Role of Hematological Indices as Predictors of Systemic Inflammation in Dermatology

Abstract

Hematological indices have long been used in a variety of conditions as indicators of systemic inflammation. The neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) are inexpensive tools in predicting severity and disease activity in acute coronary syndrome, malignancy, coronavirus disease 2019 (COVID-19) illness, and other autoimmune conditions. Evidence suggests that such hematological indices can be used in various dermatoses as predictors of systemic inflammation, to assess the severity of the disease, prognosis, and treatment response.

Introduction

Neutrophils account for the initial nonspecific immunity in systemic inflammation. The lymphocytes act protectively by regulating the inflammatory immune response. The neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and a few other hematological indices are simple, easily obtainable inexpensive indicators of systemic inflammation. Increased neutrophils, accompanied by reduced lymphocytes, are commonly seen in cardiovascular diseases, malignancies, cystic fibrosis, and stress. Sepsis may be associated with lymphopenia due to increased lymphocyte apoptosis.[1] Mean platelet volume (MPV) is a routinely used determinant of platelet size. An increase in platelet production is associated with a reduction in MPV. Alterations in the proportions of platelets are associated with inflammation, abnormal synthesis, increased wear, and due to the effect of activating factors. Platelets with higher mean volume (>15 fl) have high reactivity that undergoes rapid activation and thrombosis.[2]

The NLR is a simple and effective method for assessing the prognosis and predicting mortality in acute coronary syndrome, malignancies, and coronavirus disease 2019 (COVID-19). In this review, we discuss the utility of such hematological indices in various dermatoses as predictors of systemic involvement and their significance [Table 1].

Table 1.

Significance of hematological inflammatory markers in various dermatoses

Dermatological condition	Significance of inflammatory markers
Henoch–Schonlein purpura	Elevated NLR is suggestive of systemic (renal and GI) involvement
Kawasaki disease	Raised pretreatment NLR, PLR, and CRP serve as indicators of IVIG-resistant disease
Giant cell arteritis	Elevated PLR is associated with the disease
Systemic lupus erythematosus	NLR and PLR are useful markers of disease activity and predictors of lupus nephritis
Dermatomyositis	Elevated NLR and PLR predict ILD and serve as markers of disease activity. Elevated NLR determines mortality due to associated ILD in DM/PM and CADM
Systemic sclerosis	Raised NLR and PLR predict digital ulcers and ILD and MLR marks disease activity
Sjögren’s syndrome	NLR and RDW correlate with disease activity. PLR correlates with disease activity and neurologic involvement
Behçet’s disease	NLR and PLR correlate with disease activity. NLR, PLR, and MPV predict systemic (GI, joint, and ocular) involvement.
Chronic urticaria	Posttreatment reduction in NLR and PLR, and elevation in MPV indicate the therapeutic efficacy of omalizumab
Erythema nodosum	Elevated NLR marks the disease and differentiates the secondary from the idiopathic disease
Hidradenitis suppurativa	LHR, NHR, and MHR are elevated in the disease. MHR correlates with disease duration. CRP, MCV, and RDW correlate with disease severity
Vitiligo	MHR, MLR, and RDW serve as indicators of systemic inflammation and oxidative stress. MLR and MHR correlate with disease activity and duration
Psoriasis	Elevated NLR and PLR mark the disease and may correlate with disease severity. They also predict psoriatic arthritis
Pemphigus vulgaris	NLR and PLR correlate positively, while MPV and PNR correlate negatively with the disease activity. NLR and PLR also predict early relapse
Bullous pemphigoid	NLR and PLR mark the disease activity
Atopic dermatitis	NLR and PLR mark systemic inflammation and correlate with the disease activity and duration. NLR also differentiates the intrinsic disease from the extrinsic type
Erythema nodosum leprosum	NLR and PLR serve as diagnostic indicators of the disease with good sensitivity, specificity, and diagnostic accuracy
Condyloma acuminata	NLR and PLR are associated with greater wart size and SIN transformation. PLR is also a risk determinant of greater wart size, recurrence, and high-grade SIN
Cutaneous metastatic melanoma	NLR serves as the prognostic marker of overall survival
Alopecia areata	CRP is useful as a diagnostic marker of the disease
Familial Mediterranean fever	NLR marks disease activity and determines subclinical inflammation in asymptomatic patients
PFAPA syndrome	NLR and PLR mark disease activity and low MPV determines subclinical inflammation in the asymptomatic phase
Adult-onset still disease	Elevated NLR and PLR and reduced MPV are associated with the disease.

