Association of the neutrophil‐lymphocyte ratio with outcome in sick hospitalized neonatal foals

Abstract

Background

The neutrophil‐lymphocyte ratio (NLR) in human medicine is an objective biomarker that reflects prognosis. The NLR as an independent biomarker to help predict nonsurvival in hospitalized neonatal foals has not been thoroughly interrogated.

Objectives/Hypothesis

Retrospectively evaluate if the NLR at admission is associated with nonsurvival in sick hospitalized foals <4 days old. We hypothesized that a lower NLR will be associated with nonsurvival.

Animals

One thousand one hundred ninety‐six client‐owned foals <4 days old of any breed and sex: 993 hospitalized foals and 203 healthy foals.

Methods

Retrospective multicenter study. Medical records of foals presenting to 3 equine referral hospitals were reviewed. Foals were included if they had complete CBCs, sepsis scores, and outcome data. The NLR was calculated by dividing the absolute neutrophil count by the absolute lymphocyte count. Data were analyzed by nonparametric methods and univariate analysis.

Results

Of the 993 sick hospitalized foals, 686 were sick nonseptic and 307 were septic. The median NLR was lower in sick hospitalized foals (median [95% confidence interval], 3.55 [0.5‐13.9]) compared with healthy foals (6.61 [3.06‐18.1]). Septic foals had the lowest NLR (2.00 [0.20‐9.71]). The NLR was lower in nonsurviving (1.97 [1.67‐2.45]) compared with surviving foals (4.10 [3.76‐4.33]). Nonsurviving septic foals had the lowest NLR (1.47 [1.70‐3.01]). Foals with a NLR of <3.06 or <1.6 at admission had odds ratio of 3.21 (2.24‐4.29) and 4.03 (2.86‐5.67) for nonsurvival, respectively.

Conclusions and Clinical Importance

A NLR < 3.06 at admission in sick hospitalized foals is readily available and clinically useful variable to provide prognostic information.

Abbreviations

AUC

Area under the curve

CI

confidence interval

IgG

immunoglobulin G

NLR

neutrophil‐to‐lymphocyte ratio

OR

odds ratio

ROC

receiver operating characteristic

SNS

sick nonseptic

1. INTRODUCTION

The case fatality rate of equine neonates admitted to referral hospitals ranges from 20% to 60%. ¹ , ² , ³ Identifying hospitalized neonatal foals at higher risk of dying early in hospitalization is essential for providing timely therapies, improving survival, enhancing owner communication, and making clinical decisions. Application of clinical, physical, and biochemical variables to predict survival in sick foals often relies on inaccurate scoring systems, multiple variables, advanced and specialized tests, and correct identification of the primary disorder or pathogen. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ Having access to a rapidly available objective clinicopathologic variable that can be used to accurately estimate the likelihood of survival in hospitalized foals at admission would be a valuable tool for clinicians.

In human medicine, the neutrophil‐lymphocyte ratio (NLR) is an emerging biomarker with broad clinical utility to predict disease severity and outcome in different disease states including infections, inflammatory disorders, and cancer. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ The NLR is easily obtained from a peripheral CBC by dividing the absolute neutrophil count by the absolute lymphocyte count. Calculating the NLR within 24 h of admission in intensive care units can be a good predictor of survival in critical patients. ¹⁰ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ The NLR captures the dynamic relationship between neutrophils and lymphocytes in response to neuroendocrine stress and immune‐inflammatory imbalances, which likely contributes to its clinical applicability.

The NLR is not a new concept in equine neonatology; however, its clinical utility has been limited to anecdotal experience and a few reports investigating the NLR in foals. ¹⁹ , ²⁰ , ²¹ , ²² The reports suggest that foals with sepsis have a lower NLR compared with healthy foals, and a low NLR carries a poor prognosis; however, the reports are limited to either a small number of foals or a subset of foals (ie, premature or noninfected). No large‐scale (>200 foals), multicenter study has been conducted to investigate whether the NLR, independent of other clinical variables, could provide clinically useful prognostic information. We proposed that the NLR at admission could have prognostic utility for predicting death in foals ≤4 days of age that present to tertiary equine referral hospitals. We postulated that the NLR at admission will predict nonsurvival in hospitalized foals independent of sepsis score classification. The main objectives of our study were (1) to define the NLR at admission in sick nonseptic (SNS; sepsis score ≤11) and septic (sepsis score ≥ 12) foals and compare them to healthy foals; (2) to determine whether the NLR at admission differs between surviving and nonsurviving foals; and (3) to evaluate whether the NLR can predict nonsurvival in sick hospitalized foals. We also explore whether there is a correlation between the NLR and prematurity, serum immunoglobulin G (IgG), and cortisol concentrations. We hypothesized that hospitalized foals would have a lower NLR compared with healthy foals and that a lower NLR will reflect disease severity and be associated with nonsurvival.

