Systemic inflammation in optic neuropathies: comparative insights from hematological and composite indices

Abstract

Purpose

To compare complete blood count (CBC)-derived inflammatory indices among patients with multiple sclerosis-associated optic neuritis (MS-ON), non-arteritic anterior ischemic optic neuropathy (NAION), idiopathic intracranial hypertension-related optic neuropathy (IIH-ON), and healthy controls, while accounting for demographic confounding.

Methods

This retrospective study included 135 participants: 36 patients with MS-ON, 29 with NAION, 34 with IIH-ON, and 36 healthy controls. Six composite inflammatory indices, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), pan-immune-inflammation value (PIV), and hemoglobin-albumin-lymphocyte-platelet (HALP) score, were calculated from routine laboratory measurements obtained at diagnosis before treatment. Log-transformed indices were compared using analysis of covariance adjusted for age and sex. False discovery rate correction was applied across the six indices. Receiver operating characteristic (ROC) analysis was performed for indices showing significant adjusted group effects.

Results

Significant differences in age and sex distribution were observed among the groups. After adjustment for age and sex, overall group effects remained for NLR (p = 0.008) and SIRI (p = 0.033), although only NLR remained significant after false discovery rate correction (q = 0.047). Bonferroni-adjusted pairwise comparisons demonstrated higher NLR in MS-ON than in healthy controls (adjusted geometric means 2.83 vs. 1.80, p = 0.005). NLR discriminated MS-ON from controls with an area under the ROC curve (AUC) of 0.681 (95% CI, 0.554–0.807), while SIRI showed a comparable AUC of 0.661 (95% CI, 0.535–0.787). Among conventional hematological parameters, red cell distribution width remained elevated in IIH-ON, whereas serum albumin remained lower in MS-ON and IIH-ON after adjustment.

Conclusions

In the primary analysis, NLR was the only index that retained significance after adjustment for age, sex, and multiple testing, with higher values in MS-ON than in healthy controls. The sensitivity of several other indices to robust inference and the modest ROC performance warrant cautious interpretation and prospective validation.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12886-026-05433-0.

Introduction

Optic neuropathies can cause visual loss through inflammation, ischemia, or sustained optic disc swelling [1]. Multiple sclerosis-associated optic neuritis (MS-ON) involves inflammatory demyelination; non-arteritic anterior ischemic optic neuropathy (NAION) is associated with impaired optic nerve head perfusion; and idiopathic intracranial hypertension-related optic neuropathy (IIH-ON) develops in the setting of papilledema caused by elevated intracranial pressure. These mechanisms differ, but peripheral blood markers may reflect accompanying systemic responses. An association between NAION and systemic inflammation has been proposed [2]. Whether blood-derived inflammatory indices show distinct patterns across these three conditions after accounting for demographic differences remains unclear.

In recent years, a variety of systemic inflammatory biomarkers derived from routine complete blood count (CBC) parameters have gained attention as accessible and cost-effective indicators of immune and inflammatory status. Composite indices such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), pan-immune-inflammation value (PIV), and hemoglobin-albumin-lymphocyte-platelet (HALP) score have been investigated in a wide range of neurological, cardiovascular, oncological, and ophthalmic diseases [3–5]. These markers are thought to reflect the balance between pro-inflammatory and anti-inflammatory pathways, immune activation, and overall systemic inflammatory burden. Although individual inflammatory indices have been evaluated in a few optic nerve disorders lately, comparative data across different types of optic neuropathies remain limited [6–9]. Moreover, it is unclear whether these recently proposed composite inflammatory biomarkers differ meaningfully between inflammatory, pressure-related, and ischemic optic neuropathies, or whether they may help elucidate shared or distinct inflammatory profiles underlying these conditions. Identifying such differences could enhance understanding of disease mechanisms and potentially support risk stratification or adjunctive diagnostic evaluation.

