ANCA-associated vasculitis (AAV) is defined by the 2012 Chapel Hill Consensus Conference Nomenclature of Vasculitides (CHCC 2012) as necrotizing vasculitis, predominantly affecting small vessels, with few or no immune deposits.1 AAV is a systemic inflammatory disease associated with myeloperoxidase (MPO) ANCA or proteinase 3 (PR3) ANCA, although approximately 10% of patients are ANCA negative.2

Clinical features may arise as nonspecific symptoms such as fatigue, anorexia, weight loss, arthralgias, myalgias and fever, with other clinical manifestations depending on the organs affected by vasculitis. Clinical and pathological findings subdivide AAV in three different clinical syndromes: microscopic polyangiitis (MPA), granulomatosis with polyangiitis (GPA) and eosinophilic granulomatosis with polyangiitis (EGPA).3

Renal involvement is frequent in AAV, particularly in patients with MPA and GPA, and clinical presentation typically includes microscopic hematuria, low grade to sub-nephrotic proteinuria and/or a decline of kidney function frequently presenting as rapidly progressive glomerulonephritis.4,5

Prognosis has evolved favorably since the introduction of corticosteroids first and cyclophosphamide after, with mortality rates dropping from 80% in the first year to approximately 20% in most studies on a five year follow up.6 However, end-stage kidney disease and short-term mortality within one year after diagnosis remain important issues to address, particularly in those who present with rapidly progressive glomerulonephritis and alveolar hemorrhage.7

Predictors of clinical outcomes in AAV have paramount importance in determining which patients may benefit from more aggressive immunosuppressive therapy. Therefore, genetic, histopathological and clinical predictors have been researched, although validation in multicenter randomized controlled trials of these predictors remains scarce.8–10

The neutrophil-lymphocyte ratio (NLR) is a biomarker of systemic inflammation which has shown prognostic ability in multiple clinical settings, namely in progression of chronic kidney disease,11 acute kidney injury,12 cancer patients13 and in renal transplantation.14 Whether the same prognostic ability can be extended to AAV remains unknown.

The aim of this study was to evaluate the prognostic ability of the NLR at admission in patients presenting with de novo AAV for development of severe infection and mortality within the first year.

We registered 45 cases of AAV. The mean age at diagnosis was 67.5±12.1 years, 23 patients (51.1%) were male and the vast majority were Caucasian (97.8%, n=44).

Eighty percent of patients (n=36) presented with rapidly progressive renal failure (RPRF), 25 patients (55.6%) required hemodialysis start, and 13 patients (28.9%) presented alveolar hemorrhage. The mean Birmingham Vasculitis Activity Score (BVAS) at presentation was 26.0±10.4.

At admission, mean serum creatinine (SCr) was 4.92±2.52mg/dL, hemoglobin was 9.57±1.72g/dL, NLR was 8.54±6.83, erythrocyte sedimentation rate (ESR) was 76.90±33.85mm/h and C-reactive protein was 13.2±5.8mg/dL. Twenty-nine patients (64.4%) were ANCA-MPO positive, 7 patients (15.6%) ANCA-PR3 positive and 9 patients (20%) were negative ANCA vasculitis.

Concerning induction therapy, 35 patients (77.8%) were treated with cyclophosphamide, 35 patients (77.8%) with corticosteroid pulses, eight patients (17.8%) with rituximab for induction therapy, of which five were treated with both cyclophosphamide and rituximab. One-third of patients (33.3%, n=15) underwent plasmapheresis.

NLR and severe infection

A ROC curve was produced to evaluate the discriminative ability of the NL ratio for severe infection at 3 months after starting induction immunossupression. The AUC for prediction of severe infection at 3 months was of 0.0794 (95% CI 0.647–0.900, p<0.0001) (Fig. 1). The optimal cut-off was assessed to be ≥4.04, which has a sensitivity of 95% and specificity of 52%.

Fig. 1.

AUC of the risk model for the prediction of severe sepsis ate 3 months with NL ratio in ANCA-vasculitis patients.

(0.11MB).

This cut-off was also used to assess disease activity. However, there was no significant association between NLR ≥4.04 and BVAS (p=0.647).

In this retrospective cohort, we demonstrated that a higher NLR at admission was independently associated with development of severe infection within 3 months of starting induction therapy in AAV patients. Additionally, development of severe infection was associated with one-year mortality.