2. MATERIALS AND METHODS

2.1. Animal and inclusion criteria

A total of 1196 foals ≤4 days of age of any breed and sex from 2009 to 2019 foaling seasons were included in our retrospective study, of which 993 were hospitalized and 203 were healthy. Sick hospitalized foals were admitted to 3 equine referral hospitals (The Ohio State University Equine Center, Rood and Riddle Equine Hospital, and Hagyard Equine Medical Institute) as the primary patient and did not include foals that were healthy companions to sick mares. Foals were excluded from our study if they did not have the following recorded at the time of hospital admission: (1) a CBC with both neutrophil and lymphocyte counts (10³/μL); (2) documentation of either discharge or nonsurvival (death or euthanasia); (3) sepsis score. Sick hospitalized foals with a sepsis score ≥12 were considered septic, whereas foals that presented for illnesses with sepsis scores ≤11 were classified as SNS. ²³ , ²⁴ In our study, sepsis score was used as a surrogate for disease severity as a higher sepsis score reflects more severe systemic disturbances. Positive blood culture results did not automatically classify a foal as septic if all other clinical variables were not consistent with sepsis, as healthy foals can have positive blood cultures. ²⁵ It was not the goal of our study to determine whether the NLR correlates with blood culture results. Healthy foals were evaluated at a nearby breeding farm or the hospital and were considered healthy based on normal physical examination as determined by a licensed clinician, hematology, serum IgG concentrations (>800 mg/dL), and sepsis scores <3. This cutoff was based on clinical judgment and the scoring system in the 3 equine referral hospitals of our study. Some healthy foals presented as companions to their sick mares or were born in the hospital as part of a foaling program. Sick hospitalized foals were categorized into survivors (n = 786) and nonsurvivors (n = 207). Survivors were foals discharged alive, and nonsurvivors were foals that died or were euthanized because of grave medical prognosis. Foals euthanized for financial reasons were not included in our study. Prematurity was defined based on gestational age where <320 days of gestation indicates prematurity.

Our study was approved by The Ohio State University Institutional Animal Care and Use Committee and research was carried out following institutional and The US Department of Agriculture guidelines on the use of animals in veterinary research.

2.2. Hematology, serum IgG, and cortisol measurements

Blood was collected in plain tubes from each foal at admission, allowed to clot, centrifuged, and stored at −80°C until analysis. A complete blood cell count was obtained at admission using EDTA‐containing tubes and analyzed by a hematology system (ADVIA). Serum IgG concentrations (in milligrams per deciliter) were determined by turbidimetry using an automated system (Roche COBAS c501). Serum cortisol concentrations (in micrograms per deciliter) were determined with a validated radioimmunoassay. ²

2.3. Statistics

Data were tested for normality by the Shapiro‐Wilk statistics and were not normally distributed. Thus, values are presented as median and range and descriptive statistics were used to determine the 5% to 95% confidence interval (CI). Sick hospitalized foals were classified as having low, high, and normal NLR based on 5% to 95% CI from healthy foals. Comparisons among 3 groups (healthy vs SNS vs septic) were performed by Kruskal‐Wallis analysis of variance (ANOVA) with Dunn's post hoc analysis, whereas comparisons between 2 groups (survivors vs nonsurvivors) were analyzed with the Mann‐Whitney U test. The receiver operating characteristic (ROC) curve was used to determine sensitivity and specificity of the NLR to predict disease severity and nonsurvival. Based on the ROC, the best cutoff points to predict either severe disease or nonsurvival were ascertained. Univariate logistic regression and crude odds ratio (OR) were calculated to determine association between NLR and nonsurvival. The dependent variable was nonsurvival, and the referent was normal NLR range from healthy foals. Correlation between individual variables was assessed by the Spearman coefficient (ρ). Statistical analyses were performed using IBM SPSS Statistics 28.0 (IBM Corporation, Armonk, New York, USA) and Prism 9.0 (GraphPad Software, Inc., La Jolla, California, USA). Figures were created using Prism 9.0 (GraphPad Software, Inc., La Jolla, California, USA) and Inkscape (Inkscape Project, https://inkscape.org). Significance was set at P < .05.

