In a cohort of 32 patients with newly diagnosed PMR, 16 (50%) had a disease course characterized by resistance to low-dose glucocorticoid therapy, with baseline NLR proving a distinct predictor of this outcome. A trend towards significance was also observed between NLR and disease activity, with a statistically significant correlation found between PLR and PMR-AS both at baseline and during follow-up. All participants were steroid-naïve at study enrolment, glucocorticoid therapy thus having no bearing upon baseline haematologic parameters.
An increasing body of evidence now refutes historical opinion that PMR is a self-limiting, perpetually steroid-responsive entity. In a recent population-based study, around 25% of patients required more than 4 years of continuous prednisolone.(18) Significant morbidity is known to arise from this treatment paradigm whereby glucocorticoid monotherapy is advocated in the first instance and steroid-sparing agents are only initiated following recurrent relapse.(11, 19) On whole-body magnetic resonance imaging (MRI) however, a complete patient-reported response to prednisolone has been shown to correlate with an extra-capsular pattern of inflammation, thereby confirming the existence of discrete phenotypic differences between glucocorticoid-responsive and -resistant PMR cases.(20) Identification of baseline NLR as a predictor of later glucocorticoid-resistance in this study therefore has certain clinical relevance. Not only does it add to the paucity of recognized adverse prognostic indicators in this condition (such as female sex, peripheral joint involvement and high ESR)(2), but it also further characterizes the subset of patients with relapsing PMR whom may benefit from early disease modifying anti-rheumatic drug (DMARD) initiation.
Relying upon systemic markers of inflammation to assess disease activity in PMR can be problematic. Normal CRP and ESR results are observed in 14 and 27% of relapses respectively.(21) There is consequently an unmet need for a novel biomarker that accurately reflects the underlying inflammatory state in PMR. In research settings, the PMR-AS is considered a valid and reliable measure of disease activity; a value ≥ 9.35 is associated with a sensitivity of 96.6% and specificity of 90.7% for flare diagnosis.(22) The composite nature of this scoring system is however less conducive to use in everyday clinical practice. In this study, a statistically significant relationship existed between PMR-AS and both NLR and PLR at baseline, with PLR also being found to correlate with disease activity during follow-up. We therefore propose that these indices, which can be simply calculated from an FBE, represent readily accessible and inexpensive biomarkers with the potential to measure PMR disease activity more reliably than conventional inflammatory markers like CRP and ESR.
Several other studies have recently investigated trends in haematologic parameters in PMR. In their retrospective analysis of 94 PMR patients contrasted with 242 RA patients, Jung et al. found NLR and PLR to be significantly higher among PMR cases.(8) These levels subsequently diminished with treatment and were found to correlate with other measures of disease activity including CRP, however an association between NLR or PLR and a relapsing disease course was not identified in this instance. A Japanese study has however reported thrombocytosis as an adverse prognostic factor in PMR.(23) Using hierarchical cluster analysis, participants with a platelet count > 450 × 109/L were found to be less likely to exhibit a response to glucocorticoid therapy at one month. Finally, a change in leukocyte dynamics among patients with PMR and GCA compared with healthy and infection controls has been newly documented by van Sleen et al., with a shift towards the production of myeloid-lineage leukocytes noted prior to the commencement of prednisolone.(24) This bias ultimately persisted during treatment with a weaning schedule of glucocorticoid therapy and was even maintained in patients achieving drug-free remission.
Taken together, these results suggest that the inflammatory milieu which characterises PMR directly impacts neutrophil and platelet production within the bone marrow. As other authors have previously hypothesised, this likely occurs secondary to high levels of the key cytokine interleukin-6, given this pro-inflammatory mediator is additionally an independent regulator of granulopoiesis and stimulates thrombopoiesis through thrombopoietin.(25, 26) It is therefore entirely conceivable that haematologic parameters reported on an FBE may provide an accurate representation of underlying disease activity in PMR.
Whilst the precise mechanism behind the association of high NLR with adverse survival in cancer patients is not known, neutrophilia as an inflammatory response is understood to suppress other immune cells including lymphocytes, activated T cells and natural killer cells, whose infiltration of tumors is otherwise associated with positive cytotoxic treatment outcomes.(4) In RA, where a recent meta-analysis has demonstrated a consistent relationship between NLR and PLR, and the presence of active disease, these indices are similarly thought to reflect the pathogenic mechanisms at play rather than being surrogate measures of the body’s overall inflammatory state.(27)
The pathogenesis of PMR remains comparatively unclear though – it is hypothesized to arise from an aberrant immune response that follows an interaction between environmental factors, possibly viral, and the innate immune system in genetically predisposed individuals.(28) Activated macrophages predominate in both synovial and arterial samples taken from PMR patients. In the related condition GCA however, neutrophils have been implicated as the promotors of an escaped proinflammatory disease phenotype.(29) More specifically, a subset of neutrophil granulocytes capable of suppressing T cell activity following the institution of high-dose prednisolone therapy have been demonstrated to lose this ability upon glucocorticoid weaning in the context of rising IL-17 and IL-6 levels. Loss of this neutrophil suppressor function is thought to result in unchecked T cell proliferation within the vessel wall and eventual disease relapse.
Whilst this mechanism provides a biologically plausible explanation for a predominance of neutrophils enabling an adverse prognostic outcome in PMR, baseline neutrophil counts for glucocorticoid-resistant patients in our study were not substantially different from their glucocorticoid-responsive counterparts. Rather, lymphocyte counts were lower in this subset thereby producing higher NLR values. In general, no difference in total lymphocyte numbers is observed when PMR cases are compared with healthy controls.(30) Circulating B-cell lymphopaenia secondary to redistribution or intravascular marginalization of predominantly pro-inflammatory B-effector cells has however been established to exist in newly diagnosed, steroid-naïve PMR.(31) This disturbance in B cell homeostasis corrects with glucocorticoid therapy and is sustained in clinical remission but reoccurs upon disease relapse. Unfortunately, we did not undertake lymphocyte subset testing as part of this study, hence it is unclear whether lower lymphocyte counts at baseline in glucocorticoid-resistant patients represented an exaggeration of this previously described B cell phenomenon. Further investigation of lymphocyte subsets and their association with glucocorticoid resistance in PMR is now planned.
There are several other limitations to this study. It involves a relatively small number of participants and consequently the results may not be generalizable to a larger, more heterogenous PMR cohort. As glucocorticoids have the potential to impact upon measured neutrophil and platelet counts, it is similarly not known if NLR can predict glucocorticoid-resistant disease in PMR if calculated from an FBE obtained once the patient has commenced prednisolone. Finally, this study provides proof-of-concept for the utility of NLR and PLR in measuring PMR disease activity and predicting treatment response. For use in everyday practice however, cut-off values to aid clinical decision-making would be ideal. Unfortunately, this was beyond the scope of the current pilot and hence further validation of these biomarkers in a large, multi-center trial is necessary.