Association between neutrophil to lymphocyte ratio and mood polarity in adolescents admitted to an inpatient psychiatric ward

Abstract

Inflammatory processes are associated with mood disorders, but data on pediatric patients is scarce. The aim of this study was to investigate a possible association between elevated neutrophil - lymphocyte ratio (NLR) — a marker of inflammation, and mood polarity (manic/depressed), in adolescents, admitted between 2010-2015 due to a mood disorder episode, to an adolescent inpatient ward. Electronic medical records of 305 patients (aged 10-19 years, 60.6% males) admitted during the study period due to a mood disorder episode, were reviewed. Of these, 63 were diagnosed with manic episodes and 242 with depressive episodes. Multivariate analyses were used to compare NLR between and within the two groups, co-varying for age, sex, and antipsychotic use. NLR was significantly higher in the manic episode group compared to the depression one. Moreover, in inpatients with multiple hospitalizations, the NLR was higher during their manic episodes than during their non-manic states. These results suggest that, as has been reported in adults with bipolar disorder, inflammatory mechanisms may be involved in adolescents’ mood disorders as well, particularly in the manic episodes. Thus, clinicians may consider adding anti-inflammatories as part of the treatment of these patients.

Introduction

The role of the immune system in the pathophysiology of psychiatric disorders has received much attention over the years (Dantzer et al., 2008). Presence of circulatory inflammatory markers has been reported in numerous psychiatric disorders in adults (Gariup et al., 2015). However, the potential role of inflammation in pediatric age groups has been less investigated (Mills et al., 2013). Early studies on association between inflammation and psychopathology focused on mood, especially major depressive disorder (MDD) (Maes, 1995,Du Preez et al., 2016). Patients with depression have been found to have higher circulatory levels of pro-inflammatory cytokines, acute phase proteins, chemokines and cellular adhesion molecules (Kohler et al., 2017). A strong association between inflammatory markers and mania was also found (Mazza et al., 2018,Inanli et al., 2019,Melo et al., 2019). However, there are few studies on the association between the inflammatory system and bipolar disorder in children and adolescents. The lifetime prevalence of bipolar disorder among adolescents was found to be between 1.9% and 4.3%. It was also found to be higher among girls (3.3%) than among boys (2.6%) (Merikangas et al., 2010).

The neutrophil/lymphocyte ratio (NLR) has been investigated as a proxy marker for systemic inflammatory response (Aydin Sunbul et al., 2016,Mentis et al., 2016). It was also reported that NLR is higher in adult patients with MDD than in healthy controls (Demir et al., 2015). NLR was found to correlate with C-Reactive Protein (CRP) levels - an established biomarker of inflammation in aggressive schizophrenia patients (Barzilay et al., 2016). Furthermore, in adults with bipolar disorder, both NLR and platelet to lymphocyte ratio were found to be significantly higher during manic episodes compared to non-manic states. (Demir et al., 2015,Aydin Sunbul et al., 2016,Demircan et al., 2016). These findings suggest involvement of inflammatory processes in the pathophysiology of bipolar disorder in adults, especially in the manic state (Kalelioglu et al., 2015,Munkholm et al., 2015,Wang et al., 2016).

A study in adolescent inpatients with non-affective psychosis also reported elevated NLR (Bustan et al., 2018). Literature on the association between severe mood disorders and inflammation in children and adolescents with neuropsychiatric disorders is limited (Goldstein et al., 2011,Mitchell and Goldstein, 2014,Binici et al., 2018,Ozyurt and Binici, 2018). Cevher Binici et al. found no differences in NLR between adolescents with bipolar disorder and controls. They, however, suggest that further studies are needed to investigate changes in this parameter during the various episodes of bipolar disorder (Cevher Binici et al., 2018).