3. RESULTS

3.1. Study sample demographics

Of the sick hospitalized foals (n = 993), 31% (n = 307/993) were classified as septic and 69% (n = 686/993) as SNS. The median sepsis score for hospitalized foals was 9 (range: 0‐29) and for septic and SNS foals were 14 (range: 12‐29) and 7 (range: 0‐11), respectively. The nonsurvival rates for septic and SNS foals were 43% (n = 135/307) and 10% (n = 71/686), respectively. The median age of septic and SNS foals was 5 h (range: 0‐96 h) and 12 h (0‐96 h), respectively. There was no significant difference in age between septic and SNS foals (P = .14). Of the sick hospitalized foals, 80% (n = 795/993) had recorded prematurity information, 14% (n = 113/795) were premature, and 86% (n = 682/795) were full‐term. The nonsurvival rates for premature and full‐term foals were 45% (n = 50/113) and 17% (n = 113/682), respectively.

Breeds represented in the total foal cohort included Thoroughbred (n = 676), Standardbred (n = 160), Quarter Horse (n = 62), Saddlebred (n = 22), American Paint Horse (n = 12), Warmblood (n = 11), Arabian (n = 9), Friesian (n = 7), Belgian (n = 6), American Miniature Horse (n = 4), Andalusian (n = 2), Appaloosa (n = 2), Rocky Mountain Horse (n = 2), and Gypsy Vanner (n = 1). Breed was not recorded for 217 foals. In our study cohort, 46% of foals were colts (551/1193), 39% were fillies (463/1193), and 15% (179/1193) had no recorded sex.

3.2. NLR, neutrophil, and lymphocyte counts for healthy and hospitalized foals

The median NLR, neutrophil count, and lymphocyte count for healthy and sick hospitalized foals are presented in Table 1. We stratified hospitalized foals into septic and SNS and determined their median NLR, neutrophil count, and lymphocyte count. The NLR was significantly lower in sick hospitalized (septic and SNS), septic, and SNS foals compared with healthy foals (Table 1 and Figure 1A). Septic foals had lower neutrophil counts than healthy foals (Table 1 and Figure 1B). No significant differences were found in the lymphocyte count between groups of foals (Table 1 and Figure 1C).

TABLE 1.

Admission NLR, neutrophil count, and lymphocyte count in healthy and sick hospitalized foals (median, range, and 5%‐95% confidence interval).

	Healthy (n = 203)	Hospitalized (n = 993)	SNS (n = 686)	Septic (n = 307)
Median NLR (range) 5%‐95%	6.61 (1.83‐27.0) 3.06‐18.1	3.55*** (0.04‐39.5) 0.50‐13.9	4.35*** (0.04‐39.5) 1.01‐15.8	2.00***‡ (0.04‐33.8) 0.20‐9.71
Median neutrophil count (103/μL) (range) 5%‐95%	8.40 (1.88‐27.9) 4.75‐13.4	5.20*** (0.01‐29.4) 0.49‐12.6	6.20*** (0.04‐21.5) 1.40‐12.9	2.80***‡ (0.01‐29.4) 0.20‐11.8
Median lymphocyte count (103/μL) (range) 5%‐95%	1.29 (0.30‐4.30) 0.50‐2.30	1.39 (0.03‐8.10) 0.04‐3.27	1.40 (0.08‐6.70) 0.40‐3.01	1.33 (0.03–8.10) 0.35‐3.97

Note: ***P < .0001 compared with healthy foals; ‡P < .0001 compared with SNS foals. The italic values define the P values for significance within each table.

FIGURE 1.

The neutrophil‐lymphocyte ratio (NLR) (A), neutrophil count (B), and lymphocyte count (C) in healthy, sick hospitalized, sick nonseptic (SNS), and septic foals based on sepsis score. (***P < .0001; ns, no significance). Each circle represents an individual foal. Sick hospitalized foals included both SNS and septic foals.

