded healthcare system [68].

Disease severity assessments in real-world settings often rely on clinical opinion rather than validated composite measures [69]; therefore, the perception that a patient may not have a high CVD risk due to a low frequency of CVD risk factors may have impacted the rheumatology HCPs decision to undertake a formal risk assessment in our study. Patients who have extra articular manifestation or autoantibodies have an increased CVD risk [53, 70, 71]. For this reason, rheumatology HCPs may decide that a patient has a certain level of risk attributable to these RA disease specific factors alone and therefore feel that despite evidence to the contrary [72], a formal CVD risk assessment using a composite measure is not required.

We found no documented evidence of any communication from a rheumatology HCP to patients’ GPs to have a CVD risk assessment performed in primary care. Research conducted among GPs has reported their willingness to manage, or co-manage, RA-related comorbidities including CVD [45, 73]. European centres in the Netherlands [74] and Denmark [75] have developed shared-care pathways between rheumatologists and GPs to better communicate risk to patients and their GPs. Dedicated cardiology/ rheumatology clinics [76] have also been developed to better target and monitor at-risk patients; however, implementing such clinics worldwide may be problematic due to country-specific reimbursement regulations and differing healthcare policies [77]. Furthermore, even with a dedicated CVD risk screening programme, targeted patient advice and timely communication to GPs, initiating preventative care can still be problematic despite adequate communication to patients themselves and their general physicians about RA-related CVD risk [65]. Our findings report care provided by rheumatology HCPs who work in a publicly funded hospital within a partially integrated healthcare system. In this context, routine monitoring of blood profiles by GPs is not undertaken on all patients with RA. While in other countries collaborative care between rheumatologists and cardiologists may be common practice, it is not routinely implemented in Ireland. We examined all patients’ medical records seeking evidence of cardiology referrals or collaboration, of which we found none.

CVD risk progression, or lack thereof, may be due to interventions already undertaken by GPs as part of primary care management. For example, patients who were hypertensive or who had hypercholesterolaemia early in their disease may have had antihypertensive or lipid lowering medications commenced by their GPs as part of standard primary care management without communication to patients’ rheumatology team. In our sample, evidence of lipid lowering medications were noted in just one of 15 patients who had laboratory evidence of dyslipidaemia, with antihypertensive medications noted in use in two of 17 patients with hypertension. Our findings correspond with other studies reporting poor communication of CVD risk factors between rheumatology HCPs and GPs in the routine care of patients with RA [20, 45, 78].

A major aspect of CVD risk assessment in patients with RA is medication management including patient use of medications known to increase CVD risk. The duration of corticosteroid and NSAID use may not be adequately reflected in written or typed medical notes. Rheumatology HCPs often document current daily use without noting duration of these medications. This becomes relevant when there is prolonged use of either medication resulting in cumulative CVD risk to the patient. GPs often prescribe corticosteroids and NSAIDs in primary care to treat disease flares without consultation with the patent’s rheumatologists, and often for reasons other than control of RA. Improved collaboration between primary care and the rheumatology team would also allow for a more thorough understanding of patients’ use of both corticosteroids and NSAIDs and subsequent impact on CVD risk.

The findings from our study add to the real-world evidence highlighting issues with CVD risk guideline implementation in the routine care of patients with RA. While other papers report the presence of subclinical atherosclerosis in patients with RA and the value of diagnostic imaging techniques in advanced risk stratification [79, 80], we found a complete lack of basic guideline-recommended CVD risk assessment in this patient cohort. Ausserwinkler et al. [81] call for a pragmatic approach to CVD risk management in patients with RA which entails patient education about CVD risk, treating to a target of disease remission, minimal use of corticosteroid, and initiation of a statin where indicated. We believe, this can be achieved by implementing EULAR guidelines into practice. As our review highlights non-adherence to these guidelines in a routine rheumatology care setting, research needs to be undertaken to gain an understanding of the barriers to CVD risk assessment and guideline implementation in the care of patients with RA.

Because this was a retrospective study, data collection was limited to what was documented in patients’ medical records. As is standard in RCRs, our methodology focused on recorded interventions. It is possible that lifestyle advice provided by rheumatology HCPs was not documented and, therefore, not captured in our analysis. In the absence of documented evidence, such interventions could not be included in the dataset.

The small sample size of 21 patients and recruitment from a single centre limit the generalisability of our findings to broader populations or other healthcare settings. However, this review provides evidence of adherence to guidelines which is currently lacking in literature. The primary aim of this study was to identify patterns in treatment that may warrant further investigation. To reduce sampling bias and enhance the relevance of our findings to real-world practice, we included all patients who received a new diagnosis of RA in 2018, allowing us to reflect their routine care over a five-year period. Notably, other studies in the rheumatology literature have used retrospective designs with similar sample sizes to explore routine care and have yielded valuable insights [82‐84]. While our findings may be transferable to other centres and are relevant to both rheumatology HCPs, organisational leaders, and service managers, we advise caution in interpreting our results given the limited sample size.