Mood disorders are often associated with suicidality. Notably, measures of affective temperament-types were reported to be associated with lifetime suicide attempt more strongly than diagnosis of major affective disorders (Baldessarini et al., 2017). It has also been suggested that people’s Internet searches for information and news regarding suicide may be linked to risk for suicidality. An Italian national study, however, found that adults’ online searches for suicide-related terms in Italy are more likely to be linked to factors other than suicidiality, such as personal interest and suicide bereavement (Solano et al., 2016).

Despite the often above-mentioned association, the current study did not investigate the relationship between NLR values and suicidal behaviors or measures of affective temperament-types. Rather, this study focused on the association between manic states and NLR, compared to depressed states, in adolescent inpatients with severe mood disorder episodes, where NLR serves as a possible proxy-marker for inflammation. Association between higher NLR and manic state was investigated in patients with multiple admissions too.

The hypothesis was that, similar to adults, in adolescents too mania is associated with increased inflammation leading to higher NLR.

Methods

Design

This study employed a retrospective cross-sectional design. Data were retrieved from the electronic medical records of all consecutive admittances between March 2010 and September 2015, in a single adolescent inpatient ward at the Geha Mental Health Center (GMHC) in Israel. This is a regional mental health center, with a catchment area of approximately 900,000 inhabitants. The GMHC review board approved the study and waived the need for signed informed consent due to the retrospective nature of the study.

Subjects

All adolescent inpatients aged 10-19 years, who were hospitalized in the adolescent ward due to major affective episodes (depressive, manic or mixed episodes), as define in the DSM-5, were included in the study. Blood samples had been drawn within three days from admission (a routine practice in the ward) and analyzed for complete blood count and blood chemistry. For patients whose record contained more than one admission during the study period, only the first hospitalization was included in the study.

Patients were classified as having a manic episode (n=63) or as having depression (n=242). Patients with mixed episodes were not included in the study. Four child and adolescent psychiatrists (DHBD, YB, RB and MA) reviewed, each separately, the patients’ records and full consensus of the diagnosis were required for inclusion in the study.

Exclusion criteria consisted of presence of mental retardation, neurological diseases, evidence of inflammatory diseases and treatment with agents that may affect the immune system, like lithium. Two patients with leukocytosis (WBC count>14,000/mm³), one patient with fever (body temperature>37.9°C) and three patients treated with lithium at admission, were excluded.

Variables

Sociodemographic and clinical data retrieved from patients’ electronic medical records included age at hospitalization, sex, body mass index (BMI), medications at admission, family history of mood disorders and duration of hospitalization. Laboratory data included blood count parameters: total leukocyte counts, platelet counts, absolute neutrophil counts, absolute lymphocyte counts and albumin.

NLR for each patient was calculated as follows:

$$\text{NLR}=\frac{numberofneutrophils}{numberoflymphocytes}$$

Statistical analysis

SPSS for Windows ver. 24.0 (IBM Corp., IBM SPSS Statistics for Windows, version 21.0, Armonk, New York) was used for statistical analyses. Descriptive statistics are expressed as mean ± standard deviation (SD) for continuous variables and rates (%) for categorical variables. For univariate analyses, two-tailed Student's t-tests, Mann-Whitney U test or Chi-square test, were used as appropriate. Multivariate analysis was performed using binary logistic regression analyses with NLR as the main independent variable and mania as a dependent variable controlling for age, sex, and antipsychotic use as covariates.

An additional exploratory analysis was conducted within a subset of patients, hospitalized due to manic episodes, who were subsequently re-hospitalized during the study period but not due to an affective/psychotic episode (n=13). We compared NLR between the two hospitalizations (manic and non-manic) using paired t-test. In all analyses, significance was set at p < 0.05.

Results

Clinical and demographic data of adolescent inpatients with affective disorders are shown in Table 1. The mania (n=63) group was comprised of adolescents diagnosed with a manic episode at admission, while at discharge most (55/63, 87.3%) were diagnosed with bipolar or schizoaffective disorders as major diagnosis.

Table 1-.

Demographic, clinical and laboratory parameters of adolescent inpatients admitted due to affective exacerbation by state.