3.3. NLR and disease severity

To determine that the NLR reflects disease severity as defined by the sepsis score, ROC curves were generated for sick hospitalized (SNS and septic), SNS, and septic foals (Figure 2A‐C). A NLR < 5.27 at admission had a sensitivity of 68% and specificity of 70% in hospitalized foals (AUC = 0.74; P < .0001); a NLR < 5.90 had a sensitivity of 68% and specificity of 56% in SNS (AUC = 0.70; P < .0001); and a NLR < 3.55 had a sensitivity of 72% and specificity of 90% in septic foals to predict disease severity (AUC = 0.86; P < .0001).

FIGURE 2.

Receiver operating characteristic curves for neutrophil‐lymphocyte ratio (NLR) and disease severity on admission in sick hospitalized, sick nonseptic (SNS), and septic foals. (A) NLR < 5.27 was associated with disease severity in sick hospitalized foals; (B) NLR < 5.9 was associated with disease severity in SNS foals; (C) NLR < 3.55 was associated with disease severity in septic foals. Sick hospitalized foals included both SNS and septic foals.

3.4. NLR and nonsurvival in SNS and septic foals

To determine the relationship between the NLR and nonsurvival in sick hospitalized foals, we analyzed sick hospitalized foals collectively and then stratified them based on SNS and septic status. In sick hospitalized and septic foals, the NLR at admission was significantly lower in nonsurvivors compared with survivors (Table 2 and Figure 3A, P = .0001). In SNS foals, the NLR was not different between survivors and nonsurvivors (Table 2 and Figure 3A, P = .10). Additionally, in sick hospitalized, SNS, and septic foals, there was a significant difference in the neutrophil counts between survivors and nonsurvivors (Table 2 and Figure 3B). In sick hospitalized and SNS foals, lymphocyte counts were not different between survivors and nonsurvivors, but in septic foals, there was an increase in lymphocyte count in nonsurvivors compared with survivors (Table 2 and Figure 3C).

TABLE 2.

The NLR, neutrophil count, and lymphocyte count at admission of surviving and nonsurviving foals (median, range, and 5%‐95% confidence interval).

	Hospitalized (n = 993)	SNS (n = 686)	Septic (n = 307)
Median NLR survivors (range) 5%‐95%	4.10*** (0.09‐39.5) (3.76–4.33)	4.42 (0.09‐39.5) (4.15‐4.71)	2.6*** (0.09‐33.7) (2.18‐3.33)
Median NLR nonsurvivors (range) 5%‐95%	1.97 (0.04‐30.9) (1.67–2.45)	3.75 (0.04–30.9) (2.71‐5.12)	1.47 (0.04‐11.4) (1.08‐1.81)
Median neutrophil counts (103/μL) Survivors (range) 5%‐95%	5.70*** (0.04‐29.3) (5.40‐6.00)	6.35** (0.10‐20.20) (5.90‐6.70)	3.10* (0.04–29.3) (2.70‐3.50)
Median neutrophil counts (103/μL) Nonsurvivors (range) 5%‐95%	3.10	5.12	2.20
	(0.02‐21.52)	(0.04‐21.50)	(0.02‐17.7)
	(2.30‐3.8)	(4.23‐6.20)	(1.70–3.01)
Median lymphocyte counts (103/μL) Survivors (range) 5%‐95%	1.37 (0.03‐8.00) (1.00‐1.43)	1.40 (0.08–6.70) (1.30‐1.44)	1.22* (0.03‐8.00) (1.3‐1.41)
Median lymphocyte counts (103/μL) Nonsurvivors (range) 5%‐95%	1.40 (0.10‐6.00) (1.24‐1.50)	1.22 (0.13‐3.69) (0.90‐1.43)	1.50 (0.10‐6.00) (1.3‐1.76)

Note: ***P < .0001; **P < .001; *P < .01 survivors compared with nonsurvivors. The italic values define the P values for significance within each table.

FIGURE 3.

Comparison between the neutrophil‐lymphocyte ratio (NLR) (A), neutrophil count (B), and lymphocyte count (C) of surviving versus nonsurviving foals. Each circle represents an individual foal (***P < .0001; **P < .001; *P < .01; ns, no significance). SNS, sick nonseptic.