The association of this marker with prognosis in this setting reinforces the role of inflammation in ANCA-vasculitis. Although the events leading to vasculitis initiation remain unproven, several hypothesis have been posed, which highlight the role of genetic, infectious and environmental factors in the pathogenesis of AAV.18

Inflammatory cytokines, such as tumor necrosis factor α and interleukin 1, play a role through priming neutrophils, inducing increased expression of ANCA antigens on their surfaces, enabling ANCAs (formed through prolonged exposure of antigens to the immune system or infection-induced molecular mimicry) to bind them, causing neutrophil activation. Activated neutrophils alter the expression of cell surface adhesion molecules, making possible the attachment to vascular endothelial cells and are responsible for releasing lytic enzymes from granules, reactive oxygen species and neutrophil extracellular traps (NETS), thus provoking damage to the endothelial cells. Moreover, activated neutrophils release factors that activate the alternative complement pathway, generating C5a which amplifies the inflammatory reaction.19–22

NLR has been described in recent literature as a significant marker of inflammation and a useful predictor of the development of acute kidney injury in different settings, namely in the emergency department, sepsis-associated AKI,23,24 contrast-induced AKI,25 after cardiovascular surgery 26 and major abdominal surgery.27

Neutrophil counts are usually proportionally elevated to inflammation, and activated neutrophils are associated with the pathogenesis of AAV.22 Also lymphocyte counts may be diminished in inflammatory states, namely in active AAV in which lymphopenia results from extensive recruitment of peripheral T cells to the affected tissues.28,29 Thus, the NLR could hypothetically reflect the inflammatory load of AAV patients.

Furthermore, neutrophilia and lymphopenia can be a result of immunosuppression or a result of systemic infection developed during immunosuppressive treatment.30 Therefore, to reduce bias the ratio should be calculated at disease presentation before any therapeutic measure takes place.

Validation of this practical, inexpensive and easily calculated biomarker has been explored in AAV patients, with early studies suggesting a positive relation between the NLR and BVAS, acute phase reactants and worse renal outcomes.31 Other studies, however, have shown conflicting results between the ratio and disease activity as well as extrarenal manifestations.32

Recently, Ahn et al. retrospectively analyzed 160 AAV patients and reported that an elevated NLR was correlated with vasculitis activity estimated by BVAS at diagnosis and predicted relapse during follow-up (Patients having NLR ≥5.9 exhibited severe AAV more frequently than those having NLR<5.9 at diagnosis (relative 2.189, p=0.023).).33 In our study, we have not confirmed this association. Park et al. also proposed a platelet to lymphocyte ratio (PLR) to predict AAV inflammatory burden, in which a PLR ≥272.0 was able to predict severe AAV at diagnosis.34

The assessment of clinical predictors of complications associated with the immunosuppressive treatment of AAV has been barely done. On a literature review, there have been no reports of NLR used as a predictor of severe infections in AAV patients. Our study therefore reflects the importance of the NLR as an indirect and sensitive predictor of inflammation and outcomes in this setting.

Although novel therapeutic options have been developed for patients with AAV,35,36 long-term safety still composes a major concern, specifically in terms of infectious complications and neoplasms, since these therapeutics do not appear to significantly decrease the risk of developing such adverse events.37

In fact, infections were responsible for hospitalization of more than 30% of patients treated with immunosuppressive therapy on a five-year follow up and the leading cause of death in those patients.7,38

Age over 65 years, pulmonary involvement and the Five Factor Score (FFS) constitute baseline independent predictors of severe infection according to a retrospective analysis of five major randomized controlled trials (RCTs) of the French Vasculitis Study Group.39 A higher BVAS at diagnosis has also shown prognostic ability in assessing development of serious infection in AAV patients.40 Smoking status and baseline serum creatinine at diagnosis may also play a role as risk factors for infectious complications.41

Data regarding infectious complications related to treatment modalities and strategies remain controversial. Although no therapeutic regimen was significantly associated with severe infection, patients treated with corticosteroids and cyclophosphamide as induction therapy tended to have had more severe infections according to the French Vasculitis Study Group.39,42 The role of rituximab has also been extensively studied in recent years, with two major RCTs showing that serious infections rates were similar between rituximab and the groups including patients treated with iv cyclophosphamide.35,36 Furthermore, concerns regarding corticosteroids adverse effects have led to the development of steroid-sparing protocols with good results, namely reduced risk of infectious complications and without an increase in progression to end-stage renal disease or mortality.43,44

