Monocyte–Lymphocyte Ratio and Dysglycemia: A Retrospective, Cross-Sectional Study of the Saudi Population

Background: Abnormalities in fasting blood glucose (FBG) resulting in hypoglycemia (OG), impaired fasting glycemia (IFG), or hyperglycemia (HG) arise from disordered metabolic regulation caused in part by inflammation. To date, there is a dearth of evidence regarding the clinical utility of the monocyte–lymphocyte ratio (MLR), an emerging inflammatory index, in the management of dysglycemia. Methods: This retrospective, cross-sectional study explored MLR fluctuations as a function of glycemic control in 14,173 Saudi subjects. Data collected from 11 August 2014 to 18 July 2020 were retrieved from Al-Borg Medical Laboratories. Medians were compared by Mann–Whitney U or Kruskal–Wallis tests and the prevalence, relative risk (RR), and odds ratio (OR) were calculated. Results: MLR was significantly elevated in IFG (p < 0.0001) and HG (p < 0.05) groups compared to the normoglycemia (NG) group, and individuals with elevated MLR (>0.191) had significantly increased FBG (p < 0.001). The risk of IFG (RR = 1.12, 95% CI: 1.06–1.19, p < 0.0002) and HG (RR = 1.10, 95% CI: 1.01–1.20, p < 0.0216) was significantly increased if MLR was elevated, and individuals with elevated MLR were 1.17 times more likely to have IFG (OR = 1.17, 95% CI: 1.08–1.26, p < 0.0002) and 1.13 times more likely to have HG (OR = 1.13, 95% CI: 1.02–1.24, p < 0.0216). Conclusion: Elevated MLR is correlated with and carries a greater risk for IFG and HG. However, large prospective cohort studies are needed to establish the temporal relationship between MLR and FBG and to examine the prognostic value of this novel marker.


Introduction
Maintenance of plasma glucose levels requires complex metabolic regulation involving insulin, glucose uptake, glycogenesis, gluconeogenesis, glucagon, and glycogenolysis. Hypoglycemia (OG) is most often iatrogenic in nature, with type I diabetes mellitus (T1DM) patients being at a higher risk than those with T2DM. Other causes, however, include alcohol consumption, critical illness, and non-islet cell tumors [1]. On the other hand, disturbances in glucose homeostasis resulting in impaired fasting glycemia (IFG) or hyperglycemia (HG) often arise secondary to obesity, aging, fat-and carbohydrate-rich diets, sedentary lifestyle, or genetic predisposition. In these cases, increased fasting blood glucose (FBG) is associated with an inflammatory state characterized by insulin resistance, cytokine release, fatty acid exhaustion, mitochondrial dysfunction, and endoplasmic reticulum stress [2].
It has been demonstrated that macrophages and lymphocytes accumulate in the milieu of adipocytes [3], highlighting a central role of immune cells in obesity-induced

Statistical Analysis
The data were not normally distributed as revealed by D'Agostino and Pearson test and Kolmogorov-Smirnov test (p < 0.0001) and hence nonparametric tests were used for statistical analysis. Two groups were analyzed by Mann-Whitney U test while three or more were analyzed by Kruskal-Wallis test followed by Dunn's multiple comparisons test. Figures show medians ± interquartile range (IQR). Association between two variables was determined by Spearman's correlation and by calculations of the relative risk (RR) and odds ratio (OR). Sensitivity and specificity were examined by ROC curve analysis and area under the curve (AUC) determination. GraphPad Prism v9.2.0 (GraphPad Software, Inc., San Diego, CA, USA) was used for statistical analysis, and the cutoff for significance was set at a p value of <0.05.
These findings highlight the gender disparity observed in DM which could be, at least in part, explained by sex-specific monocyte physiology and differential sex steroid availability and its influence on insulin sensitivity and glucose homeostasis [12,30].

Statistical Analysis
The data were not normally distributed as revealed by D'Agostino and Pearson test and Kolmogorov-Smirnov test (p < 0.0001) and hence nonparametric tests were used for statistical analysis. Two groups were analyzed by Mann-Whitney U test while three or more were analyzed by Kruskal-Wallis test followed by Dunn's multiple comparisons test. Figures show medians ± interquartile range (IQR). Association between two variables was determined by Spearman's correlation and by calculations of the relative risk (RR) and odds ratio (OR). Sensitivity and specificity were examined by ROC curve analysis and area under the curve (AUC) determination. GraphPad Prism v9.2.0 (GraphPad Software, Inc., San Diego, CA, USA) was used for statistical analysis, and the cutoff for significance was set at a p value of <0.05.