To determine whether NLR could predict nonsurvival, ROC curves were generated for sick hospitalized (SNS and septic), SNS, and septic foals (Figure 4A‐C). The NLR in sick hospitalized foals had a sensitivity of 76% and a specificity of 84% (AUC = 0.86; P < .0001; Figure 4A), in SNS foals had a sensitivity of 68% and a specificity of 67% (AUC = 0.73; P < .001; Figure 4B), and in septic foals had a sensitivity of 86% and a specificity of 86% (AUC = 0.92; P < .0001; Figure 4C) to predict nonsurvival.

FIGURE 4.

Receiver operating characteristic curve for neutrophil‐lymphocyte ratio (NLR) and neutrophil counts and nonsurvival on admission in sick hospitalized, sick nonseptic (SNS), and septic foals; (A) NLR < 4.1 was associated with nonsurvival in sick hospitalized foals; (B) NLR < 5.5 was associated with nonsurvival in SNS foals; (C) NLR < 3.9 was associated with nonsurvival in septic foals; (D) neutrophil count of <6.24 (10³/μL) was associated with nonsurvival in sick hospitalized foals; (E) neutrophil count of <6.45 (10³/μL) was associated with nonsurvival in SNS foals; and (F) neutrophil count of <5.65 (10³/μL) was associated with nonsurvival in septic foals. Sick hospitalized foals included both SNS and septic foals.

Because there was a statistical difference in neutrophil counts between survivors and nonsurvivors in sick hospitalized foals, we determined whether neutrophil counts could predict nonsurvival using ROC curves (Figure 4D‐F). The neutrophil count in sick hospitalized foals had a sensitivity of 80% and a specificity of 84% (AUC = 0.86; P < .0001; Figure 4D), in SNS foals had a sensitivity of 76% and specificity of 75% (AUC = 0.81; P < .0001; Figure 4E), and in septic foals had a sensitivity of 91% and specificity of 87% (AUC = 0.95; P < .0001; Figure 4F) to predict nonsurvival.

3.5. NLR and nonsurvival independent of sepsis score

We wanted to determine if the NLR in sick hospitalized foals is a useful objective biomarker independent of sepsis score classification or other physical and clinicopathologic markers. Using the 5% to 95% confidence interval NLR range from 203 healthy foals (3.06‐18.1), we divided the sick hospitalized foals into 3 groups: low (NLR < 3.06, n = 427), normal (NLR 3.06‐18.1, n = 542), and high (NLR >18.1, n = 24) and compared nonsurvival within each group (Table 3). The low‐NLR group had the highest percentage of nonsurvivors (31%) compared with normal‐ and high‐NLR groups (13% and 8%, respectively). Within the low‐NLR group, nonsurviving foals had significantly lower NLR compared with surviving foals (P < .001). In the normal‐ and high‐NLR groups, there was no difference between the NLR of survivors versus nonsurvivors.

TABLE 3.

Nonsurvival in sick hospitalized foals based on NLR classification (low, normal, high NLR; median, range, and 5%‐95% confidence interval).

	Hospitalized foals (n = 993)	<3.06 (n = 427)	3.06‐18.1 (n = 542)	>18.1 (n = 24)
Nonsurvival, n (%)	207 (21)	135 (31)	70 (13)	2 (8)
Median NLR Survival (range) 5%‐95%	4.10*** (0.09–39.5) (3.76‐4.33)	1.91*** (0.09‐3.04) (0.33‐3.00)	5.60 (3.06‐18.1) (3.25‐13.7)	22.5 (18.5‐39.5) (18.7‐38.0)
Median NLR Nonsurvival (range) 5%‐95%	1.97 (0.04–30.9) (1.67‐2.45)	1.20 (0.04‐2.99) (0.09‐2.74)	5.17 (3.16‐17.4) (3.27‐13.5)	27.4 (23.7‐31) (24.0‐31.0)

Note: The italic values define the P values for significance within each table.

^*** P < .0001 survivors compared with nonsurvivors.

We wanted to determine whether foals with a low NLR were more likely to not survive compared with sick hospitalized foals with an NLR within the normal range. Using univariate logistic regression, we analyzed the OR of nonsurvival of 2 different NLR cut off values based on the lower end of the 5% to 95% confidence interval from healthy (NLR < 3.06) and nonsurviving foals (NLR < 1.67) and found the OR for nonsurvival to be 3.21 (95% confidence interval: 2.24‐4.29, P = <.001) and 4.03 (95% confidence interval: 2.86‐5.67, P = <.001), respectively.