Therefore, the aim of the present study was to compare systemic inflammatory biomarkers derived from CBC and serum albumin measurements among patients with MS-ON, IIH-ON, and NAION, as well as healthy controls. By examining multiple hematological indices within a standardized retrospective framework, we sought to determine whether these markers differ between optic neuropathy subtypes and to explore their potential relevance in neuro-ophthalmic disease assessment.

Materials and methods

Study design and population

This retrospective observational study was conducted at a single tertiary referral center and included patients diagnosed with MS-ON, IIH-ON, or NAION, together with a group of healthy individuals who served as controls. Medical records of patients evaluated between January 2005 and January 2025 were reviewed. Ethical approval was obtained from the Dokuz Eylul University Ethics Committee (decision number 2025/42 − 09, dated 1 December 2025). The study was conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was waived owing to the retrospective design.

Patients were eligible for inclusion if they had a confirmed diagnosis of one of the specified optic neuropathies and had a CBC and serum albumin measurement obtained at the time of diagnosis, prior to the initiation of any treatment. Control participants were selected from individuals attending the ophthalmology clinic for routine examination who had no documented ocular or systemic disease and for whom CBC and serum albumin results were available from a preoperative assessment. No laboratory tests were ordered specifically for this retrospective study.

Exclusion criteria comprised a refractive error greater than ± 3.0 diopters, a history of intraocular surgery, inflammatory ocular disease, or any other concomitant ocular pathologies. Patients with high refractive error were excluded because optic nerve head appearance and retinal nerve fiber layer assessment may be altered in such eyes, which could affect diagnostic classification. Participants were also excluded if they had received corticosteroids or any other systemic treatment capable of influencing hematological parameters. Additional exclusion criteria comprised the presence of active infection, chronic inflammatory or autoimmune disease, malignancy, or any systemic condition or medication use at the time of diagnosis that could alter inflammatory markers.

A total of 99 patients (36 with MS-ON, 29 with NAION, and 34 with IIH-ON) and 36 healthy controls were included, giving 135 participants in the final analysis.

Diagnostic criteria

Participants were retrospectively classified according to contemporary diagnostic criteria for each optic neuropathy subtype. NAION was diagnosed in the presence of acute, painless monocular visual loss accompanied by optic disc edema, a relative afferent pupillary defect, and characteristic visual field defects. Cases classified as arteritic anterior ischemic optic neuropathy were not included. Diabetes and hypertension were not categorical exclusion criteria. At presentation, none of the included NAION patients had known uncontrolled diabetes or hypertension or fundus signs of diabetic or hypertensive retinopathy. IIH-ON required a diagnosis of idiopathic intracranial hypertension according to the revised Friedman criteria [10] together with papilledema and documented optic nerve dysfunction, evidenced by reduced visual acuity and/or a corresponding visual field defect. The ophthalmological evaluation included visual acuity, fundus examination, optical coherence tomography of the optic nerve head and peripapillary retinal nerve fiber layer, and fluorescein angiography when clinically indicated to distinguish papilledema from other optic disc abnormalities [11]. MS-ON was defined as optic neuritis presenting as the first demyelinating event in patients who met the 2010 McDonald criteria [12] for multiple sclerosis during the subsequent diagnostic evaluation. None of the MS-ON patients was receiving disease-modifying therapy at the time of blood sampling.

Laboratory measurements and inflammatory indices

Venous blood samples were obtained at the time of diagnosis as part of routine clinical evaluation. For MS-ON and NAION, sampling occurred during the acute presenting episode, before treatment. For IIH-ON, samples were obtained at the diagnostic evaluation while papilledema and optic nerve dysfunction were present, before treatment; the onset of IIH-ON was not required to fall within a fixed acute-phase interval. Laboratory admission timestamps recorded in the electronic hospital information system were retrospectively reviewed; only samples collected between 08:30 and 10:30 were included to standardize morning sampling and increase the likelihood of fasting conditions. Hematological parameters comprised red blood cell count, white blood cell count, neutrophil, lymphocyte, monocyte and eosinophil counts, platelet count, hemoglobin concentration, red cell distribution width (RDW), and serum albumin levels.

