Abstract:Patients with rheumatoid arthritis (RA) die at a younger age than their peers in most hospital series [1], while an increase in standardized mortality rate (SMR) has not been a consistent feature in recent studies of RA patients from date of diagnosis (inception cohorts) [2]. The reasons for this apparent discrepancy are likely to be the presence of more severe disease in hospitalbased series and the advent of earlier and more effective treatment in recent years. However, an increased mortality from vascular disease in RA has been consistently reported in both established and inception studies [1, 3], and an increased SMR for respiratory disease and lymphoreticular cancers has also been recognized in most series. A detailed study in this month’s issue addresses the important question of whether patients with RA treated early and actively with disease-modifying anti-rheumatic drugs (DMARDs) still have increased mortality rates [4]. The study assesses the effect of modern out-patient intervention on all-cause mortality in a large inception cohort over the first 7 yrs of their disease, and examines whether this relates to disease severity or duration. Detailed assessment of the causes of death offers new insights into the relevance of both cardiovascular and respiratory contributions to mortality and attempts to identify risk factor profiles. Although increasing age and male gender have been recognized as predictors for an earlier death in RA, few other clues as to an individual’s risk of premature death have been identified. However, recent work has shown direct links between rheumatoid disease activity, increased comorbidity and early death [5, 6]. The Early Rheumatoid Arthritis Study (ERAS) group [4] reports on 1429 patients recruited within 2 yrs of diagnosis over the 11 yrs from 1986. Although 84% received therapy with DMARDs, the all-cause SMR was significantly raised. After increased age, the odds ratios for death were most elevated in those patients with extra-articular disease and in those with established comorbidity. Early death was most often due to cardiovascular disease or interstitial pulmonary fibrosis, although other vascular and respiratory causes were also over-represented. Cardiovascular disease accounted for 31% of all deaths while pulmonary problems (including respiratory infection and lung cancer) were responsible for almost 29%. Ischaemic heart disease (IHD) accounted for a quarter of all deaths and carried an SMR of 1.49. It was also the most common cause of death at a young age and had a worse prognosis than in patients without RA. There was a relationship between disease activity as assessed by erythrocyte sedimentation rate (ESR) and probability of death from IHD, a finding consistent with previous reports linking increased markers of inflammation with coronary artery disease [7]. This suggests that DMARD therapy should protect RA patients from developing IHD, and reinforces the need for such therapy to be commenced early in the disease and used aggressively where possible, aiming for remission within the first year of disease. The link between cardiac disease and other therapeutic agents is less clear. Oral steroids did not have any clear effect on cardiovascular mortality but high-dose or long-term therapy might be expected to adversely affect the traditional risk factor profile by increasing blood pressure and body mass index and worsening the lipid profile. The influence of non-steroidal anti-inflammatory drugs (NSAIDs) on cardiovascular health has been much debated lately. COX-2 drugs especially have been linked with an increase in vascular events (in the order of one additional event per 100 patient years), with the data leading to the ultimate withdrawal of rofecoxib. However, a recent major meta analysis of the data on all NSAIDs report that many nonselective agents such as ibuprofen and diclofenac are also associated with an increased risk of adverse cardiovascular events, although it was unclear as to whether this effect was linked to dose or duration of treatment [8]. Perhaps the most striking aspect of the ERAS report is the high mortality resulting from interstitial pulmonary fibrosis (IPF). This accounted for 6% of all deaths and exceeded the combined number of deaths from all other extra-articular features of RA. It especially affected those patients with a younger age at onset of RA. Given that features of IPF can be shown using highresolution computed tomography (HRCT) in up to 20% of RA patients [9], this is a significant finding. Although there is some evidence that the natural history of IPF in RA is better than that in cryptogenic fibrosis [10], there is a paucity of data in the literature on this condition. Recent concern has been expressed that the natural history is worse in patients on biological agents following reports linking death from accelerated IPF in patients receiving etanercept, infliximab or adalimumab. Indeed the BSRBR confirms that deaths from IPF are increased nearly 3fold in patients on anti-tumor necrosis factor (anti-TNF) agents when compared with controls on methotrexate. Methotrexate itself has not been shown to cause deterioration in lung function in large series of patients, but nor does it appear to have any beneficial effect on IPF. Its ability to cause pneumonitis places patients with reduced respiratory reserve at increased risk, and is a cogent argument for the assessment of pulmonary function prior to its use, with patients demonstrating a restrictive defect requiring greater definition using HRCT to identify those patients with significant underlying lung disease [11]. Perhaps the choice of DMARD in patients with proven IPF should be driven by consideration of the lung disease as much as by the degree of articular involvement? Drug-related deaths were reported in the ERAS report, and these related largely to the use of NSAIDs which were responsible for eight fatalities resulting from upper gastrointestinal haemorrhage or perforation. Most of these patients had been on long-term therapy with NSAIDs but some had only recently commenced these agents. There is good evidence to implicate Helicobacter pylori infection in upper gastrointestinal ulceration, and recent work has shown that the presence of this bacterial agent together with NSAID exposure in RA patients greatly increases their risk of having an endoscopically visualized peptic ulcer [12]. Drugs may also have contributed to the greatly increased SMR (350) recorded for renal failure in the study, given the propensity of NSAIDs to precipitate renal failure in patients with reduced glomerular filtration rates. Taking these findings together with those linking NSAIDs with increased cardiovascular mortality, it is time for rheumatologists and primary care practitioners to reduce dependence on NSAIDs, particularly in patients with a high 10-yr risk of cardiovascular disease, a history of dyspepsia or evidence of reduced renal function. The risk of death as a result of infection is increased in patients with RA and this is reflected in the ERAS report. Septicaemia had Rheumatology 2007;46:183–184 doi:10.1093/rheumatology/kel332 Advance Access publication 26 September 2006

Sphingolipid and other constituents from almond nuts (Prunus amygdalus Batsch).

To do no harm — and the most good — with AI in health care

Clinician Perspectives on Decision Support and AI-based Decision Support in a Pediatric ED

An ethics assessment tool for artificial intelligence implementation in healthcare: CARE-AI

Optimizing the Clinical Direction of Artificial Intelligence With Health Policy: A Narrative Review of the Literature

Trustworthy Augmented Intelligence in Health Care

FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare

Synthesis and antiproliferative evaluation of pyrazolo[1,5-a]-1,3,5-triazine myoseverin derivatives.

Effects of Orally Effective Hypoglycaemic Agents from Plants on Alloxan Diabetes

The T9000 transputer

Applying Artificial Intelligence to Clinical Decision Support in Mental Health: What Have We Learned?

The Ethical Considerations of Artificial Intelligence in Clinical Decision Support

Interaction of intensity and order regarding painful events

The Architecture of the PCM-1

Current Challenges and Future Opportunities for XAI in Machine Learning-Based Clinical Decision Support Systems: A Systematic Review

What kills patients with rheumatoid arthritis?

Reporting guideline for the early stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI

Identifying Facilitators and Barriers to Implementation of AI-Assisted Clinical Decision Support in an Electronic Health Record System

Artificial Intelligence in Primary Health Care: Perceptions, Issues, and Challenges

Building Trust: Developing an Ethical Communication Framework for Navigating Artificial Intelligence Discussions and Addressing Potential Patient Concerns

Artificial intelligence and clinical decision support: clinicians’ perspectives on trust, trustworthiness, and liability