FBG Is Significantly Increased in Individuals with Elevated MLR
Stratified by MLR, FBG is consistently elevated in H-MLR irrespective of gender which propounds a state of systemic inflammation synchronous with glucose build up.

Correlation between MLR and FBG
Although our simple linear regression model shown in Figure 2G found no relationship between MLR and FBG, the two variables seem to be related but fluctuations in FBG cannot be exclusively explained by those in MLR (R 2 = 0.0003, p < 0.0345). Analysis of the ROC curve indices ( Figure 2H) revealed an AUC of 0.5257 (p < 0.0001) reflecting poor diagnostic accuracy of MLR for HG.
Several factors may have been involved in the significant, but indirect association between MLR and FBG, most notably other inflammatory mediators, such as cytokines, fetuin-A, chemerin, and vaspin which lead to β-cell failure and insulin resistance [31]. The requirement of these factors for the observed association between MLR and FBG must be established in future studies.  These findings highlight the gender disparity observed in DM which could be, at least in part, explained by sex-specific monocyte physiology and differential sex steroid availability and its influence on insulin sensitivity and glucose homeostasis [12,30].  Figure 2E,F, respectively.

FBG Is Significantly Increased in Individuals with Elevated MLR
Stratified by MLR, FBG is consistently elevated in H-MLR irrespective of gender which propounds a state of systemic inflammation synchronous with glucose build up.

Correlation between MLR and FBG
Although our simple linear regression model shown in Figure 2G found no relationship between MLR and FBG, the two variables seem to be related but fluctuations in FBG cannot be exclusively explained by those in MLR (R 2 = 0.0003, p < 0.0345). Analysis of the ROC curve indices ( Figure 2H) revealed an AUC of 0.5257 (p < 0.0001) reflecting poor diagnostic accuracy of MLR for HG.
Several factors may have been involved in the significant, but indirect association between MLR and FBG, most notably other inflammatory mediators, such as cytokines, fetuin-A, chemerin, and vaspin which lead to β-cell failure and insulin resistance [31]. The requirement of these factors for the observed association between MLR and FBG must be established in future studies.  Figure 3B-D. When either males or females were analyzed alone, no significant differences in MLR were found among the different glycemic states except in female young adults ( Figure 3J) who exhibited significant elevation in MLR from 0.186 (0.151-0.230) in the NG group to 0.194 (0.157-0.237) in the IFG group (p < 0.05).

Differential Influence of Age and Gender on MLR
Altogether, these results support the influence of age on glycemic control with older groups of both genders exhibiting significant MLR elevations in IFG and HG. Further, MLR is a more sensitive marker in adults and elderlies as it was significantly elevated in IFG in these groups. Age-related differences in immune regulation resulting in an aggravated inflammatory response and energy depletion by lymphocytes [32,33] may account for this observation.

Elevated MLR Is Associated with Increased Risk of IFG and HG
Analysis of the overall and within-group prevalence of H-MLR in the study population showed that it was less prevalent in the NG group and more prevalent in the OG, IFG, and HG groups with increases of 28.0%, 9.39%, and 5.40%, respectively (Table  1). Moreover, as shown in Table 2, elevated MLR was associated with increased risk of IFG (RR = 1.12, 95% CI: 1.06-1.19, p < 0.0002) and HG (RR = 1.10, 95% CI: 1.01-1.20, p < 0.0216). Additionally, elevated MLR was 1.17 times more likely to fall into the IFG group (OR = 1.17, 95% CI: 1.08-1.26, p < 0.0002) and 1.13 times more likely to suffer from HG (OR = 1.13, 95% CI: 1.02-1.24, p < 0.0216).
Collectively, these findings indicate that increased MLR predisposes to IFG and HG, possibly due to the inflammatory potential of monocytes.  Altogether, these results support the influence of age on glycemic control with older groups of both genders exhibiting significant MLR elevations in IFG and HG. Further, MLR is a more sensitive marker in adults and elderlies as it was significantly elevated in IFG in these groups. Age-related differences in immune regulation resulting in an aggravated inflammatory response and energy depletion by lymphocytes [32,33] may account for this observation.

Elevated MLR Is Associated with Increased Risk of IFG and HG
Analysis of the overall and within-group prevalence of H-MLR in the study population showed that it was less prevalent in the NG group and more prevalent in the OG, IFG, and HG groups with increases of 28.0%, 9.39%, and 5.40%, respectively (Table 1). Moreover, as shown in Table 2, elevated MLR was associated with increased risk of IFG (RR = 1.12, 95% CI: 1.06-1.19, p < 0.0002) and HG (RR = 1.10, 95% CI: 1.01-1.20, p < 0.0216). Additionally, elevated MLR was 1.17 times more likely to fall into the IFG group (OR = 1.17, 95% CI: 1.08-1.26, p < 0.0002) and 1.13 times more likely to suffer from HG (OR = 1.13, 95% CI: 1.02-1.24, p < 0.0216).  Collectively, these findings indicate that increased MLR predisposes to IFG and HG, possibly due to the inflammatory potential of monocytes.