3.6. The NLR and prematurity

In our subset of foals, there was no statistical difference in the NLR between premature and full‐term foals (Table 4 and Figure 5A, P = .22). In terms of nonsurvival, there was no statistical difference between premature and full‐term foals (Table 4 and Figure 5A, P = .9). In premature and full‐term foals, there was a weak statistical difference in neutrophil counts (Table 4 and Figure 5B, P = .02) and no statistical difference in lymphocyte counts (Table 4 and Figure 5C, P = .17).

TABLE 4.

NLR, neutrophil, and lymphocyte counts at admission in foals based on prematurity.

	Premature (n = 113)	Full‐term (n = 682)
Median NLR Survivors (range) 5%‐95%	3.09 (0.05‐28.6) (2.25‐3.90)	3.54 (0.04‐39.5) (3.29‐3.85)
Median NLR Nonsurvivors (range) 5%‐95%	1.48 (0.05‐13.1) (0.9‐2.5)	2.19 (0.04‐17.4) (1.74‐2.71)
Median neutrophil counts (103/μL) Survivors (range) 5%‐95%	3.50* (0.06‐29.4) (2.70‐5.50)	5.22* (0.01‐20.3) (4.90‐5.50)
Median neutrophil counts (103/μL) Nonsurvivors (range) 5%‐95%	2.00 (0.06‐21.5) (1.40‐2.90)	3.20 (0.01‐12.4) (2.20‐4.10)
Median lymphocyte counts (103/μL) Survivors (range) 5%‐95%	1.20 (0.03‐4.50) (1.10‐1.37)	1.40 (0.09‐8.00) (1.32‐1.50)
Median lymphocyte counts (103/μL) Nonsurvivors (range) 5%‐95%	1.34 (0.27‐4.50) (0.80‐2.70)	1.43 (0.10‐5.90) (0.50‐2.76)

Note: The italic values define the P values for significance within each table.

^* P < .01 premature compared with full‐term.

FIGURE 5.

Comparison between the neutrophil‐lymphocyte ratio (NLR) (A), neutrophil count (B), and lymphocyte count (C) of premature versus full‐term foals. Each circle represents an individual foal (*P < .02, ns, no significance).

3.7. Correlation between NLR and sepsis score, IgG, and cortisol

We tested whether the sepsis score was correlated with NLR and found a negative correlation between sepsis score and NLR in sick hospitalized (ρ = −0.37; P < .0001), SNS (ρ = −0.13; P < .0001), and septic foals (ρ = −0.24; P < .0001).

Low serum IgG concentrations are associated with higher odds of nonsurvival in hospitalized foals. ²⁶ The median IgG concentration in hospitalized sick foals was 750 mg/dL (range: 0‐4000), and when foals were further subdivided into SNS and septic, the IgG concentration was 844 mg/dL (range: 0‐4000) and 528 mg/dL (range: 0‐3922), respectively. However, there was no correlation between IgG concentrations and NLR in sick hospitalized (ρ = −0.001; P < .8), SNS (ρ = −0.05; P < .2), or septic foals (ρ = 0.04; P < .48). We found a positive correlation between IgG concentrations and low NLR (ρ = 0.10; P = .03).

Because cortisol dynamics can influence neutrophil and lymphocyte counts, we evaluated whether there was a correlation between NLR and cortisol concentrations. Cortisol concentrations were measured upon admission in 225 foals. Of these foals, 143 were SNS with 11% (n = 16/143) not surviving, and 82 were septic with 50% (n = 41/82) not surviving. The median cortisol concentration for all 225 foals was 7.60 μg/dL (range: 0.47‐59.3) and for septic and SNS foals was 7.63 μg/dL (range: 1.00‐59.3) and 7.36 μg/dL (range: 0.47‐54.4), respectively. In these foals, there was no correlation between the NLR and cortisol concentrations in SNS (ρ = 0.075; P = .37) or septic foals (ρ = 0.13; P = .23), no correlation between cortisol concentrations and low (ρ = 0.03; P = .77) or high NLR (ρ = −0.033; P = .71), and no correlation between NLR and cortisol concentrations in nonsurviving foals (ρ = 0.242; P = .09).