		Manic state (N=63)	State of depression (N = 242)	Statistics	t (df)/U test/ Pearson χ2 (df)	p-values
Demographics & Clinical Parameters	Age, mean (SD)	15.9 (1.6)	15.1 (1.8)	Mann-Whitney	9.56	0.002
	Males, n (%)	38 (60.3%)	147 (51.6%)	Chi-square	10.56 (1)	0.002
	Smokers, n (%)	11 (17.5%)	63 (26.1%)	Chi-square	2.29 (2)+	0.318
	BMI [Kg/m2], mean (SD)	23.7 (4.8)	22.7 (4.8)	Mann-Whitney	7.75	0.143
	On antipsychotics at admission	25 (39.7%)	58 (24%)	Chi-square	6.23 (1)	0.013
	Length of Hospitalization [days]	94.3 (98.9)	57.7 (76.9)	Mann-Whitney	10.15	<0.001
	Family History of Mood disorders	16 (30.8%)	34 (18.9%)	Chi-square	3.36 (1)	0.066
Laboratory Parameters	Albumin (g/dL)	4.68 (0.28)	4.66 (0.3)	t-test	−0.39 (303)	0.698
	WBC *103/mm3, mean (SD)	7.36 (2.2)	7.06 (1.8)	t-test	−1.11 (303)	0.27
	Platelets *103/mm3, mean (SD)	257(57)	266 (62)	t-test	1.09 (303)	0.275
	Neutrophil/Lymphocyte ratio (NLR), mean (SD)	2.36 (1.70)	1.87 (1.00)	Mann-Whitney	8.94	0.035

⁺Smoking status data was missing for 27% of the sample population

BMI=Body mass index; NLR=neutrophil to leukocyte ratio; WBC= white blood cells

The patients in the mania group were significantly older, consisting of a higher rate of males, received more antipsychotic medications at admission and had longer hospitalization stays compared to the depression group (Table 1). No differences were found between those groups regarding smoking habits, BMI or family history of mood disorders (Table 1). Analysis of the blood counts revealed that the patients with mania had significantly elevated (26%) NLR compared to those with depression (2.36 ± 1.70 vs. 1.87±1.00 respectively, p=0.001). No differences were observed in leukocyte and platelet counts or in serum albumin levels (all p’s>0.27, Table 1).

Multivariate analysis was performed to assess the effect of covariates on the association between NLR and mania. A binary logistic regression with mania as the dependent variable was applied. The model included NLR, age, male sex and antipsychotic treatment as independent variables. The logistic regression model was statistically significant [χ² (4) = 31.424, p =0.000]. It explained 15.3% (Nagelkerke R2 value) of the variance in the mania state. After controlling for all other factors in the model, the association of the NLR value with manic polarity was significant (odds ratio= 1.288, 95% confidence interval 1.019-1.628, p=0.034). Full model statistics are shown in Table 2.

Table 2:

Details of the model statistics

Parameter	Odds ratio	95% confidence interval	p-value
NLR	1.288	1.019-1.628	0.034
Age	1.288	1.075-1.542	0.006
Male sex	2.566	1.421-4.635	0.002
Antipsychotic treatment	2.123	1.146-3.932	0.017

NLR = neutrophil to lymphocyte ratio

In order to explore the possibility that in an individual, manic polarity is associated with higher NLR, we analyzed data from a subset of 13 patients (mean age at admission 14.9 years, 7 females), who had been hospitalized due to a manic episode and were re-hospitalized during the study period for reasons other than mania, depression or psychosis, namely, non-affective behavioral problems. NLR at admission due to a manic episode compared to a corresponding NLR at re-admission due to a non-manic state (n=13; mean time between admissions 264±260 days), revealed higher NLR (+33%) at admission due to mania (2.0±0.8 vs. 1.5±0.5, respectively, paired-t(12)= 2.341, p<0.05). In this subset of patients, no differences were found in total white blood cells and platelet counts between the two types of admission (Table 3).

Table 3.

