- Email: [email protected]
The neutrophil-lymphocyte ratio in early rheumatoid arthritis and its ability to predict subsequent failure of triple therapy
ARTICLE IN PRESS Seminars in Arthritis and Rheumatism 000 (2019) 1 4
Contents lists available at ScienceDirect
Seminars in Arthritis and Rheumatism journal homepage: www.elsevier.com/locate/semarthrit
The neutrophil-lymphocyte ratio in early rheumatoid arthritis and its ability to predict subsequent failure of triple therapy Daniel Boulosa,*, Susanna M Proudmanb,c, Robert G Metcalf b, Leah McWilliamsb, Cindy Hallb, Ian P Wicksa,* a
Department of Rheumatology, Royal Melbourne Hospital, Melbourne Health, Level 7, 300 Grattan Street, Parkville, Melbourne, Victoria, Australia Department of Rheumatology, Royal Adelaide Hospital, Adelaide, Australia c Discipline of Medicine, University of Adelaide, Adelaide, Australia b
A R T I C L E
I N F O
Keywords: Arthritis Rheumatoid Neutrophils Lymphocytes Severity of Illness Index Prognosis
A B S T R A C T
Objectives: To assess whether the neutrophil-lymphocyte ratio (NLR) and platelet-lymphocyte ratio (PLR) can predict those who subsequently require escalation of disease modifying therapy because of continued disease activity in rheumatoid arthritis (RA). Methods: Patients with newly diagnosed RA were recruited from the Early Arthritis Clinic at the Royal Adelaide Hospital. All patients commenced “triple-therapy” with a standardised protocol of methotrexate, sulfasalazine and hydroxychloroquine, and were reviewed every three to six weeks. DMARD therapy was adjusted according to a pre-defined algorithm if not in low disease activity. The NLR, PLR and other markers of disease activity including ESR, CRP and DAS28 were collected, as well as current therapy. The primary outcome measure was failure of triple-therapy to maintain low-disease activity (DAS28<3.2) at 12 months. Results: Two-hundred and twenty-two patients met inclusion criteria. The mean age was 54.2 § 15.4 years, with a mean disease duration of 22.3 § 25.0 weeks. Forty-five (20%) patients had failed triple therapy by one year. The mean baseline NLR was significantly higher in those who failed triple therapy compared with those who did not (3.7 § 2.8 vs. 2.9 § 1.5; p = 0.02), however, the PLR was not significantly different. A baseline NLR>2.7 was an independent predictor of treatment failure (OR 2.65, CI 1.23 5.72, p = 0.01) whilst the PLR, ESR, CRP and DAS-28ESR were not. Conclusion: The NLR is significantly increased in those who subsequently fail triple-therapy for RA, and it outperformed conventional markers of disease activity. The NLR may offer an inexpensive, objective and reproducible prognostic marker in RA. Further studies are justified to confirm its potential role in guiding the management of RA. © 2019 Elsevier Inc. All rights reserved.
Introduction Rheumatoid arthritis (RA) is a systemic, chronic inflammatory disorder which primarily affects synovial joints. RA is thought to be autoimmune in nature, although activation of innate immunity and of myeloid cells in particular, is characteristic [1]. Treated RA often conforms to a relapsing-remitting course, with disease flares associated with elevated inflammatory markers including the C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR). The paradigm of “treat to target” (TTT), has revolutionized management of RA, with multiple randomized controlled trials demonstrating that a TTT approach is more likely to result in disease remission, and less joint damage [2]. A number of composite tools have been developed to measure disease activity and to help guide therapy, such as the Disease Activity Score of 28 joints (DAS28), which combines
* Corresponding authors. E-mail address: [email protected] (D. Boulos). https://doi.org/10.1016/j.semarthrit.2019.05.008 0049-0172/© 2019 Elsevier Inc. All rights reserved.
inflammatory markers with other clinical parameters and patient reported outcomes. Neutrophils play a critical role in the chronic inflammation of RA, making up more than 90% of the cells in synovial fluid analysis from inflamed joints, producing multiple inflammatory mediators and tissue damaging enzymes, as well as amplifying adaptive immunity [3]. In inflammatory disorders such as multiple sclerosis, lymphoma and autoimmune thyroiditis, the neutrophil-lymphocyte ratio (NLR) has shown potential as a cheap and effective surrogate marker for the underlying inflammatory state and hence, disease activity [4 6]. The recent meta-analysis published by Erre et al. [7] highlights contemporary interest in the prognostic utility of the NLR in RA; reporting that both the NLR and PLR were predictive of the presence of RA when compared to controls. In this study, we retrospectively reviewed the records of patients with RA who attended an Early Arthritis clinic (EAC) at the Royal Adelaide Hospital to assess whether the NLR could accurately predict the need for escalation in disease modifying rheumatic drug (DMARD) therapy and hence, disease progression.