Six composite indices were calculated from the raw CBC and albumin values. Because published values for several of these indices vary with the measurement units used, the units applied here are stated explicitly. Neutrophil, lymphocyte, monocyte and platelet counts were expressed in ×10³/µL, and hemoglobin and albumin in g/dL:

• NLR = neutrophils / lymphocytes.

• PLR = platelets / lymphocytes.

• SII = (neutrophils × platelets) / lymphocytes.

• SIRI = (neutrophils × monocytes) / lymphocytes.

• PIV = (neutrophils × platelets × monocytes) / lymphocytes.

• HALP = (hemoglobin × albumin × lymphocytes) / platelets.

Statistical analysis

Analyses were performed in Python 3.12 using the NumPy, SciPy, and pandas libraries. All tests were two-tailed, and p < 0.05 was considered statistically significant.

Continuous variables were assessed for normality using the Shapiro-Wilk test, histograms, and Q-Q plots, while homogeneity of variance was evaluated with Levene’s test. Because all six composite inflammatory indices were markedly right-skewed, natural logarithmic transformation was applied before inferential analyses. The effects of transformation on distributional assumptions are presented in Supplementary Table S1. Composite indices are reported as median (interquartile range) on the original scale, whereas hematological parameters are presented as mean ± standard deviation; untransformed values are provided in Supplementary Table S2.

Age and sex were compared using one-way analysis of variance and the chi-square test, respectively. Unadjusted group comparisons were performed using one-way analysis of variance with Tukey’s honestly significant difference test for pairwise comparisons. Because age and sex differed significantly between groups, the primary analyses used analysis of covariance with group and sex as fixed factors and age as a covariate. A group-by-sex interaction was not included because of the small number of male participants in the IIH-ON group. Adjusted group comparisons were based on estimated marginal means evaluated at the grand mean age and averaged across sexes; results for the log-transformed indices are presented as back-transformed adjusted geometric means, with Bonferroni correction for pairwise comparisons.

The six composite inflammatory indices were prespecified as the primary outcomes and corrected for multiple testing using the Benjamini-Hochberg false discovery rate procedure. Hematological parameters were analyzed as exploratory outcomes without correction for multiple comparisons. Three sensitivity analyses were pre-specified and are presented in Supplementary Table S3: HC3 robust inference, a model adjusting for age alone, and a repeat analysis restricted to female participants with adjustment for age, the last of these serving to eliminate any residual confounding by the marked sex imbalance between groups.

Receiver operating characteristic analysis was performed for those indices showing a significant adjusted group effect, comparing MS-ON patients with healthy controls. Areas under the curve, their 95% confidence intervals, and the associated p values were obtained using the DeLong method, and the two curves were compared using the DeLong test for correlated areas. Optimal cut-off values were identified by maximizing Youden’s index, with the corresponding sensitivity and specificity reported.

Results

Study population

A total of 135 participants were analyzed: 36 patients with MS-ON, 29 with NAION, 34 with IIH-ON, and 36 healthy controls. Baseline demographic characteristics are presented in Table 1. Age differed significantly among the four groups (p < 0.001). Sex distribution also differed significantly (p < 0.001), mainly because of the marked female predominance in the IIH-ON group (91.2%) compared with controls (50.0%), MS-ON (66.7%), and NAION (44.8%). Therefore, all subsequent analyses were adjusted for age and sex.

Table 1.

Baseline demographic characteristics of the study participants

Characteristic	Control (n = 36)	MS-ON (n = 36)	NAION (n = 29)	IIH-ON (n = 34)	p
Age, years	48.1 ± 9.2	38.2 ± 10.0	55.4 ± 6.4	38.1 ± 6.4	< 0.001
Sex, female	18 (50.0)	24 (66.7)	13 (44.8)	31 (91.2)	< 0.001
Sex, male	18 (50.0)	12 (33.3)	16 (55.2)	3 (8.8)

Data are mean ± SD or n (%). Age was compared using one-way ANOVA and sex distribution using the chi-square test

