• DMARDs (biologic)
• DMARDs (synthetic)
• Rheumatoid Arthritis
• Treatment
• Outcomes research

For many years, rheumatologists have used the term ‘disease-modifying antirheumatic drugs’ (DMARDs) to distinguish agents that interfere with the disease process leading to rheumatoid arthritis (RA) and influence the natural course of the disease.1–3 DMARDs inhibit joint damage, suppress the acute phase response, decrease autoantibody levels and exert effects on long-term functional outcome beyond those on signs and symptoms alone, as conveyed merely by symptomatic drugs. ‘Slow acting antirheumatic drugs’ (SAARDs)4 was an alternative term because—in contrast to the symptomatic drugs—it frequently took several weeks to months until the efficacy of these agents became recognisable,3 and even longer to reach maximal effects. However, when using biologicals that do affect structural changes, acute phase reactant and rheumatoid factor levels, rapid clinical and serological improvement is seen in a considerable number of patients, and therefore, the term SAARDs is not appropriate any more as an alternative for DMARDs.

Until about a dozen years ago, the group of DMARDs comprised solely of synthetic drugs and included methotrexate (MTX), sulfasalazine, gold salts, antimalarials and, as a newcomer, leflunomide. With the advent of biological compounds, the numbers and types of DMARDs have increased and the term ‘biological DMARDs (bDMARDs)’ came into use. Since bDMARDs were developed to be directed highly specifically at particular well-defined molecules expressed on cells or secreted into the extracellular space, they were frequently also named ‘targeted therapies’. However, the ‘old’ DMARDs now were deprived of their name; ‘non-biological DMARDs’ constituted solely an antonym to the biologics. Terms like ‘traditional’, ‘conventional’, ‘chemical’ or ‘synthetic’ were used interchangeably for this group of pharmaceuticals. And an expression like ‘non-targeted’ DMARDs does not only have a degree of inherent intricacy but also does not do full justice to the ‘old’ DMARDs since their targets may already have become known, such as in the case of leflunomide, or will be found in the future.5

For all these reasons, we propose to capture these old chemicals under the umbrella of the term ‘conventional synthetic DMARDs’ (csDMARDs): conventional here implies that they have entered the treatment armament for RA in a conventional historic way that involved fortuitous and empiric findings of disease-modifying efficacy, such as in the case of methotrexate and antimalarials,6–10 or pathogenetic assumptions, which turned out not to be appropriate, such as in the case of gold salts or sulfasalazine.11–15 This would be in clear contrast to the development of a chemical compound by complex discovery mechanisms to selectively interfere with a specific molecule (the target), such as the active site of an enzyme presumed to be of pathogenetic importance.16 So why then would leflunomide, whose target we know,5 be comprised under csDMARDs? Because it was not primarily developed to target dihydroorotate dehydrogenase, but because this enzyme was found in search of a target for the drug—that is what we still do with MTX, sulfasalazine and others without having found an ultimate molecular target.

Two recent evolutions, however, have elicited the need to revisit the term DMARDs from several new angles of perspective, namely, the approvals of tofacitinib17 and of biosimilars in the field of rheumatology.18

Biosimilars are biological agents that essentially copy the primary, secondary and tertiary structure of a parent originator compound, ideally even including its post-translational modifications, and convey similar efficacy and safety as the original protein19; indeed, an infliximab biosimilar has already passed approval by the European Medical Agency for the treatment of RA and all other indications of infliximab.18 ,20 More biosimilars in rheumatology are currently in development,19 and it will, therefore, be important to assess their long-term efficacy and safety in comparison with the respective innovator products. To this end, a name that captures all of the biosimilars may be helpful. We therefore propose the term ‘bsDMARDs’ for ‘biosimilar DMARDs’, and they would constitute a subgroup of the bDMARDs. Without doubt, different biosimilars of a specific compound, such as infliximab, rituximab or etanercept, may differ minimally in composition, although by definition of the European and US authorities, major structural variations or dissimilarities in efficacy and safety would not be acceptable19 and there is already now significant variability regarding post-translational modifications among individual batches of original biological agents produced by the parent companies.19 ,21 Nevertheless, this aspect may require further account, and one can envisage that an additional terminology may have to be introduced, such as bs-infliximab_1,2,3,… or bs-infliximab_A,B,C,…or similar to better account for potential subtle differences among bsDMARDs. Contrasting biosimilars, the biological original or biological originator agents may be subsumed as ‘boDMARDs’. While one could consider it to be sufficient naming all original biological compounds just ‘bDMARDs’ and this may, indeed, continue to be the case in everyday interactions, we would then lack an overarching term for all biological products, given the fact that obviously also bsDMARDs constitute biological agents and thus ‘b’ DMARDs.