ARTICLE IN PRESS 2
D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
Methods
Results
Subjects
Participants
Patients were recruited from the EAC at the Royal Adelaide Hospital, a tertiary referral centre in South Australia. Newly referred, consenting patients who were greater than 18 years of age, fulfilled the 1987 revised American College of Rheumatology (ACR) criteria for RA [8] and were DMARD-naïve were included. Those who had been on a DMARD other than an antimalarial for a period of greater than one month, those with a known sensitivity or contraindication to DMARDs, or who were on glucocorticoids at first review, were excluded. This study was approved by the RAH Research Ethics Committee.
A total of 262 consecutive patients with RA who fulfilled the eligibility criteria were recruited from the EAC. Of these, 31 were taking oral glucocorticoids at baseline and nine did not have sufficient investigations at initial review to assess NLR, resulting in 222 patients in the final analysis. Of these, 62 (28%) participants were male, and the mean age was 54.2 § 15.4 years (Table 1). The mean duration of polyarthritis was 22.3 § 25.0 weeks prior to first review in the EAC. Forty-five (20%) patients went on to fail triple therapy by their one-year review. The NLR & PLR
Study design Patients commenced ‘triple therapy’ at the initial consultation, according to the following regimen: 10 mg methotrexate (MTX) orally weekly with concurrent folic acid, 500 mg sulfasalazine (SSZ) daily, increasing to 1 g twice daily over four weeks and 200 mg hydroxychloroquine twice daily. Participants were reviewed every three weeks for the first twelve weeks, every six weeks until the target of low disease activity (LDA) (DAS28<3.2) was reached and then every three months thereafter. Triple therapy was increased to maximal doses of MTX 25 mg/week, SSZ 3 grams/day, HCQ 400 mg/day if tolerated. MTX was switched to parenteral dosing if the target was not reached or there was gastrointestinal disturbance. In the event of failure of this combination to achieve LDA, leflunomide (LFL) was added (see Supplementary material). Parenteral glucocorticoids were allowed if clinically appropriate; however, oral steroids were avoided. Information about adverse effects secondary to DMARD therapy was collected at each visit and the offending agent either reduced or withdrawn at the physician’s discretion. Data collection Baseline demographics, rheumatoid factor and anti-citrullinated peptide antibody serology, smoking status, disease duration and comorbidities were retrieved. Full blood examination- including the NLR, inflammatory markers, disease activity as measured by the DAS28, and details of current and past therapy were collected every three months for one year and then annually hereafter. The primary outcome measure was failure of triple therapy to maintain LDA and progression to either LFL therapy or biologic if Australian Pharmaceutical Benefits Scheme criteria were met [9] within one year. Secondary outcome measures included: failure of triple therapy at six months, correlation with DAS28-ESR score and the platelet lymphocyte ratio (PLR).
The baseline NLR was significantly higher in those who subsequently failed triple therapy at one year when compared to those who did not (3.7 vs. 2.8 respectively; p <0.01). This remained true for the log of the NLR (0.50 vs. 0.40; p = <0.01) (Fig. 1). The baseline NLR did not correlate with the baseline ESR, CRP or DAS28-ESR (p values 0.18, 0.85, 0.38 respectively), however, it strongly correlated with the baseline PLR (r = 0.62, p<0.01). The baseline PLR did not correlate with the ESR (r = 0.10, p = 0.25), CRP (r = 0.05, p = 0.56) or DAS28-ESR (r = 0.11, p = 0.20). The predictive value of the baseline NLR The cut-off value for the baseline NLR to predict subsequent failure of triple therapy according to a ROC curve was 2.70, with a sensitivity of 67% and specificity of 58% (Fig. 2). The positive predictive value was 37% with a negative predictive value of 83%. The area under the receiver operating curve was 0.63 (95% CI 53 73%). Logistic regression analysis demonstrated that an NLR cut-off of 2.70 was an independent predictor of subsequent failure of triple therapy (b: 0.98, SE: 0.39, OR: 2.65, 95% CI 1.23 5.72, p = 0.01). Variables which were not predictive of failure of triple therapy included: age, gender, presence of diabetes, RF, anti-CCP, smoking status, ESR, CRP, DAS28-ESR and the PLR (See supplementary material). Discussion Neutrophils are short lived, bone marrow-derived cells and are the most abundant peripheral blood leucocyte. Neutrophil number and lifespan are under tight control. Cytokines such as G-CSF, IL-6, ILTable 1 Baseline demographics and markers of disease activity Parameter
Maintenance Triple Therapy n = 177
Failed Triple Therapy n = 45
Evidence of difference (p-value)
54.1 § 15.4 130 (73) 74 (42) 103 (58) 96 (54) 23.1 § 27.7
54.4 § 15.1 30 (67) 23 (51) 32 (71) 23 (51) 19.3 § 9.9
0.94 0.36 0.31 0.13 0.74 0.39
5.4 § 1.9 2.0 § 0.7 306.0 § 78.6 2.9 § 1.5 171.4 § 84.5 21.4 § 33.1 33.5 § 24.6 5.5 § 1.3
5.8 § 3.9 2.0 § 0.8 305.1 § 76.5 3.7 § 2.8 184.1 § 78.6 14.3 § 17.7 27.2 § 21.6 5.3 § 1.2
0.46 0.87 0.96 0.02 0.41 0.17 0.12 0.28
Statistical analysis Continuous variables were described using mean § standard deviation (SD) or median and interquartile range (IQR) if the data were parametric or non-parametric, respectively. Categorical data were summarised as absolute numbers and percentages. Receiver operating curves (ROC) and Youden’s index [10] were used to determine cut-off values of the NLR. Univariate analysis was applied to clinical and biochemical predictors of failure of triple therapy. Logistic regression analysis was applied to predictors with a p-value<0.10. Student’s t-test was used to compare continuous variables and Spearman’s coefficient was used to assess the correlation between NLR and secondary outcome measures. Given the variable nature of ratios and the hypothesized skewing of the variable in question, logarithmic transformation of the NLR was also derived. p-values 0.05 were considered significant. All statistical analyses were performed using IBM SPSS Grad Pack version 24.0 (Chicago, USA).