NLR=neutrophil-to-lymphocyte ratio, GI=gastrointestinal, PLR=platelet-to-lymphocyte ratio, CRP=C-reactive protein, IVIG=intravenous immunoglobulin, ILD=interstitial lung disease, DM/PM=dermatomyositis/polymyositis, CADM=clinically amyopathic dermatomyositis, MLR=monocyte-to-lymphocyte ratio, RDW=red cell distribution width, MPV=mean platelet volume, LHR=lymphocyte-to-high-density lipoprotein ratio, NHR=neutrophil-to-high-density lipoprotein, MHR=monocyte-to-high-density lipoprotein ratio, MCV=mean corpuscular volume, PNR=platelet-to-neutrophil ratio, SIN=squamous intraepithelial neoplasia, ELR=eosinophil-to-lymphocyte ratio, LDH=lactate dehydrogenase, PFAPA=periodic fever, aphthous stomatitis, pharyngitis, and adenitis

Cutaneous vasculitis

Henoch–Schonlein purpura (HSP)

HSP is a common cutaneous vasculitis exhibiting palpable purpura, arthritis, and abdominal pain. Extracutaneous manifestations indicate severe disease and poor prognosis. Lei et al. observed a correlation between NLR and disease severity in HSP cases with renal and/or gastrointestinal (GI) involvement and suggested that it can predict systemic involvement. They also observed that patients with elevated NLR without organ involvement initially revealed positive urine/stool findings subsequently.[3] Kim et al. observed female gender and elevated NLR as risk indicators for kidney involvement in HSP.[4] Jaszczura et al. evaluated IgA vasculitis in children and found NLR as a valuable tool in recognizing patients at risk of systemic involvement.[5] An Indian study by George et al. found elevated NLR being significantly associated with GI manifestations and serving as a predictor of systemic involvement in IgA vasculitis.[6]

Kawasaki disease (KD)

KD is an acute febrile disease of childhood featuring medium-sized arterial vasculitis. Coronary artery involvement occurs in about a quarter of untreated cases. Intravenous immunoglobulin (IVIG) is the cornerstone in the treatment of acute KD. Patients with IVIG-resistant disease apparently pose a greater risk of coronary artery disease. There are many risk scoring systems available to predict IVIG resistance utilizing common laboratory parameters. In regard to hematological indices as predictors of IVIG-resistant KD, Takeshita et al. observed that the raised pretreatment levels of NLR and PLR independently predicted IVIG resistance.[7] Chang et al. retrospectively reviewed KD patients and designed a scoring system to predict coronary artery involvement in children. It included elevated C-reactive protein (CRP) (>103 mg/L), elevated NLR (>3.5), male gender, and IVIG resistance.[8] A meta-analysis in assessing the of role NLR as a marker to recognize IVIG-resistant KD revealed that NLR estimation after initial IVIG treatment might be a detector of IVIG-resistant KD.[9]

Giant cell arteritis (GCA)

GCA is primarily a disease seen in the elderly featuring large vessel vasculitis. The temporal artery is frequently involved, and biopsy of the same is the gold standard in diagnosing GCA. Among the suspected GCA cases subjected to temporal artery biopsy, Dohn et al. found PLR being significantly elevated in GCA than in non-GCA patients, which could be considered as an additional diagnostic feature of GCA.[10]

Connective tissue diseases

Systemic lupus erythematosus (SLE)

SLE is an autoimmune connective tissue disease affecting multiple organs. Conventionally, antinuclear antibody screening and autoantibody profiling are employed in the diagnosis and monitoring of SLE. A positive correlation between NLR and CRP, erythrocyte sedimentation rate (ESR), and SLE disease activity index (SLEDAI) was noted by Qin et al. They also observed NLR and PLR being significantly elevated in SLE cases with nephritis and demonstrated a correlation between the presence of Sjögren’s-syndrome-related antigen A (SSA) and Sjögren’s-syndrome-related antigen B (SSB) antibodies and PLR.[11] El-said et al. noted significantly higher PLR in SLE patients compared with controls.[12] A meta-analysis by Ma et al. found significantly higher NLR and PLR in SLE than in controls, which also positively correlated with SLEDAI.[13] Hence, such hematological indices can serve as simple and inexpensive measures in assessing and predicting SLE disease activity.