MS-ON, multiple sclerosis-associated optic neuritis; NAION, non-arteritic anterior ischemic optic neuropathy; IIH-ON, idiopathic intracranial hypertension-related optic neuropathy

Hematological parameters and serum albumin

Unadjusted comparisons of hematological parameters are summarized in Table 2. Significant overall group differences were observed for red blood cell count, white blood cell count, hemoglobin, neutrophil count, platelet count, RDW, and serum albumin (all p < 0.05). Pairwise comparisons indicated higher red blood cell counts in NAION, higher neutrophil counts in MS-ON and IIH-ON, higher platelet counts and RDW in IIH-ON, and lower serum albumin levels across all optic neuropathy groups than in healthy controls. No significant differences were observed for lymphocyte, monocyte, or eosinophil counts.

Table 2.

Hematological parameters and serum albumin across the study groups

Parameter	Control (n = 36)	MS-ON (n = 36)	NAION (n = 29)	IIH-ON (n = 34)	p (unadj.)	Post-hoc, unadjusted (Tukey)	p (adj.)	p (adj., HC3)	Post-hoc, adjusted (Bonferroni)
RBC (×10⁶/µL)	4.69 ± 0.39	4.50 ± 0.52	4.93 ± 0.54	4.52 ± 0.54	0.003	MS-ON vs. NAION: 0.004; NAION vs. IIH-ON: 0.008	0.189	0.219	—
WBC (×10³/µL)	7.19 ± 1.48	8.65 ± 2.73	8.61 ± 3.68	8.75 ± 2.32	0.040	—	0.058	0.004	—
Hemoglobin (g/dL)	13.81 ± 1.14	13.21 ± 1.77	13.70 ± 1.69	11.96 ± 1.75	< 0.001	Control vs. IIH-ON: <0.001; MS-ON vs. IIH-ON: 0.008; NAION vs. IIH-ON: <0.001	0.046	0.048	—
Neutrophils (×10³/µL)	4.14 ± 1.09	5.74 ± 2.73	5.47 ± 3.27	5.68 ± 1.87	0.014	Control vs. MS-ON: 0.023; Control vs. IIH-ON: 0.035	0.030	< 0.001	—
Lymphocytes (×10³/µL)	2.29 ± 0.57	2.09 ± 0.89	2.37 ± 0.75	2.28 ± 0.59	0.399	—	0.347	0.491	—
Monocytes (×10³/µL)	0.52 ± 0.14	0.56 ± 0.25	0.57 ± 0.23	0.57 ± 0.18	0.749	—	0.656	0.558	—
Eosinophils (×10³/µL)	0.17 ± 0.10	0.14 ± 0.13	0.13 ± 0.14	0.17 ± 0.16	0.531	—	0.312	0.365	—
Platelets (×10³/µL)	259.83 ± 48.60	253.69 ± 67.40	249.69 ± 72.37	316.68 ± 136.83	0.006	Control vs. IIH-ON: 0.038; MS-ON vs. IIH-ON: 0.017; NAION vs. IIH-ON: 0.016	0.057	0.138	MS-ON vs. IIH-ON: 0.038
RDW (%)	13.54 ± 0.59	14.25 ± 1.31	14.30 ± 1.57	15.28 ± 1.99	< 0.001	Control vs. IIH-ON: <0.001; MS-ON vs. IIH-ON: 0.018; NAION vs. IIH-ON: 0.039	< 0.001	< 0.001	Control vs. IIH-ON: <0.001
Albumin (g/dL)	4.45 ± 0.27	4.19 ± 0.40	4.20 ± 0.33	4.07 ± 0.27	< 0.001	Control vs. MS-ON: 0.005; Control vs. NAION: 0.011; Control vs. IIH-ON: <0.001	< 0.001	< 0.001	Control vs. MS-ON: 0.003; Control vs. IIH-ON: <0.001