Demographics Age (years) Female Never smoker RF positive CCP positive Disease duration (weeks) Disease Activity Neutrophils (x109/L) Lymphocytes (x109/L) Platelets (x109/L) NLR PLR CRP (mg/L) ESR (mm/hr) DAS28-ESR
Data as mean § SD for continuous variables and (%) for categorical variables. NLR- Neutrophil-to-Lymphocyte Ratio; PLR Platelet-to-Lymphocyte Ratio; CRP reactive Protein; ESR Erythrocyte Sedimentation Rate.
C-
ARTICLE IN PRESS D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
Fig. 1. Logarithmic transformation of the baseline NLR for those who passed or subsequently failed triple therapy at 12 months.
17, GM-CSF, TNF and IL-1 contribute to neutrophil production, activation and survival [3]. Neutrophils are phagocytic and microbiocidal, releasing tissue remodelling enzymes and reactive oxidative species, as well as cytokines and chemokines. Neutrophils can generate extracellular traps (NETs), which can be proinflammatory and provide a potential source of autoantigens, triggering autoimmunity [11]. A number of studies have suggested that joint destruction and disease activity in inflammatory arthritis directly correlate with the recruitment of neutrophils in the synovium [12]. Our study aimed to determine whether the baseline NLR could predict the need for subsequent DMARD escalation because of continued disease activity in a cohort of patients with RA. Importantly, these
3
patients were all attending the one EAC, which employs protocolised treatment and a TTT approach aiming for low disease activity. It is worth appreciating that in this stringently treated cohort, LDA was achieved in 80% of participants with triple-therapy, supporting the findings of O’Dell et al. [13] reinforcing the efficacy of combination therapy after monotherapy failure. In this retrospective, observational study, we found that a higher NLR at baseline correlated with those who failed triple therapy by one year. The NLR has been shown to be predictive of poorer outcomes in oncology and other inflammatory disorders [4 6]. In contrast, this biomarker has not been extensively investigated in rheumatic diseases. As far as we are aware, this is the first study to examine real-world outcome of the management of RA in relation to the NLR. Our findings concur with those of Uslu et al. [14], who analysed a cross-sectional cohort of 104 RA patients, and found that those with active disease had a higher NLR, but not PLR, compared to those in disease remission. In a similar fashion, it appears that the NLR correlates with treatment response, with Ghang et al. [15] demonstrating that the NLR decreased significantly with tocilizumab therapy in RA (2.67§1.31 vs. 1.71§0.97; p<0.01). In a search of publicly available databases of recent clinical trials in rheumatic diseases, only the RAVE trial [16] had sufficient data to calculate the neutrophil and lymphocyte profiles of responders and nonresponders in 112 patients with ANCA-associated vasculitis. In this randomised-controlled trial, there was no significant difference between the baseline NLR of responders and non-responders to treatment at six months (7.69 § 6.93 vs. 6.92 § 4.15; p = 0.54). However, this may have been influenced by the large proportion of patients in this trial who were receiving glucocorticoids at baseline (88%). Of note, there was a trend towards higher cumulative glucocorticoid doses in the responder group (332 § 322.5 vs. 290 § 228.9 mg; p = 0.49). In our study, a higher NLR did not appear to correlate with conventional markers of RA disease activity such as the ESR, CRP and DAS28. In contrast, Uslu et al. found that the NLR correlated with the DAS28 in
Fig. 2. Receiver operating curve demonstrating sensitivity and specificity of baseline NLR in patients who subsequently failed triple therapy at one year.
ARTICLE IN PRESS 4
D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
their study and Fu et al. [17] demonstrated correlation with all three disease indices. The reasons for these discrepancies are not clear. Those patients receiving glucocorticoids at baseline were excluded from our analysis because of the ‘redistribution’ of neutrophils that occurs with administered glucocorticoids. This exclusion may have removed patients with higher levels of disease activity from our cohort. Why the NLR and PLR are highly related to each other, but only the NLR correlated with responsiveness to triple therapy, is also unclear. It is also worth noting that whilst the baseline NLR was predictive of treatment failure, the specificity and sensitivity were modest (58 and 67% respectively), however, the NPV (83%) is comparable with other screening tests [18]. Finally, uncertainty around the Youden estimate may have led to overestimation of confidence intervals. Despite these considerations, a high NLR at baseline was identified in our study as an independent predictor of subsequent failure of triple therapy (odds ratio 2.65, p = 0.01 and an area under the curve of 0.63). This predictive property was unique to the NLR and was not seen with conventional markers of disease activity, nor the PLR. Further study of this simple biomarker is warranted, with exploration of the NLR over time and its ability to direct ongoing management decisions. The NLR may have utility as a sole biomarker or as part of a multiparameter algorithm for the consideration of the early initiation of biologic therapy in early RA and perhaps other rheumatic diseases. Further prospective studies of its prognostic potential in treatment algorithms for managing RA are warranted. Declaration of interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Terry Speed and Leon di Stefano who aided with statistical methods and study design. Funding Melrose Health for supporting ongoing studies. Supplementary materials Supplementary material associated with this article can be found in the online version at doi:10.1016/j.semarthrit.2019.05.008.