Therefore, corticosteroids sparing regimens, rituximab over cyclophosphamide or mixed protocols where cyclophosphamide is used in reduced cumulative doses are currently used and being developed to achieve a mainstay in therapeutic options, providing the same efficacy achieved with previous regimens but at a reduced adverse events rate. In our cohort, no differences were found between treatment regimens and outcomes, nevertheless the small sample may have resulted in these results.

Mortality also remains an important issue in patients with AAV. Despite important improvements after initiation of corticosteroids and cytotoxic therapy as noted above, an important portion of patients with AAV still decease as a consequence of disease activity or more frequently as consequence of severe infection in the first year.45 On the long term, cardiovascular disease, malignancy and mortality account for the major causes of death in this population.7

Age, secondary infection, initial renal function and pulmonary involvement of AAV act as independent predictors of mortality, as reported in a retrospective study with almost 400 patients.46 Higher BVAS, lower hemoglobin and higher white cell count at presentation were also described as negative outcome predictors in AAV patients.7

Since early mortality plays an important role in this pathology, identifying predictors of premature mortality may help in stratifying those who may be in need for intensive care unit (ICU) admission and more tailored therapy.47 Acute kidney injury, smoking, hemoptysis and respiratory failure were all predictors of ICU admission, with those in need for blood transfusions and mechanical ventilation being at increased risk of early mortality, within 28 days from admission.48 Although multiple clinical disease activity parameters significantly correlate with one-year mortality, a high FFS and the occurrence of severe infection during the first month correlate better and constitute the most solid predictors of one-year mortality.49

In our study, the NLR was a strong predictor of infection. Indeed, a NLR ≥4.04 is associated with a significant risk of severe infection at 3 months. More importantly, assessment of this score emphasizes the low risk of developing severe infection in patients who have NLR <4.04, which could guide therapy decision. Nevertheless, infection was not associated with mortality in the multivariate analysis, which could be a result of the reduced number of patients in our population. Therefore, further studies are required to corroborate the NLR as a predictor of severe infection and mortality in AAV patients.

We address some of the important limitations of this study. Firstly, the single-center nature of our study limits generalizability. Secondly, the retrospective design and the small cohort of patients could have contributed to disregarding some potential confounders with prognostic importance. Thirdly, the NLR was not age-adjusted, which might have overestimated the importance of the ratio in older patients, who are prone to exhibit a pro-inflammatory state. Finally, we did not discriminate infection source and did not address cause-specific mortality as well.

Despite these limitations, our study has several notable strengths. To the best of our knowledge, this is the one of the first studies evaluating the association between the NLR and prognosis in AAV patients. We demonstrated the clinical significance of the NLR at diagnosis in estimating the prognosis of AAV patients. Indeed, the high specificity of the NLR reflects that a high ratio is predictive of severe infection within 3 months. Also, we provided an optimal cutoff of NLR to predict severe infection.

We highlight the need for further understanding of AAV in order to identify at-risk patients, guide immunossupression and ultimately improve patient outcomes.

In conclusion, we demonstrated that the NLR at admission was independently associated with development of severe infection within 3 months of starting induction therapy in AAV patients. This ratio can be quickly assessed from a routine blood sample and can prove useful in identifying patients at risk of infection. Recognizing high risk patients is crucial to guide immunosuppression and improve patient outcomes.

The study was approved by the Ethical Committee at the Centro Hospitalar Universitário Lisboa Norte, EPE, in agreement with institutional guidelines. Informed consent was waived by the Ethical Committee due to the retrospective and non-interventional nature of the study.

Not applicable.

Please contact author for data requests.

No funding was received for this study.

JAF and JG made substantial contributions to the study concept and design, analysis and interpretation of data, and were involved in drafting the manuscript and revising it critically for important intellectual content. JAF, ID, JG participated in the acquisition of data. SJ and JAL were involved in the critical revision of the manuscript and approval of the final version to be submitted.

There is no conflict of interest. The results presented in this paper have not been published previously in whole or part, except in abstract format.