Dermatomyositis

Dermatomyositis is an autoimmune inflammatory myopathy associated with typical cutaneous manifestations and/or internal organ involvement. The diagnosis and severity assessment of dermatomyositis is essentially based on the estimation of muscle enzymes and autoantibody profiling. Yang et al. observed significantly raised NLR and PLR in dermatomyositis cases without interstitial lung disease (ILD) than in controls, as well as higher NLR and PLR in dermatomyositis cases with ILD than in those without. Furthermore, they also noted NLR positively correlating with CRP and ESR and the global activity scores significantly and positively correlating with CRP, NLR, and PLR.[14] In a retrospective study to characterize the factors associated with one-year mortality in clinically amyopathic dermatomyositis with acute ILD, Zou et al. noted that NLR, serum ferritin, serum lactate dehydrogenase, and high-resolution computed tomography (HRCT) scores were statistically significant risk factors. The overall HRCT scores and serum ferritin were independent risk factors on multivariate analysis and were also significant for categorizing non-survivors.[15] In another retrospective study, Ha et al. observed elevated NLR (and CRP-to-albumin ratio) being significantly associated with non-survivors with polymyositis/dermatomyositis-associated ILD. The mortality rate was also noted to be higher in such patients exhibiting elevated NLR (and CRP-to-albumin ratio) above the optimal cutoff values. Furthermore, elevated NLR was also one among the independent predictors of mortality. Therefore, it appears that baseline estimation of NLR may be a simple and inexpensive prognostic indicator in dermatomyositis.[16]

Systemic sclerosis (SSc)

SSc is an autoantibody-mediated collagen vascular disease exhibiting autoimmunity-induced inflammation, vascular endothelial damage, and tissue fibrosis. Cutaneous sclerosis, digital ulcers, and pulmonary involvement are predominant features of the disease. In a study to analyze the role of NLR and PLR in SSc, Kim et al. observed both indices being significantly elevated in patients than in healthy controls. Furthermore, PLR and NLR were also raised in patients with digital ulcers and ILD, respectively, than in those without these features. Multivariable logistic regression models showed PLR being independently associated with digital ulcers and both the PLR and NLR being independently associated with ILD.[17] A cross-sectional study concluded that NLR and monocyte-to-lymphocyte ratio (MLR) could be associated with cutaneous and vascular manifestations of SSc and also with the disease severity and activity, with the monocyte count in itself being an independent predictor of high disease activity and severity.[18]

Sjögren’s syndrome

Sjögren’s syndrome is an autoantibody-mediated collagen vascular disease characterized by dry mouth and dry eyes. Yildiz and Gökmen noted NLR being significantly elevated in patients of primary Sjögren’s syndrome than the controls. They also evaluated the European Alliance of Associations for Rheumatology (EULAR) Sjögren’s syndrome disease activity index (ESSDAI) in the patients and observed that the scores were significantly raised in those with neurological involvement and positively correlated with PLR.[19] Similarly, Hu et al. noted increased red cell distribution width (RDW) and NLR in the cases correlated positively with the disease activity as indicated by the Sjögren syndrome disease activity index.[20]

Behçet’s disease

Behçet’s disease is a vasculitic disease showing recurrent genital/oral aphthae and uveitis. It is also associated with various types of skin lesions and severe disease may exhibit life-threatening neurological and vascular events. The diagnosis of Behçet’s disease is essentially clinical, and there are no specific serologic biomarkers for diagnosis or assessing disease activity. Hammad et al. noted significantly raised NLR in active compared to inactive disease. Both PLR and NLR correlated with Behçet’s disease current activity form with the NLR showing stronger correlation. The NLR with mucocutaneous manifestations and both PLR and NLR with gastrointestinal and joint manifestations were significantly associated as well. No correlation between the indices and serious manifestations (ocular and neurological) were seen.[21] Avci et al. noted significantly raised NLR, PLR, and MPV in Behçet’s disease with anterior uveitis than in those without and in healthy controls. Among the three, NLR seemed to be superior as a predictor of ocular involvement.[22] Therefore, NLR seems useful in determining the disease activity and organ involvement in Behçet’ disease.