References [1] Cojocaru M, Cojocaru IM, Silosi I, Vrabie CD, Tanasescu R. Extra-articular manifestations in rheumatoid arthritis. Mædica 2010;5(4):286–91. [2] Hørslev-Petersen K, Hetland ML, Ørnbjerg LM, Junker P, Pødenphant J, Ellingsen T, et al. Clinical and radiographic outcome of a treat-to-target strategy using methotrexate and intra-articular glucocorticoids with or without adalimumab induction: a 2-year investigator-initiated, double-blinded, randomised, controlled trial (OPERA). Ann Rheum Dis 2016;75(9):1645–53. [3] Wright HL, Moots RJ, Edwards SW. The multifactorial role of neutrophils in rheumatoid arthritis. Nat Rev Rheumatol 2014;10:593. [4] Demirci S, Demirci S, Kutluhan S, Koyuncuoglu HR, Yurekli VA. The clinical significance of the neutrophil-to-lymphocyte ratio in multiple sclerosis. Int J Neurosci 2016;126(8):700–6. [5] Wang J, Zhou M, Xu J-Y, Yang Y-G, Zhang Q-G, Zhou R-F, et al. Prognostic role of pretreatment neutrophil-lymphocyte ratio in patients with diffuse large B-cell lymphoma treated with RCHOP. Medicine 2016;95(38):e4893. [6] Keskin H, Kaya Y, Cadirci K, Kucur C, Ziypak E, Simsek E, et al. Elevated neutrophillymphocyte ratio in patients with euthyroid chronic autoimmune thyreotidis. Endocr Regul 2016;50(3):148–53. [7] Erre GL, Paliogiannis P, Castagna F, Mangoni AA, Carru C, Passiu G, et al. Metaanalysis of neutrophil-to-lymphocyte and platelet-to-lymphocyte ratio in rheumatoid arthritis. Eur J Clin Invest 2019;49(1):e13037. [8] Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31(3):315–24. [9] Department of Human Services. Rheumatoid arthritis initial applications (PB109): Australian Government. Available from: http://www.medicareaustralia.gov.au/ provider/pbs/drugs1/files/4141-2-rheumatoid-arthritis-pbs-authority-application-continuing.pdf. [10] Ruopp MD, Perkins NJ, Whitcomb BW, Schisterman EF. Youden index and optimal cut-point estimated from observations affected by a lower limit of detection. Biom J Biom Zeitschrift 2008;50(3):419–30. [11] Wright HL, Moots RJ, Edwards SW. The multifactorial role of neutrophils in rheumatoid arthritis. Nat Rev Rheumatol 2014;10(10):593–601. [12] Sadik CD, Kim ND, Iwakura Y, Luster AD. Neutrophils orchestrate their own recruitment in murine arthritis through C5aR and FcgammaR signaling. Proc Natl Acad Sci USA 2012;109(46):E3177–85. [13] O'Dell JR, Mikuls TR, Taylor TH, Ahluwalia V, Brophy M, Warren SR, et al. Therapies for active rheumatoid arthritis after methotrexate failure. N Engl J Med 2013;369(4):307–18. [14] Uslu AU, Kucuk A, Sahin A, Ugan Y, Yilmaz R, Gungor T, et al. Two new inflammatory markers associated with Disease Activity Score-28 in patients with rheumatoid arthritis: neutrophil-lymphocyte ratio and platelet-lymphocyte ratio. Int J Rheum Dis 2015;18(7):731–5. [15] Ghang B, Kwon O, Hong S, Lee C-K, Yoo B, Kim Y-G. Neutrophil-to-lymphocyte ratio is a reliable marker of treatment response in rheumatoid arthritis patients during tocilizumab therapy. Modern Rheumatol 2017;27(3):405–10. [16] Grayson PC, Carmona-Rivera C, Xu L, Lim N, Gao Z, Asare AL, et al. Neutrophilrelated gene expression and low-density granulocytes associated with disease activity and response to treatment in ANCA-associated vasculitis. Arthritis Rheumatol (Hoboken, NJ) 2015;67(7):1922–32. [17] Fu H, Qin B, Hu Z, Ma N, Yang M, Wei T, et al. Neutrophil- and platelet-to-lymphocyte ratios are correlated with disease activity in rheumatoid arthritis. Clin Lab 2015;61(3 4):269–73. [18] Trevethan R. Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and practice. Front Public Health 2017;5:307.