4. DISCUSSION

The objective of our study was to compare the NLR between healthy and sick hospitalized foals and to determine whether the NLR at admission was associated with nonsurvival. Our results revealed that the NLR is significantly different between sick hospitalized and healthy foals, confirming previous reports that sick hospitalized foals have a lower NLR than healthy foals and septic foals have a lower NLR compared with SNS foals. ²² Independent of sepsis score and prematurity, nonsurviving foals have a lower NLR compared with surviving foals. A lower NLR increases the odds of nonsurvival. Our data align with another study where nonsurviving foals had a lower NLR than surviving foals. ⁴ Our results, however, report a lower NLR in nonsurviving foals of 1.97 compared with 4.2, ⁴ which could be because of the larger number of foals in our study. Nevertheless, collectively, our data support that a lower NLR reflects more systemic disturbances and demonstrates that irrespective of the underlying disease or presenting complaint, foals with a low NLR have a worse prognosis. Our data build on previous studies that found that a low NLR is a poor prognostic indicator in a subset of foals (premature and noninfected). ¹⁹ , ²⁰

In contrast to the foals in our study, in humans, cats, and dogs, a high NLR is commonly associated with severe disease and worsening of disease progression. ¹¹ , ¹² , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ In these species, a high NLR is driven by neutrophilia and concurrent lymphopenia. ³² , ³³ Septic foals often have lower neutrophil counts and similar lymphocyte counts compared with SNS and healthy foals, thus suggesting that neutropenia might be a more important contributor to critical illness than neutrophilia and lymphopenia.

Our study provides more evidence that the NLR correlates with severe clinical disease as defined within the confines of the sepsis scoring system, thus supporting its clinical value. This finding is consistent with reports in human medicine demonstrating the sensitivity of the NLR in the diagnosis of severe illness including bacteremia, infection, and sepsis. ¹⁰ , ¹⁴ , ³⁴ , ³⁵ We propose that the NLR can be used to help identify foals with critical immune‐inflammatory imbalances, which can be valuable in the clinical setting. Our study utilized sepsis score as a surrogate for disease severity, and it was not the goal of our study to determine if the NLR can distinguish the underlying cause of the critical illness.

In human medicine, it has been proposed that the NLR has poor discriminatory value because it reflects the intensity of immune‐inflammatory reaction and supraphysiological stress response independent of underlying etiology. ⁸ , ¹⁴ However, in foals, it is possible that the type of bacterial infection could influence the NLR, as Gram‐negative bacteria induce a different pattern of inflammatory response than Gram‐positive bacteria. ³⁶ Additionally, in humans, changes in the NLR occur within 4 to 8 h of an acute insult and are thought to promptly reflect acute systemic inflammatory responses, ⁸ the NLR is a better predictor of bacteremia than C‐reactive protein, ¹⁰ and a combination of biomarkers along with the NLR can enhance the diagnosis of sepsis. ³⁴ Whether the NLR is a better predictor of systemic inflammation compared with other acute phase proteins was not investigated in our study.

The results of our study demonstrate that the NLR is correlated with nonsurvival with relatively high sensitivity and specificity. This result echoes the clinical utility of the NLR in human medicine to provide prognostic information in a diverse range of disease processes, including sepsis, neoplasia, pancreatitis, pneumonia, appendicitis, and acute stroke. ¹⁴ Here, nonsurviving foals had a lower NLR than surviving foals, and the low NLR was largely driven by neutropenia. This result is not surprising, given that previous studies have associated neutropenia with nonsurvival in hospitalized foals. ⁴ , ⁵ , ⁶ , ³⁷ Our results also demonstrated that neutrophil counts at admission have similar sensitivity and specificity to the NLR to predict nonsurvival, which supports our point that neutrophils are the major driver of the NLR in foals. Adding lymphocyte counts (among other variables) to a modified sepsis score did not improve its sensitivity or specificity for diagnosing sepsis in foals, ⁷ which further supports that neutrophils counts drive the prognostic utility of the NLR.

We found a weak correlation between serum IgG concentrations and a low NLR. Serum IgG concentrations are associated with disease severity and outcome in sick foals. ²⁶ , ³⁸ The weak correlation might be because the NLR captures rapid changes in the numbers of neutrophils and lymphocytes and not functionality, whereas IgG promotes immune cell function and could reflect the competency of the immune response. Additionally, a weak correlation suggests that other factors are important to the development of a low NLR besides IgG concentrations. Furthermore, assessing 1 time point might not reflect the relationship between serum IgG concentrations and the NLR.