Data are mean ± SD. Unadjusted comparisons: one-way ANOVA with Tukey HSD post-hoc tests. Adjusted comparisons: ANCOVA with group and sex as fixed factors and age as covariate; pairwise comparisons of estimated marginal means with Bonferroni correction

p adjusted (HC3) denotes heteroscedasticity-consistent p-values, reported for all parameters because between-group variances differed substantially for several variables (e.g. white blood cell count: variance ratio 6.2 between the control and NAION groups, Bartlett p < 0.001). Median-centred Levene’s test, which is deliberately insensitive to extreme observations, did not flag every such instance; robust p-values are therefore reported throughout rather than selectively

RBC, red blood cells; WBC, white blood cells; RDW, red cell distribution width. — indicates no significant pairwise comparison

After adjustment for age and sex, the group effect on red blood cell count was no longer significant (p = 0.189). Significant overall group effects remained for hemoglobin (p = 0.046), neutrophil count (p = 0.030), RDW (p < 0.001), and albumin (p < 0.001), whereas platelet count was no longer significant (p = 0.057). Pairwise comparisons showed higher platelet counts and RDW in IIH-ON than in MS-ON (p = 0.038) and healthy controls (p = 0.001), respectively, while albumin remained lower in MS-ON (p = 0.003) and IIH-ON (p < 0.001) than in controls. With HC3 robust inference, the overall WBC group effect became significant (p = 0.004 versus p = 0.058 in the primary model), whereas the platelet group effect remained nonsignificant (p = 0.138 versus p = 0.057); the remaining results are shown in Table 2.

Composite inflammatory indices

Distributional diagnostics for the six composite inflammatory indices are presented in Supplementary Table S1, with raw distributions in Supplementary Table S2. Because all indices were markedly right-skewed (Shapiro-Wilk p < 0.001), logarithmic transformation was applied for inferential analyses, whereas descriptive statistics are presented as medians (Q1-Q3).

Group comparisons are summarized in Table 3a, with adjusted geometric means shown in Table 3b. In unadjusted analyses, significant overall group differences were observed for NLR, SII, HALP, SIRI, and PIV (all p < 0.05), whereas PLR was not significant (p = 0.092). After adjustment for age and sex, only NLR (p = 0.008) and SIRI (p = 0.033) remained significant, although only NLR retained significance after Benjamini-Hochberg correction (q = 0.047). Bonferroni-adjusted pairwise comparisons identified a single significant difference, with higher NLR in MS-ON than in healthy controls (adjusted geometric means 2.83 vs. 1.80; p = 0.005).

Table 3a.

Composite systemic inflammatory indices across the study groups

Index	Control (n = 36)	MS-ON (n = 36)	NAION (n = 29)	IIH-ON (n = 34)	p (unadj.)	p (adj.)	p (adj., HC3)	q (BH-FDR)	Post-hoc, adjusted (Bonferroni)
NLR	1.71 (1.48–2.35)	2.45 (1.66–3.95)	2.21 (1.69–2.80)	2.38 (2.01–2.99)	0.003	0.008	< 0.001	0.047	Control vs. MS-ON: 0.005
PLR	115.2 (98.3–142.0)	123.9 (87.0–175.8)	111.9 (84.7–141.6)	124.7 (103.1–167.0)	0.092	0.396	0.420	0.396	—
SII	456.9 (355.3–609.0)	550.2 (373.3–1208.9)	569.5 (412.8–714.0)	637.7 (510.3–1032.5)	0.004	0.052	0.006	0.104	—
SIRI	0.92 (0.69–1.29)	1.24 (0.84–1.73)	1.05 (0.84–1.43)	1.41 (0.91–1.84)	0.014	0.033	0.006	0.100	—
PIV	248.9 (156.7–351.4)	303.1 (191.7–499.8)	267.1 (206.4–356.8)	345.8 (272.8–525.1)	0.013	0.085	0.021	0.127	—
HALP	0.53 (0.43–0.62)	0.42 (0.30–0.66)	0.55 (0.41–0.70)	0.37 (0.26–0.50)	0.002	0.106	0.029	0.127	—

Data are median (Q1–Q3) on the untransformed scale. All inferential analyses were performed on natural-log-transformed values. Adjusted analyses: ANCOVA with group and sex as fixed factors and age as covariate

q values were obtained by the Benjamini–Hochberg procedure applied across the six prespecified indices in the primary age- and sex-adjusted models. Only NLR retained significance in that primary analysis after false discovery rate correction. HC3 p-values are presented as sensitivity analyses

NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; SII, systemic immune-inflammation index; SIRI, systemic inflammation response index; PIV, pan-immune-inflammation value; HALP, hemoglobin-albumin-lymphocyte-platelet score

Table 3b.