Contents lists available at ScienceDirect
Seminars in Arthritis and Rheumatism journal homepage: www.elsevier.com/locate/semarthrit
The neutrophil-lymphocyte ratio in early rheumatoid arthritis and its ability to predict subsequent failure of triple therapy Daniel Boulosa,*, Susanna M Proudmanb,c, Robert G Metcalf b, Leah McWilliamsb, Cindy Hallb, Ian P Wicksa,* a
Department of Rheumatology, Royal Melbourne Hospital, Melbourne Health, Level 7, 300 Grattan Street, Parkville, Melbourne, Victoria, Australia Department of Rheumatology, Royal Adelaide Hospital, Adelaide, Australia c Discipline of Medicine, University of Adelaide, Adelaide, Australia b
A R T I C L E
I N F O
Keywords: Arthritis Rheumatoid Neutrophils Lymphocytes Severity of Illness Index Prognosis
A B S T R A C T
Objectives: To assess whether the neutrophil-lymphocyte ratio (NLR) and platelet-lymphocyte ratio (PLR) can predict those who subsequently require escalation of disease modifying therapy because of continued disease activity in rheumatoid arthritis (RA). Methods: Patients with newly diagnosed RA were recruited from the Early Arthritis Clinic at the Royal Adelaide Hospital. All patients commenced “triple-therapy” with a standardised protocol of methotrexate, sulfasalazine and hydroxychloroquine, and were reviewed every three to six weeks. DMARD therapy was adjusted according to a pre-defined algorithm if not in low disease activity. The NLR, PLR and other markers of disease activity including ESR, CRP and DAS28 were collected, as well as current therapy. The primary outcome measure was failure of triple-therapy to maintain low-disease activity (DAS28<3.2) at 12 months. Results: Two-hundred and twenty-two patients met inclusion criteria. The mean age was 54.2 § 15.4 years, with a mean disease duration of 22.3 § 25.0 weeks. Forty-five (20%) patients had failed triple therapy by one year. The mean baseline NLR was significantly higher in those who failed triple therapy compared with those who did not (3.7 § 2.8 vs. 2.9 § 1.5; p = 0.02), however, the PLR was not significantly different. A baseline NLR>2.7 was an independent predictor of treatment failure (OR 2.65, CI 1.23 5.72, p = 0.01) whilst the PLR, ESR, CRP and DAS-28ESR were not. Conclusion: The NLR is significantly increased in those who subsequently fail triple-therapy for RA, and it outperformed conventional markers of disease activity. The NLR may offer an inexpensive, objective and reproducible prognostic marker in RA. Further studies are justified to confirm its potential role in guiding the management of RA. © 2019 Elsevier Inc. All rights reserved.
Introduction Rheumatoid arthritis (RA) is a systemic, chronic inflammatory disorder which primarily affects synovial joints. RA is thought to be autoimmune in nature, although activation of innate immunity and of myeloid cells in particular, is characteristic [1]. Treated RA often conforms to a relapsing-remitting course, with disease flares associated with elevated inflammatory markers including the C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR). The paradigm of “treat to target” (TTT), has revolutionized management of RA, with multiple randomized controlled trials demonstrating that a TTT approach is more likely to result in disease remission, and less joint damage [2]. A number of composite tools have been developed to measure disease activity and to help guide therapy, such as the Disease Activity Score of 28 joints (DAS28), which combines
* Corresponding authors. E-mail address: [email protected] (D. Boulos). https://doi.org/10.1016/j.semarthrit.2019.05.008 0049-0172/© 2019 Elsevier Inc. All rights reserved.
inflammatory markers with other clinical parameters and patient reported outcomes. Neutrophils play a critical role in the chronic inflammation of RA, making up more than 90% of the cells in synovial fluid analysis from inflamed joints, producing multiple inflammatory mediators and tissue damaging enzymes, as well as amplifying adaptive immunity [3]. In inflammatory disorders such as multiple sclerosis, lymphoma and autoimmune thyroiditis, the neutrophil-lymphocyte ratio (NLR) has shown potential as a cheap and effective surrogate marker for the underlying inflammatory state and hence, disease activity [4 6]. The recent meta-analysis published by Erre et al. [7] highlights contemporary interest in the prognostic utility of the NLR in RA; reporting that both the NLR and PLR were predictive of the presence of RA when compared to controls. In this study, we retrospectively reviewed the records of patients with RA who attended an Early Arthritis clinic (EAC) at the Royal Adelaide Hospital to assess whether the NLR could accurately predict the need for escalation in disease modifying rheumatic drug (DMARD) therapy and hence, disease progression.
ARTICLE IN PRESS 2
D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
Methods
Results
Subjects
Participants
Patients were recruited from the EAC at the Royal Adelaide Hospital, a tertiary referral centre in South Australia. Newly referred, consenting patients who were greater than 18 years of age, fulfilled the 1987 revised American College of Rheumatology (ACR) criteria for RA [8] and were DMARD-naïve were included. Those who had been on a DMARD other than an antimalarial for a period of greater than one month, those with a known sensitivity or contraindication to DMARDs, or who were on glucocorticoids at first review, were excluded. This study was approved by the RAH Research Ethics Committee.