Influence of omalizumab on hematological markers in chronic spontaneous urticaria (CSU)

CSU refers to recurrent itchy wheals occurring daily or almost every day for at least six weeks. The NLR and PLR are known determinants of inflammation in CSU. Acer et al. retrospectively assessed pretreatment and posttreatment values of various hematological parameters in CSU patients receiving omalizumab. They noted a significant reduction in NLR and PLR along with several other parameters and an increase in eosinophil count and MPV at three months posttreatment.[23] Altrichter et al.[24] noted a reduction in interleukin-31 levels and white blood cells following omalizumab therapy. Apart from the anti-IgE effect, omalizumab also appears to influence the other inflammatory markers and their monitoring may be indicative of therapeutic response/efficacy of omalizumab in CSU.

Erythema nodosum (EN)

EN is a reactive inflammatory disorder of the subcutis characterized by tender red subcutaneous nodules preferentially involving the lower extremities. It may be precipitated by various infectious and non-infectious diseases as well as by certain drugs (secondary EN). In certain instances, no identifiable precipitating factor can be found (idiopathic EN). Hayran et al. revealed significantly elevated NLR in all EN patients than in controls. Further, elevated NLR (along with other hematological indices) was identified as a risk factor for secondary EN, which may differentiate it from idiopathic EN.[25]

Hidradenitis suppurativa

Hidradenitis suppurativa is a chronic, relapsing inflammatory disorder affecting the apocrine glands exhibiting tender nodules, abscesses, and draining sinuses, especially in the intertriginous areas. The Hurley staging system is widely used to assess the clinical severity of the disease. Çetinarslan et al. noted significantly raised lymphocyte-to-high-density lipoprotein ratio (LHR), neutrophil-to-HDL ratio (NHR), and monocyte-to-HDL ratio (MHR) in patients than in controls, with the MHR positively correlating with disease duration. However, none of these indices (including NLR and PLR) positively correlated with the Hurley stating. Other common indices such as mean corpuscular volume (MCV), CRP, and RDW, however, correlated with disease severity. Hence, the authors concluded that CRP is more valuable than the newer hematological indices and MHR is probably the more useful among the newer indices in hidradenitis suppurativa.[26]

Vitiligo

Vitiligo is a disorder manifesting as varying degrees of cutaneous/mucosal depigmentation and can cause significant stress and stigmatization. Oxidative stress and autoimmunity in genetically susceptible individuals contribute to disease induction and progression. Both these factors contribute to systemic inflammation targeting the melanocytes.[27] Dermirbas et al. opined that MHR and MLR can indicate oxidative stress and systemic inflammation in vitiligo as they noted significantly higher values of MHR, PLR, MLR, and plateletcrit in patients as compared to the controls together with a correlation between MHR and MLR levels, disease duration, and disease severity as per the vitiligo extent tensity index score.[28] Similarly, Sarac et al. opined that RDW may serve as an indicator of systemic inflammation vitiligo.[29]

Psoriasis

Psoriasis is a common relapsing and remitting inflammatory papulosquamous disorder, frequently affecting the extensor aspects of the extremities and scalp. The pathogenesis involves a complex interaction between antigen-presenting cells, T-lymphocytes, and others that incite an inflammatory response in the epidermis. Psoriatic arthritis (PsA) is known to be associated with about 20% of patients. Kim et al. evaluated NLR and PLR (together with cellular counts and ESR) values in cases of psoriasis and PsA. Both were significantly elevated in PsA than in psoriatic and healthy individuals and also significantly predicted the presence of PsA among the patients with psoriasis. Among the psoriatic patients, the PLR and NLR correlated positively with psoriasis area severity index scores.[30] Pooled results of a systematic review and meta-analysis showed the PLR and NLR were significantly raised in psoriasis patients, but they only showed association with the presence, but not with the severity of psoriasis.[31] As to the NLR and PLR as determinants of treatment response, An et al. noted a statistically significant reduction in NLR and PLR (as well as MPV and CRP) levels following biological therapy in psoriasis and opined that these indices could be used in the assessment of treatment response with biologics and in monitoring the disease course.[32] Based on the association between atherosclerosis (due to platelet activation), psoriasis, and PsA, Canpolat et al. assessed the differences in MPV values among psoriasis patients, PsA, and healthy controls. They noted MPV being significantly elevated in psoriasis patients compared to the controls along with statistically higher levels in patients with arthritis than those without. Further, the levels also correlated with disease activity and duration.[33] In a case–control study, Mondal et al. observed elevated NLR, PLR, and CRP among the psoriasis patients as compared to the controls. However, only CRP levels showed a statistically significant difference between the two groups. Values of all the three indices were also higher among the patients with joint involvement than those without, but without significant difference. In relation to disease severity, a significant correlation between CRP and psoriasis area severity index (PASI) as well as between PLR and PASI was observed.[34]