Tofacitinib, on the other hand, is not a biological drug, even though it interferes with biological pathways and especially also pathways that are further upstream targeted by some of the available biologics. It is a synthetic, chemical, orally applied compound that has been specifically designed to inhibit Janus kinases and, indeed, does so with high potency,22 and also acts as DMARD in RA.23 So while it is a synthetic DMARD, it does not fit under the term csDMARD as discussed above since it has not been developed in the conventional way (synthesising a drug and only subsequently searching for its target), but rather using an a priori targeted approach: the targeted structure was known first and drugs blocking this enzyme were then proactively searched for; in this regard, it has been developed like targeted biological agents where monoclonal antibodies, such as after specific immunisation and further molecular modification, or receptor constructs have been shaped to bind to (and usually neutralise) a specific, predefined molecular target.24–26 Therefore, we would suggest coining the term ‘targeted synthetic DMARDs’ (tsDMARDs) to account for this and all future chemical (oral) drugs that are developed by modelling them to interact with specific, well-defined molecules or known structures, particularly aiming to inhibit their active sites.16 ,27 ,28 The term tsDMARD would consequently not account for any drug that was synthesised and especially also clinically applied before its target was known; thus, even if we were able to characterise a unique target structure for MTX or sulfasalazine, they would not fulfil the definition of a tsDMARD as proposed here, and therefore also leflunomide does not satisfy this requirement. Evidently, a tsDMARD like tofacitinib, by interfering with Janus kinases, disrupts the effects of multiple cytokines, and, therefore, we have to move beyond the thinking that targeted therapies inhibit a single cytokine or similar, especially since each cytokine elicits a plethora of effects in different cell populations and can often induce other cytokines and activate multiple signal transduction pathways29; from this perspective, inhibition of a signal transduction molecule may be regarded as more or less specific than inhibition of a cytokine, but creating drugs that interfere with a particular, known, well-defined functional moiety nevertheless has to be classified as targeted drug development. Obviously, biological agents usually target extracellular or cell membrane molecules, while chemical compounds can easily enter the cell and will usually interact with intracellular structures. Importantly, these considerations are not only pertinent for RA and many developing targeted small molecules, but likewise for other diseases requiring disease modification and where traditional and novel drugs are employed or are in development, such as systemic lupus erythematosus and other rheumatic and non-rheumatic inflammatory conditions.30

The nomenclature we propose here does not only provide mechanistic distinctions; it inherently allows making distinctions in the comprehensiveness of the approval process (eg, boDMARDs vs bsDMARDs), in the complexity of drug production (eg, tsDMARDs vs csDMARDs) and in following long-term safety and effectiveness, such as in registries. It is conceivable that some current or future agents might not clearly fall into one or the other category proposed here, but this will have to be dealt with in the most plausible way whenever pertinent.

The proposed nomenclature will also allow for better future analyses of the different classes of agents in comparison with others. When it comes to recommendations such as starting therapy with a synthetic DMARD, it will also be clearer to say ‘csDMARD’ since otherwise a drug like tofacitinib might also be comprised under such a term; however, grouping it under tsDMARDs clarifies the issue. Table 1 summarises our thinking.

In conclusion, we propose a new nomenclature to define clearly what type of DMARD is being referred to.