A total of 262 consecutive patients with RA who fulfilled the eligibility criteria were recruited from the EAC. Of these, 31 were taking oral glucocorticoids at baseline and nine did not have sufficient investigations at initial review to assess NLR, resulting in 222 patients in the final analysis. Of these, 62 (28%) participants were male, and the mean age was 54.2 § 15.4 years (Table 1). The mean duration of polyarthritis was 22.3 § 25.0 weeks prior to first review in the EAC. Forty-five (20%) patients went on to fail triple therapy by their one-year review. The NLR & PLR
Study design Patients commenced ‘triple therapy’ at the initial consultation, according to the following regimen: 10 mg methotrexate (MTX) orally weekly with concurrent folic acid, 500 mg sulfasalazine (SSZ) daily, increasing to 1 g twice daily over four weeks and 200 mg hydroxychloroquine twice daily. Participants were reviewed every three weeks for the first twelve weeks, every six weeks until the target of low disease activity (LDA) (DAS28<3.2) was reached and then every three months thereafter. Triple therapy was increased to maximal doses of MTX 25 mg/week, SSZ 3 grams/day, HCQ 400 mg/day if tolerated. MTX was switched to parenteral dosing if the target was not reached or there was gastrointestinal disturbance. In the event of failure of this combination to achieve LDA, leflunomide (LFL) was added (see Supplementary material). Parenteral glucocorticoids were allowed if clinically appropriate; however, oral steroids were avoided. Information about adverse effects secondary to DMARD therapy was collected at each visit and the offending agent either reduced or withdrawn at the physician’s discretion. Data collection Baseline demographics, rheumatoid factor and anti-citrullinated peptide antibody serology, smoking status, disease duration and comorbidities were retrieved. Full blood examination- including the NLR, inflammatory markers, disease activity as measured by the DAS28, and details of current and past therapy were collected every three months for one year and then annually hereafter. The primary outcome measure was failure of triple therapy to maintain LDA and progression to either LFL therapy or biologic if Australian Pharmaceutical Benefits Scheme criteria were met [9] within one year. Secondary outcome measures included: failure of triple therapy at six months, correlation with DAS28-ESR score and the platelet lymphocyte ratio (PLR).
The baseline NLR was significantly higher in those who subsequently failed triple therapy at one year when compared to those who did not (3.7 vs. 2.8 respectively; p <0.01). This remained true for the log of the NLR (0.50 vs. 0.40; p = <0.01) (Fig. 1). The baseline NLR did not correlate with the baseline ESR, CRP or DAS28-ESR (p values 0.18, 0.85, 0.38 respectively), however, it strongly correlated with the baseline PLR (r = 0.62, p<0.01). The baseline PLR did not correlate with the ESR (r = 0.10, p = 0.25), CRP (r = 0.05, p = 0.56) or DAS28-ESR (r = 0.11, p = 0.20). The predictive value of the baseline NLR The cut-off value for the baseline NLR to predict subsequent failure of triple therapy according to a ROC curve was 2.70, with a sensitivity of 67% and specificity of 58% (Fig. 2). The positive predictive value was 37% with a negative predictive value of 83%. The area under the receiver operating curve was 0.63 (95% CI 53 73%). Logistic regression analysis demonstrated that an NLR cut-off of 2.70 was an independent predictor of subsequent failure of triple therapy (b: 0.98, SE: 0.39, OR: 2.65, 95% CI 1.23 5.72, p = 0.01). Variables which were not predictive of failure of triple therapy included: age, gender, presence of diabetes, RF, anti-CCP, smoking status, ESR, CRP, DAS28-ESR and the PLR (See supplementary material). Discussion Neutrophils are short lived, bone marrow-derived cells and are the most abundant peripheral blood leucocyte. Neutrophil number and lifespan are under tight control. Cytokines such as G-CSF, IL-6, ILTable 1 Baseline demographics and markers of disease activity Parameter
Maintenance Triple Therapy n = 177
Failed Triple Therapy n = 45
Evidence of difference (p-value)
54.1 § 15.4 130 (73) 74 (42) 103 (58) 96 (54) 23.1 § 27.7
54.4 § 15.1 30 (67) 23 (51) 32 (71) 23 (51) 19.3 § 9.9
0.94 0.36 0.31 0.13 0.74 0.39
5.4 § 1.9 2.0 § 0.7 306.0 § 78.6 2.9 § 1.5 171.4 § 84.5 21.4 § 33.1 33.5 § 24.6 5.5 § 1.3
5.8 § 3.9 2.0 § 0.8 305.1 § 76.5 3.7 § 2.8 184.1 § 78.6 14.3 § 17.7 27.2 § 21.6 5.3 § 1.2
0.46 0.87 0.96 0.02 0.41 0.17 0.12 0.28
Statistical analysis Continuous variables were described using mean § standard deviation (SD) or median and interquartile range (IQR) if the data were parametric or non-parametric, respectively. Categorical data were summarised as absolute numbers and percentages. Receiver operating curves (ROC) and Youden’s index [10] were used to determine cut-off values of the NLR. Univariate analysis was applied to clinical and biochemical predictors of failure of triple therapy. Logistic regression analysis was applied to predictors with a p-value<0.10. Student’s t-test was used to compare continuous variables and Spearman’s coefficient was used to assess the correlation between NLR and secondary outcome measures. Given the variable nature of ratios and the hypothesized skewing of the variable in question, logarithmic transformation of the NLR was also derived. p-values 0.05 were considered significant. All statistical analyses were performed using IBM SPSS Grad Pack version 24.0 (Chicago, USA).