Autoimmune bullous disorders

Pemphigus vulgaris and bullous pemphigoid are chronic remitting and relapsing autoimmune bullous disorders. Factors determining the relapse of these disorders remain mostly unelucidated.[35] Low CD4 T-cell count, B-cell repopulation, and high serum antibodies levels against desmoglein 1 (Dsg1) and desmoglein 3 (Dsg3) have been described as predictors of relapse in pemphigus vulgaris following rituximab therapy.[36] However, such an assessment is routinely not possible. Lyakhovitsky et al. opine that the NLR, PLR, MPV, and platelet-to-neutrophil ratio (PNR) can indicate treatment response with the NLR and PLR also serving as detectors of early relapse in pemphigus vulgaris. This was based on their observation of the NLR and PLR correlating positively, and the MPV and PNR correlating negatively with the disease activity, with the NLR showing a more significant correlation in patients not treated with rituximab.[37] Lin and Li observed NLR and PLR (together with monocyte count and antibodies against Dsg 1 and Dsg 3) correlating positively with the pemphigus vulgaris disease activity with the NLR, low albumin-to-globulin ratio, and high anti-Dsg1 antibody levels being associated with pemphigus area disease index scores. The authors believe these findings could be more comprehensive and objective reflectors of the pemphigus vulgaris disease severity.[38]

Sun et al. noted a positive correlation between bullous pemphigoid disease activity and the NLR and PLR. Both these indices correlated with the bullous pemphigoid disease area index (BPDAI) without damage score and the BPDAI total score, while the NLR correlated positively with BPDAI erosion/blister scores.[39]

Atopic dermatitis (AD)

AD is an inflammatory chronic dermatosis characterized by pruritic rash commonly affecting children, though a significant number of adults also suffer from the disease. Skin barrier dysfunction, impaired immune surveillance, and allergy contribute to the inflammation leading to AD. The diagnosis of AD is essentially clinical, and clinical scorings such as scoring AD (SCORAD) and eczema area severity index are employed in assessing the disease severity. To explore the role of hematologic indices as determinants of systemic inflammation in AD, Jiang and Ma found the NLR and PLR values correlating positively with SCORAD reflecting inflammation and disease severity in AD patients.[40] Batmaz noted higher PLR and NLR levels among the cases than in controls with the NLR also being higher in patients with intrinsic than in the extrinsic AD. Both the indices also correlated with disease duration.[41]

EN leprosum (ENL)

ENL, or the type 2 lepra reaction, is an immune complex-mediated phenomenon seen in lepromatous leprosy characterized by acute onset generalized, transient, tender erythematous nodules frequently accompanied by constitutional symptoms. Tanojo et al. noted high specificity and sensitivity of PLR and NLR in diagnosing ENL together with good diagnostic accuracy. Because of the lack of any standardized diagnostic criteria, these hematological indices could potentially serve as diagnostic markers of ENL.[42]

Condyloma acuminata

Condyloma acuminata refers to human papillomavirus (HPV)-induced genital epithelial hyperplastic lesions. Low-, intermediate-, and high-risk HPV strains (in terms of squamous intraepithelial neoplasia (SIN) transformation) are implicated in the causation. In a retrospective study to determine the association between peripheral inflammatory biomarkers (NLR, PLR, and systemic immune-inflammation index (SII)) and recurrence/SIN transformation in surgically treated patients of condyloma acuminata, Basim and Yuksel noted that all the biomarkers were raised in those with the greatest wart size and SIN transformation. Further, high-risk HPV strains were associated with higher SII. A significant association between recurrence and larger wart size, high-grade SIN, and higher PLR and SII was noted as well. Based on these observations, they suggest that PLR and SII can be risk predictors (in terms of recurrence and SIN transformation) in condyloma acuminata.[43]