Age- and sex-adjusted geometric means of the composite inflammatory indices

Index	Control	MS-ON	NAION	IIH-ON
NLR	1.80	2.83	2.21	2.52
PLR	115.1	131.0	107.1	132.8
SII	467.2	675.2	553.0	713.1
SIRI	0.92	1.39	1.23	1.36
PIV	238.6	330.6	307.5	383.8
HALP	0.53	0.43	0.53	0.38

Values are estimated marginal means from the log-scale ANCOVA models, back-transformed to the original scale, evaluated at the grand mean age and averaged equally over sex

Discriminatory performance

ROC analysis was performed for the indices showing a significant adjusted group effect, comparing MS-ON patients with healthy controls (36 vs. 36; Table 4; Fig. 1). NLR discriminated MS-ON from controls with an AUC of 0.681 (95% CI 0.554–0.807; p = 0.005). The optimal cut-off identified by Youden’s index was 2.59, corresponding to a sensitivity of 44.4% and a specificity of 94.4%. SIRI showed a comparable AUC (0.661, 95% CI 0.535–0.787; p = 0.013), and the difference between the two ROC curves was not significant (DeLong p = 0.637). Both indices showed modest discriminatory ability.

Table 4.

Receiver operating characteristic analysis for discriminating MS-ON (n = 36) from healthy controls (n = 36)

Index	AUC	95% CI	p	Cut-off	Sensitivity (%)	Specificity (%)	Youden index
NLR	0.681	0.554–0.807	0.005	2.59	44.4	94.4	0.389
SIRI	0.661	0.535–0.787	0.013	1.11	61.1	66.7	0.278

AUC, area under the curve. Confidence intervals and p-values were derived using the DeLong method; cut-off values were determined by Youden’s index

A DeLong test for correlated curves showed no significant difference between the two AUCs (p = 0.637)

Fig. 1.

Receiver operating characteristic (ROC) curves for (a) the neutrophil-to-lymphocyte ratio (NLR) and (b) the systemic inflammation response index (SIRI), distinguishing multiple sclerosis-associated optic neuritis (MS-ON; n = 36) from healthy controls (n = 36). NLR: area under the curve (AUC) 0.681 (95% CI 0.554–0.807; p = 0.005), optimal cutoff 2.59 (sensitivity 44.4%, specificity 94.4%). SIRI: AUC 0.661 (95% CI 0.535–0.787; p = 0.013), optimal cutoff 1.11 (sensitivity 61.1%, specificity 66.7%). The AUCs did not differ (DeLong p = 0.637). Confidence intervals and p values were obtained with the DeLong method; red diagonal lines represent AUC = 0.5. ROC analyses used unadjusted values and do not account for age and sex differences; SIRI did not retain significance after false discovery rate correction in the adjusted analysis (Table 3a)

Sensitivity analyses

Sensitivity analyses are summarized in Supplementary Table S3. With HC3 robust inference, NLR (p < 0.001) and SIRI (p = 0.006) remained nominally significant, while SII (p = 0.006), PIV (p = 0.021), and HALP (p = 0.029) became nominally significant; PLR remained nonsignificant (p = 0.420). Adjustment for age alone produced comparable findings, with nominally significant overall effects for NLR (p = 0.010), HALP (p = 0.023), SIRI (p = 0.049), and SII (p = 0.050). In the age-adjusted female-only analysis (n = 86), no significant group differences were observed, including for NLR (p = 0.113), likely reflecting the smaller sample size.