Demographics Age (years) Female Never smoker RF positive CCP positive Disease duration (weeks) Disease Activity Neutrophils (x109/L) Lymphocytes (x109/L) Platelets (x109/L) NLR PLR CRP (mg/L) ESR (mm/hr) DAS28-ESR
Data as mean § SD for continuous variables and (%) for categorical variables. NLR- Neutrophil-to-Lymphocyte Ratio; PLR Platelet-to-Lymphocyte Ratio; CRP reactive Protein; ESR Erythrocyte Sedimentation Rate.
C-
ARTICLE IN PRESS D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
Fig. 1. Logarithmic transformation of the baseline NLR for those who passed or subsequently failed triple therapy at 12 months.
17, GM-CSF, TNF and IL-1 contribute to neutrophil production, activation and survival [3]. Neutrophils are phagocytic and microbiocidal, releasing tissue remodelling enzymes and reactive oxidative species, as well as cytokines and chemokines. Neutrophils can generate extracellular traps (NETs), which can be proinflammatory and provide a potential source of autoantigens, triggering autoimmunity [11]. A number of studies have suggested that joint destruction and disease activity in inflammatory arthritis directly correlate with the recruitment of neutrophils in the synovium [12]. Our study aimed to determine whether the baseline NLR could predict the need for subsequent DMARD escalation because of continued disease activity in a cohort of patients with RA. Importantly, these
3
patients were all attending the one EAC, which employs protocolised treatment and a TTT approach aiming for low disease activity. It is worth appreciating that in this stringently treated cohort, LDA was achieved in 80% of participants with triple-therapy, supporting the findings of O’Dell et al. [13] reinforcing the efficacy of combination therapy after monotherapy failure. In this retrospective, observational study, we found that a higher NLR at baseline correlated with those who failed triple therapy by one year. The NLR has been shown to be predictive of poorer outcomes in oncology and other inflammatory disorders [4 6]. In contrast, this biomarker has not been extensively investigated in rheumatic diseases. As far as we are aware, this is the first study to examine real-world outcome of the management of RA in relation to the NLR. Our findings concur with those of Uslu et al. [14], who analysed a cross-sectional cohort of 104 RA patients, and found that those with active disease had a higher NLR, but not PLR, compared to those in disease remission. In a similar fashion, it appears that the NLR correlates with treatment response, with Ghang et al. [15] demonstrating that the NLR decreased significantly with tocilizumab therapy in RA (2.67§1.31 vs. 1.71§0.97; p<0.01). In a search of publicly available databases of recent clinical trials in rheumatic diseases, only the RAVE trial [16] had sufficient data to calculate the neutrophil and lymphocyte profiles of responders and nonresponders in 112 patients with ANCA-associated vasculitis. In this randomised-controlled trial, there was no significant difference between the baseline NLR of responders and non-responders to treatment at six months (7.69 § 6.93 vs. 6.92 § 4.15; p = 0.54). However, this may have been influenced by the large proportion of patients in this trial who were receiving glucocorticoids at baseline (88%). Of note, there was a trend towards higher cumulative glucocorticoid doses in the responder group (332 § 322.5 vs. 290 § 228.9 mg; p = 0.49). In our study, a higher NLR did not appear to correlate with conventional markers of RA disease activity such as the ESR, CRP and DAS28. In contrast, Uslu et al. found that the NLR correlated with the DAS28 in
Fig. 2. Receiver operating curve demonstrating sensitivity and specificity of baseline NLR in patients who subsequently failed triple therapy at one year.
ARTICLE IN PRESS 4
D. Boulos et al. / Seminars in Arthritis and Rheumatism 00 (2019) 1 4
their study and Fu et al. [17] demonstrated correlation with all three disease indices. The reasons for these discrepancies are not clear. Those patients receiving glucocorticoids at baseline were excluded from our analysis because of the ‘redistribution’ of neutrophils that occurs with administered glucocorticoids. This exclusion may have removed patients with higher levels of disease activity from our cohort. Why the NLR and PLR are highly related to each other, but only the NLR correlated with responsiveness to triple therapy, is also unclear. It is also worth noting that whilst the baseline NLR was predictive of treatment failure, the specificity and sensitivity were modest (58 and 67% respectively), however, the NPV (83%) is comparable with other screening tests [18]. Finally, uncertainty around the Youden estimate may have led to overestimation of confidence intervals. Despite these considerations, a high NLR at baseline was identified in our study as an independent predictor of subsequent failure of triple therapy (odds ratio 2.65, p = 0.01 and an area under the curve of 0.63). This predictive property was unique to the NLR and was not seen with conventional markers of disease activity, nor the PLR. Further study of this simple biomarker is warranted, with exploration of the NLR over time and its ability to direct ongoing management decisions. The NLR may have utility as a sole biomarker or as part of a multiparameter algorithm for the consideration of the early initiation of biologic therapy in early RA and perhaps other rheumatic diseases. Further prospective studies of its prognostic potential in treatment algorithms for managing RA are warranted. Declaration of interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Terry Speed and Leon di Stefano who aided with statistical methods and study design. Funding Melrose Health for supporting ongoing studies. Supplementary materials Supplementary material associated with this article can be found in the online version at doi:10.1016/j.semarthrit.2019.05.008.
