To the best of our knowledge, this is the first report to analyse routine inflammatory markers (ESR, CRP and leukocytosis) in treatment-naïve patients with SRDs. Most SRDs were associated with an increased ESR, whereas elevation of CRP was observed in only a few SRDs. Leukocytosis was uncommon in most patients with SRD. The degree of correlation between ESR and CRP varied markedly among SRDs. For most SRDs, WBC counts were not associated with ESR or CRP.
The human immune system responds differently to two major types of infection, namely, bacterial and viral infections. Acute bacterial infection induces production of interleukin (IL)-1, IL-6, tumour necrosis factor (TNF)-α and granulocyte-(macrophage) colony-stimulating factor (G(M)-CSF) along with preferential recruitment of myeloid cells; by contrast, viral infection promotes generation of interferon (IFN) and mobilises lymphoid cells (18–22). Interestingly, type 1 IFN suppresses innate responses to bacterial infection (23), suggesting that both anti-bacterial and anti-viral immune responses influence each other to generate “optimal” defences by (possibly) conserving limited resources. Particular SRDs such as RA and AOSD simulate an anti-bacterial response, whereas SRDs such as SLE generally mimic an anti-viral response. The type of immune response might depend on the type of cytokines that “drive” each SRD (24). For example, a large amount of type 1 IFN is generated during active SLE, which is a time when immune cells respond to human deoxyribonucleic acid/ribonucleic acid (DNA/RNA) released during tissue damage; this response is similar to responses to viral DNA/RNA. IL-1 and IL-6 are produced during both AOSD and bacterial infection, and they coordinate the “anti-bacterial” immune response (25). This complex web of different inflammatory cytokines that orchestrate autoimmune responses in SRD is further supported by findings that responses to treatments targeting a specific cytokine vary markedly between SRDs. As an example, treatments that neutralize TNF-α are quite effective in RA but not in SLE. In rare cases, TNF-α inhibitor can even exacerbate SLE disease activity (26, 27).
The profile of cytokines driving different SRDs is reflected indirectly by CRP and ESR levels. TNF-α and IL-6 increase both CRP and ESR, whereas type 1 IFN increases ESR rather than CRP. This disproportionate increase in ESR results, in part, from direct inhibition of hepatic CRP production by type 1 IFN (28). Accordingly, RA and AOSD show a concordant increase in both ESR and CRP, whereas SLE (in which production of type I IFN is dominant) shows a disproportionate increase in ESR (29–31). In addition, SLE is associated with increased production of serum immunoglobulins, which accelerates ESR (28, 32, 33). DM is also associated with type I IFN (34–36); thus patients with IIM showed low CRP levels relative to ESR, similar to SLE. Of note, we reported previously that ESR elevation is associated with interstitial lung disease and a worse outcome in patients with DM, rather than with extent of muscle inflammation; this suggests that organ involvement might influence the inflammatory marker profile in a particular SRD (37). AS did not show a significant increase in ESR or CRP when compared with OA. CPR was elevated only in 15.6% of patients with SSc, whereas ESR was elevated in 48.6%. Whether the relatively low increase in inflammatory markers results from transforming growth factor-β should be a subject to further investigations.
Strikingly, leukocytosis was observed in less than 15% of patients with SRD. Only patients with AOSD, who present with acute systemic inflammation often not distinguishable from an acute bacterial infection, had levels of leukocytosis comparable with those in patients with pneumonia. This scarcity of leukocytosis contradicts the general assumption that leukocytosis is an excellent routine marker of systemic inflammation. In addition, leukopenia was common during active SLE. Therefore, leukocytosis should not be used as the sole marker for assessing systemic inflammation in SRD.
Although ROC analysis revealed that ESR and CRP were good markers for discriminating SRD from a non-inflammatory condition (i.e., OA in this study), the cut-off values were within the normal range (or close to the upper normal limit). Therefore, inflammatory markers might have a limited value for diagnosing SRD. Indeed, in patients with AS or SSc, all three inflammatory markers had a low AUC (< 0.70). However, longitudinal changes of inflammatory markers might be still useful for monitoring disease activity.
A strength of this retrospective observational study is that we analysed inflammatory markers at the time of SRD diagnosis and prior to treatment. This eliminates treatment-related effects on inflammatory markers and thus isolates the impact of disease on these markers. However, it is unclear how these inflammatory markers vary with respect to disease progression in patients receiving or not receiving treatment. Furthermore, the association between longitudinal changes in serum cytokine, chemokine and inflammatory marker levels need to be investigated in a prospective study in which patients do and do not receive treatment.