Discussion

This study compared conventional hematological parameters and six CBC-derived inflammatory indices across patients with MS-ON, NAION, IIH-ON, and healthy controls. After adjustment for age, sex, and multiple testing, only NLR remained significantly elevated in MS-ON compared with healthy controls. No other composite index retained significance after correction for multiple testing in the primary analysis, although several overall group effects were sensitive to HC3 robust inference. Among conventional hematological parameters, increased RDW in IIH-ON and lower serum albumin in MS-ON and IIH-ON persisted after adjustment, whereas no disease-specific hematological abnormalities were identified in NAION.

The demographic differences observed in our cohort are consistent with the established epidemiology of these disorders, with older patients in the NAION group and younger, predominantly female patients in the IIH-ON group [1, 13–15]. Because many hematological parameters are influenced by age and sex, adjustment for these variables was essential to minimize demographic confounding in the between-group comparisons.

Adjustment for age and sex substantially altered the results. Several significant unadjusted differences, including red blood cell count, hemoglobin, platelet count, SII, PIV, and HALP, were no longer significant after adjustment, particularly in the NAION group. Given that NAION predominantly affects older individuals with a high burden of vascular comorbidity, these findings suggest that previously reported hematological differences may largely reflect demographic characteristics rather than disease-specific inflammation. This may also explain the inconsistent hemoglobin findings reported in previous NAION studies [16] and the elevated NLR described in studies that did not adequately account for demographic confounding [2, 6].

NLR was the only index to survive both demographic adjustment and correction for multiple testing, with higher adjusted values in MS-ON than in controls. This is consistent with earlier reports of elevated NLR in multiple sclerosis and optic neuritis [7] and with a growing body of experimental work implicating neutrophils in the early phase of MS-related neuroinflammation, including blood-brain barrier disruption and amplification of adaptive immune responses. Models of neuroinflammation such as experimental autoimmune encephalomyelitis and cuprizone-induced demyelination show early increases in neutrophil numbers with prominent meningeal accumulation at disease onset [17], and Shen et al. [18] reported that raised peripheral neutrophil counts and NLR values track disease activity in multiple sclerosis, particularly before disease-modifying treatment is started. The clinical utility of NLR was nevertheless modest. The AUC for discriminating MS-ON from healthy controls was 0.681, and the optimal cut-off (2.59) provided high specificity (94.4%) but low sensitivity (44.4%). These findings suggest that NLR may serve as a supportive marker for inflammatory optic neuropathy when elevated, but has limited value for excluding the diagnosis when normal. Given that MS-ON diagnosis relies primarily on clinical, imaging, and cerebrospinal fluid findings, NLR should be regarded as an inexpensive adjunct rather than a standalone diagnostic test.

In the primary analysis, no composite index other than NLR showed a significant Bonferroni-adjusted pairwise difference between groups. Although SIRI remained nominally significant after adjustment for age and sex (p = 0.033), it was no longer significant after false discovery rate correction (q = 0.100), and the MS-ON versus control comparison did not remain significant after Bonferroni adjustment (p = 0.059). Similar borderline findings were observed for SII and PIV, whereas PLR showed no significant group differences. Although previous studies have reported elevated SIRI in optic neuritis [19], our findings did not confirm its independent discriminatory value after correction for multiple testing, and ROC analysis showed no significant difference between the performance of NLR and SIRI (DeLong p = 0.637). The HALP score was lower in IIH-ON in the unadjusted analysis but was no longer significant after adjustment for age and sex (p = 0.106). This likely reflects attenuation of the hemoglobin and platelet differences contributing to the score, although its interpretation differed under HC3 robust inference.

Among the conventional hematological parameters, IIH-ON showed the most distinct profile. In unadjusted analyses, patients with IIH-ON had lower hemoglobin and higher RDW and platelet counts. After adjustment for age and sex, only RDW remained significantly higher than in healthy controls, while hemoglobin differences were no longer significant and the overall platelet effect was attenuated despite higher platelet counts in IIH-ON than in MS-ON.