References [1] Cojocaru M, Cojocaru IM, Silosi I, Vrabie CD, Tanasescu R. Extra-articular manifestations in rheumatoid arthritis. Mædica 2010;5(4):286–91. [2] Hørslev-Petersen K, Hetland ML, Ørnbjerg LM, Junker P, Pødenphant J, Ellingsen T, et al. Clinical and radiographic outcome of a treat-to-target strategy using methotrexate and intra-articular glucocorticoids with or without adalimumab induction: a 2-year investigator-initiated, double-blinded, randomised, controlled trial (OPERA). Ann Rheum Dis 2016;75(9):1645–53. [3] Wright HL, Moots RJ, Edwards SW. The multifactorial role of neutrophils in rheumatoid arthritis. Nat Rev Rheumatol 2014;10:593. [4] Demirci S, Demirci S, Kutluhan S, Koyuncuoglu HR, Yurekli VA. The clinical significance of the neutrophil-to-lymphocyte ratio in multiple sclerosis. Int J Neurosci 2016;126(8):700–6. [5] Wang J, Zhou M, Xu J-Y, Yang Y-G, Zhang Q-G, Zhou R-F, et al. Prognostic role of pretreatment neutrophil-lymphocyte ratio in patients with diffuse large B-cell lymphoma treated with RCHOP. Medicine 2016;95(38):e4893. [6] Keskin H, Kaya Y, Cadirci K, Kucur C, Ziypak E, Simsek E, et al. Elevated neutrophillymphocyte ratio in patients with euthyroid chronic autoimmune thyreotidis. Endocr Regul 2016;50(3):148–53. [7] Erre GL, Paliogiannis P, Castagna F, Mangoni AA, Carru C, Passiu G, et al. Metaanalysis of neutrophil-to-lymphocyte and platelet-to-lymphocyte ratio in rheumatoid arthritis. Eur J Clin Invest 2019;49(1):e13037. [8] Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31(3):315–24. [9] Department of Human Services. Rheumatoid arthritis initial applications (PB109): Australian Government. Available from: http://www.medicareaustralia.gov.au/ provider/pbs/drugs1/files/4141-2-rheumatoid-arthritis-pbs-authority-application-continuing.pdf. [10] Ruopp MD, Perkins NJ, Whitcomb BW, Schisterman EF. Youden index and optimal cut-point estimated from observations affected by a lower limit of detection. Biom J Biom Zeitschrift 2008;50(3):419–30. [11] Wright HL, Moots RJ, Edwards SW. The multifactorial role of neutrophils in rheumatoid arthritis. Nat Rev Rheumatol 2014;10(10):593–601. [12] Sadik CD, Kim ND, Iwakura Y, Luster AD. Neutrophils orchestrate their own recruitment in murine arthritis through C5aR and FcgammaR signaling. Proc Natl Acad Sci USA 2012;109(46):E3177–85. [13] O'Dell JR, Mikuls TR, Taylor TH, Ahluwalia V, Brophy M, Warren SR, et al. Therapies for active rheumatoid arthritis after methotrexate failure. N Engl J Med 2013;369(4):307–18. [14] Uslu AU, Kucuk A, Sahin A, Ugan Y, Yilmaz R, Gungor T, et al. Two new inflammatory markers associated with Disease Activity Score-28 in patients with rheumatoid arthritis: neutrophil-lymphocyte ratio and platelet-lymphocyte ratio. Int J Rheum Dis 2015;18(7):731–5. [15] Ghang B, Kwon O, Hong S, Lee C-K, Yoo B, Kim Y-G. Neutrophil-to-lymphocyte ratio is a reliable marker of treatment response in rheumatoid arthritis patients during tocilizumab therapy. Modern Rheumatol 2017;27(3):405–10. [16] Grayson PC, Carmona-Rivera C, Xu L, Lim N, Gao Z, Asare AL, et al. Neutrophilrelated gene expression and low-density granulocytes associated with disease activity and response to treatment in ANCA-associated vasculitis. Arthritis Rheumatol (Hoboken, NJ) 2015;67(7):1922–32. [17] Fu H, Qin B, Hu Z, Ma N, Yang M, Wei T, et al. Neutrophil- and platelet-to-lymphocyte ratios are correlated with disease activity in rheumatoid arthritis. Clin Lab 2015;61(3 4):269–73. [18] Trevethan R. Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and practice. Front Public Health 2017;5:307.