Blood counts of a subset of 13 patients at admission due to mania, compared to their blood counts at a subsequent re-admission due to non-affective psychosis

Blood Count Parameters	Manic episode mean (SD) (N=13)	Non-manic episode mean (SD) (N=13)	Paired-t (df=12)	p-value
WBC *103/mm3	7.56 (1.7)	7.19 (1.4)	0.92	0.37
Platelets *103/mm3	237.00 (74.0)	284.00 (97)	−1.67	0.12
NLR	2.00 (0.8)	1.5 (0.5)	2.34	0.04

WBC= white blood cells; NLR=neutrophil to lymphocyte ratio

Discussion

Major findings of the current study were as follows: NLR was significantly higher in the manic episode group (n=63) compared to the depression one (n=242). Moreover, in inpatients with multiple hospitalizations (n=13), the NLR was higher during their manic episodes than during their non-manic states. .

These findings suggest the existence, of an association between manic state and inflammation in adolescents, similar to that found in adult patients (Munkholm et al., 2015,Cevher Binici et al., 2018,Mazza et al., 2018,Mazza et al., 2019,Brinn and Stone, 2020,Fusar-Poli et al., 2021,Marazziti et al., 2021,Özdin and Usta, 2021). The current study supports the notion that inflammation is involved in the emergence of bipolar manic states, already at adolescence (Sayana et al., 2017). Although no causation can be inferred from such an association, the presence of an inflammatory marker during manic states at a young age may indicate a possible contribution of inflammatory processes to the emergence of mania in pediatric populations. Notably, our findings in adolescent inpatients support the potential role of inflammation in mood disorder pathophysiology, since pediatric populations have fewer metabolic and environmental confounders such as metabolic syndrome, cardiovascular pathology, smoking, or polypharmacy, in contrast to adults with bipolar disorder, (Salvi et al., 2020).

Limitations

The results of this study should be considered within the context of certain limitations. Common peripheral blood markers of inflammation, such as CRP or pro-inflammatory cytokines were not assessed in the current study. Instead, only NLR was used, as it has been found to be a reliable inflammatory marker for mood disorders in adults (Aydin Sunbul et al., 2016,Mentis et al., 2016).

Additionally, due the retrospective nature of the study, no clinical scales of depression or mania were used. Rather, the analysis was based strictly on chart reviews. Future prospective studies may use structured clinical interviews and psychometric scales in order to obtain a more accurate diagnostic classification as well as assessment of the severity of the clinical state.

An additional limitation is the fact that, as indicated in the Introduction, the current study did not investigate the association between NLR values and suicidal behavior or measures of affective temperament-types. Future studies should take suicidality into account.

Another limitation is the fact that all the data came from a single mental health center. The strengths of the current study lie in the relatively high number of adolescent inpatients (N=305) and in the analysis of a subset of patients (n=13) showing that elevated NLR is present only during a manic state, which points to its being an inflammatory state marker rather than a trait marker of bipolar disorder. The causal relationship of the association between inflammation and mania is still unclear however and merits further investigation (Kim et al., 2014).

Conclusion

This study reports an association between manic mood and an elevated proxy marker of inflammation, namely NLR, in adolescent inpatients. The results concur with previous studies in adult population with mood disorders (Hope et al., 2011,Munkholm et al., 2015,Mazza et al., 2018,Mazza et al., 2019,Brinn and Stone, 2020). While the association does not mean causation, the fact that an inflammatory indicator is evident at adolescence, suggests that immune factors play a role in the development of manic states of bipolar disorder already at an early age. Further studies should be conducted in adolescents, in order to identify a causal relationship between manic phases and inflammation and clarify the clinical implications of the presence of this marker for the diagnosis, prognosis and therapeutic strategies for manic episodes in bipolar disorder. Nevertheless, this study’s clinical message is that the association between elevated NLR and manic state in bipolar disorder points to the role of inflammatory processes in bipolar mania of both adolescents and adults and points to possible treatment with anti-inflammatory agents.