RDW reflects variation in erythrocyte size and is increased in iron-restricted erythropoiesis and chronic inflammation [20]. Similar findings, including anemia and reactive thrombocytosis, have been reported in patients with IIH and papilledema [8, 21], with iron deficiency proposed as a potential contributing factor in this predominantly young female population [22]. However, the underlying mechanism remains uncertain because data on iron status and body mass index were unavailable.

Serum albumin remained lower in both MS-ON and IIH-ON than in healthy controls after adjustment, making it one of the most robust conventional markers identified in this cohort. As a negative acute-phase reactant, reduced albumin may reflect low-grade systemic inflammation. In IIH-ON, obesity-related inflammation has been proposed as a contributing mechanism, although this could not be evaluated because body mass index data were unavailable [23]. In MS-ON, lower albumin may be associated with systemic inflammation and blood-brain barrier dysfunction [24], consistent with observations in other immune-mediated demyelinating disorders [9]. However, albumin is also influenced by nutritional and hydration status, which could not be assessed in this retrospective study.

The principal strength of this study is its comparative design: three etiologically distinct optic neuropathies were evaluated head-to-head against a common control group, which has rarely been done. All samples were obtained at diagnosis and before treatment, so the measurements reflect the untreated baseline inflammatory state. Analytically, the study adjusted for the demographic differences that are intrinsic to these disease groups, applied false discovery rate control across pre-specified indices, used log transformation and heteroscedasticity-consistent inference in response to explicit assumption testing, and reported pre-specified sensitivity analyses. Complete untransformed distributions are provided so that readers can assess the influence of extreme values.

Several limitations should be acknowledged. The laboratory analyzers used across the 20-year study period could not be verified retrospectively; changes in instrumentation or calibration may have introduced measurement variability. The availability of preoperative laboratory results for controls may have introduced selection bias. Previously unrecognized diabetes or hypertension in the NAION group could not be ruled out by the absence of known uncontrolled disease or retinopathy at presentation; these comorbidities were not included as covariates and may have caused residual confounding. Fasting status was not independently verifiable from the retrospective records despite restricting samples to the morning window. Several nominal group effects changed significance under HC3 robust inference, indicating sensitivity to the variance estimator. ROC analyses used unadjusted biomarker values and therefore did not account for age and sex differences between the MS-ON and control groups. The retrospective single-center design limits causal inference and generalizability, and the sample size may have been insufficient to detect small between-group differences. The marked sex imbalance in the IIH-ON group limited the precision of sex-adjusted analyses and precluded assessment of group-by-sex interactions; therefore, the association between NLR and MS-ON requires confirmation in larger, more balanced cohorts. Finally, the absence of body mass index, iron status, inflammatory biomarkers, and clinical severity measures limited mechanistic interpretation and prevented evaluation of the relationship between inflammatory indices and disease severity, while the cross-sectional design precluded longitudinal assessment.

In conclusion, in this comparative study of three optic neuropathy subtypes, NLR was the only composite inflammatory index that retained significance after adjustment for age, sex, and multiple testing in the primary analysis, showing higher levels in MS-ON than in healthy controls. However, overall group effects for several other indices were sensitive to HC3 robust inference, and their potential value remains uncertain. Among conventional hematological parameters, increased RDW in IIH-ON and lower serum albumin in MS-ON and IIH-ON were the most consistent findings. Larger prospective studies are needed to validate these findings and clarify the clinical utility of CBC-derived inflammatory markers in optic neuropathies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Concept: CDE; Design: BHU, CDE; Supervision: OK and MSB; Resource: OK and MSB; Data Collection and/or Processing: BHU, CDE, MY, and EA; Analysis and/or Interpretation: BHU and CDE; Literature Search: BHU and CDE; Writing: BHU and CDE; Critical Reviews: OK and MSB.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was obtained from the Dokuz Eylul University Ethics Committee (decision number 2025/42 − 09, dated 1 December 2025). The study was conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was waived owing to the retrospective design.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.