Cutaneous metastatic melanoma

Khoja et al. noted NLR and lactate dehydrogenase levels (before and during treatment) allowed categorization of cutaneous metastatic melanoma cases treated with ipilimumab into good, intermediate, and poor prognosis classes as they were found to be indicators of overall survival on multivariable analysis. They suggest these biomarkers may be relevant in the management of such cases.[44]

Lichen planus (LP)

LP is a chronic inflammatory papulosquamous disorder. There have been mixed results in regard to the role of hematologic inflammatory markers in this disease. Ataş et al. noted a significantly higher NLR (along with ESR and CRP) in LP as against the controls.[45] Ozlu et al. found significantly higher platelet distribution width and MHR but low MPV in LP cases compared to healthy controls.[46] Conversely, An et al. did not find any difference between the PLR and NLR values in the case and control groups, but MPV was higher among the LP cases. None of these indices showed any association with nail and oral involvement.[47]

Alopecia areata (AA)

AA is a T-cell-mediated disorder characterized by localized, extensive, or universal noncicatricial alopecia. Limited data on the levels of hematological parameters suggest that apart from CRP, many of the hematological indices are not useful as markers of the disease.[48,49]

Auto-inflammatory syndromes

Auto-inflammatory syndromes are a group of disorders occurring due to genetically mediated dysregulation in innate immunity that leads to increased activation of inflammatory mediators. The common clinical manifestations among these disorders include recurrent fever and inflammatory mucocutaneous, visceral, and joint manifestations without an underlying autoimmune or infectious cause.[50] Instances of these disorders wherein the role of hematological indices has been studied are discussed below.

Familial Mediterranean fever (FMF)

FMF is an autosomal recessive disorder commonly seen in the Mediterranean population. Celikbilek et al. observed significantly elevated NLR among active FMF patients than in controls and those in remission. They identified an optimal cutoff value of 2.63 with an overall accuracy of 71% to determine active FMF.[51] Özer et al. identified a cutoff value for NLR of 1.65 for the determination of subclinical inflammation in asymptomatic patients. This was based on their observation of significant elevation in several hematological indices (NLR, PLR, MVP, and RDW) in asymptomatic FMF cases compared to controls, with the NLR showing the strongest correlation with CRP.[52] Similarly, Dincer et al. observed elevated NLR in patients with subclinical inflammation with an optimal cutoff value of 2.94 for predicting the same.[53]

Periodic fever, aphthous stomatitis, pharyngitis, and adenitis (PFAPA) syndrome

Also known as Marshall syndrome, PFAPA syndrome is the most common auto-inflammatory syndrome of childhood. Gungorer et al. noted significantly higher NLR and PLR in patients during active disease compared to asymptomatic patients and controls. Further, a significantly lower MPV in asymptomatic patients compared to controls was also observed indicating continued subclinical inflammation in attack-free phases of the disease.[54]

Adult-onset Still disease

Adult-onset Still disease (AoSD) is a nonfamilial auto-inflammatory syndrome that can have a self-limited, intermittent, or chronic course. Ulutas et al. noted elevated NLR and PLR values and reduced MPV values in patients of AoSD without any difference among the subgroups categorized based on the disease course. They also noted significantly higher NLR and CRP values in active than inactive disease.[55] A prospective multicenter study noted NLR ≥4 in majority of the AoSD patients with high sensitivity and specificity. Based on which, the authors found its inclusion as a major criterion improving the sensitivity and accuracy of the Yamaguchi and Fautrel classification systems.[56]

Conclusion

Hematological parameters and indices serve as simple, inexpensive, and ubiquitous tools in determining systemic inflammation in various dermatoses reflecting the disease severity and prognosis. They also help in assessing responses to certain treatment modalities. An awareness of such roles of various hematologic indices in dermatology is imperative for dermatologists. Needless to say, the evidence will only grow further with more studies being carried out in reinforcing the existing evidence as well as to explore the role of various hematological indices in other dermatoses.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.