Cortisol concentrations did not correlate with the NLR. The endocrine stress response modulates the innate and adaptive immune response, and in critically ill humans, the NLR reflects a supraphysiological stress response that leads to immune dysfunction, disease progression, and death. Additionally, hypothalamic‐pituitary‐adrenal axis dysfunction exists in septic foals, ² , ³⁹ , ⁴⁰ and cortisol concentrations are higher in hospitalized and septic foals compared with healthy foals. ² , ⁴¹ It is possible that at the time of sampling, cortisol had minimal effect on neutrophil and lymphocyte counts but could have led to leukocyte shifts early in the disease process and reduced neutrophil function. It also possible that other stress hormones play important roles in modulating the NLR in this species, especially because adrenocortical function might not be fully mature in foals at birth. ⁴² Considering that cortisol was measured in a subset of foals, it is possible that our substudy sample did not capture the relationship between cortisol and the NLR. We believe this is unlikely, however, given that 225 foals is a larger sample size than most equine neonatal studies where relevance was able to be determined.

Despite some evidence in the literature that prematurity could influence the NLR, ²⁰ , ³⁷ in our study, the NLR at admission was not statistically different between premature foals and foals with normal gestational age. The reason for this lack of difference is unclear. It is possible that certain diseases have a greater influence on the NLR than prematurity. It is also possible that the larger number of foals in our study more comprehensively captures the relationship between prematurity and the NLR, and in fact, prematurity has marginal influence on the NLR.

Healthy foals have a wide NLR range, which is different than the narrow NLR range reported in healthy adult humans and dogs. ²⁹ , ⁴³ However, the NLR range in healthy neonatal animals has not been fully established, and it is possible that neonates have substantial variation in the NLR during the perinatal period. The range of NLR for healthy foals found in our study could reflect dynamic changes that occur in the early neonatal period, as well breed differences. The wide NLR range in healthy foals made it difficult to analyze whether a high NLR is associated with nonsurvival as there were limited number of sick hospitalized foals that had an NLR greater than 18.1.

The current study found that although a low NLR at admission in foals ≤4 days of age had high sensitivity and specificity for nonsurvival, the neutrophil count had a similar sensitivity and specificity and, in fact, was more sensitive and specific than the NLR when foals were further classified into SNS and septic.

Limitations of our study include the general retrospective study design, which did not permit identification of certain factors that could influence the NLR. Clinical information in our study was ascertained from clinical records spanning a 10‐year period and in many instances lack relevant information such as foaling complications, prior treatments administered, and any gestational complications. For example, previous treatments, such as antimicrobial treatment, nonsteroidal anti‐inflammatory, glucocorticoids, or plasma administration could modify the white blood cell count of foals at admission. As with other retrospective studies, causation cannot be determined, and the temporal relationship between the NLR and nonsurvival cannot be assessed. Our study also did not include foals that were euthanized for financial reasons to avoid artificially increasing the nonsurvival sample. However, based on our clinical experience, many of these foals euthanized for financial reasons are often critically ill; therefore, by excluding this cohort, we might have avoided bias in nonsurvival but artificially lowered the numbers of critically ill foals. However, because the objective of our study was to determine the prognostic utility of the NLR, removing this group enables us to capture a correlation more accurately between the NLR and nonsurvival.

5. CONCLUSION

Our study supports that the NLR is associated with nonsurvival in hospitalized foals ≤4 days of age. Calculating the NLR at admission and carefully evaluating the neutrophil count could provide clinicians with valuable clinical and prognostic information to enhance medical decisions and improve client communications. The neutrophil count is significantly associated with outcome. Given that NLR < 3.06 captures both foals that are classified as SNS and septic, we speculate that the NLR in foals might be a valuable screening tool to recognize foals that could benefit from additional diagnostic testing and aggressive supportive care early in their hospitalization.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by The Ohio State University IACUC, protocol 2008A0170. Owners gave consent for the retrieval of clinical information.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Samuels AN, Kamr AM, Reed SM, et al. Association of the neutrophil‐lymphocyte ratio with outcome in sick hospitalized neonatal foals. J Vet Intern Med. 2024;38(2):1196‐1206. doi: 10.1111/jvim.16995