MLR Is Not Influenced by HbA1c Levels
Unlike FBG, examination of MLR in light of HbA1c levels found no variance or association between the two parameters ( Figure 4A-H). While HbA1c reflects long-term (2-3 months) glycemic control, interindividual variation in this indicator may have accounted for its lack of association with MLR. Variables other than FBG, such as gender, red cell turnover, blood acidity, and race, have been described as contributing factors [34] highlighting the discord between HbA1c and FBG.

MLR Is Not Influenced by HbA1c Levels
Unlike FBG, examination of MLR in light of HbA1c levels found no variance or association between the two parameters ( Figure 4A-H). While HbA1c reflects long-term (2-3 months) glycemic control, interindividual variation in this indicator may have accounted for its lack of association with MLR. Variables other than FBG, such as gender, red cell turnover, blood acidity, and race, have been described as contributing factors [34] highlighting the discord between HbA1c and FBG.

A.
B. C.

Discussion
Inexpensive, quick, and reproducible indices are highly desired in clinical practice. In this work, we show, for the first time, that increased MLR is associated with IFG and HG in the Saudi population. We have recently reported that the neutrophil-lymphocyte ratio (NLR) is similarly elevated and more prevalent in hyperglycemic Saudis [27] which suggests concurrent systemic inflammation. MLR, on the other hand, is also significantly raised in IFG (Figure 2A-C) making it a more sensitive marker of glucose status compared to NLR. Likewise, C-reactive protein (CRP) is superior to NLR in discriminating prediabetics [35,36], which points to a possible advantage in assessing both MLR and CRP in DM management.
The Saudi Abnormal Glucose Metabolism and Diabetes Impact Study (SAUDI-DM) revealed that IFG and DM affect 22.6% and 11.9% of the Saudi population, respectively [37]. Stratified by MLR, our study found that subjects demonstrated significantly elevated FBG with increased MLR compared to those with normal MLR (Figure 2D-F). In contrast, we found no relation between MLR and HbA1c (Figure 4). The differential impact of NLR on HbA1c was also investigated in Brazilian and Turkish diabetics and no relation was established [38,39]. Since HbA1c reflects long-term glycemic control, it seems plausible to assume that it is less susceptible to transient or acute inflammatory stimuli which may explain the lack of association with NLR and MLR in this and in other various reports.
The value of MLR in the management of DM extends to common complications despite its poor diagnostic performance ( Figure 2G,H). In diabetic kidney injury, high MLR was correlated with microalbuminuria and was thus a predictive marker for diabetic nephropathy [40,41]. MLR was similarly elevated in proliferative diabetic retinopathy and may thus have a prognostic value [42] in combination with fibrinogen levels [43]. In contrast, Ilhan et al. demonstrated that NLR but not MLR was significantly elevated in diabetic macular edema [44]. Accordingly, identifying the clinical contexts in which MLR offers a prognostic value for DM complications must be the focus of future studies.
Interestingly, when stratified by both age and gender, males exhibited no significant changes in FBG levels relative to MLR ( Figure 3E-H). In DM, there exists a gender disparity in the predisposition, prevalence, and disease progression among patients. For instance, men demonstrate an earlier peak in DM prevalence than women [45], and in the Middle East, DM is more common in females than in males [46]; an observation in opposition to the vast majority of the world. Since estrogens are essential for carbohydrate and lipid utilization, sex steroids may play a protective role against DM development as highlighted by the diminished incidence of DM in menopausal women on estrogen therapy [47]. Nevertheless, large, longitudinal studies are needed to characterize sex-specific determinants of DM susceptibility and development.
Evidence from a meta-analysis points to an inflammatory basis for non-affective psychoses manifested as increased MLR in bipolar disorder, major depressive disorder, and schizophrenia [21]. A variety of cells in the brain are differentiated from monocytes including macrophages in the meninges and choroid plexus and perivascular cells. These cells act in tandem with microglia in neuropathological conditions directed by inflammatory cytokines which also promote infiltration of the brain tissue with circulating monocytes [48,49].
Regulated fluctuations in white blood cell counts are essential to produce a modest inflammatory state that maintains successful pregnancy [50]. In preeclampsia, an exaggerated immune activation mediated through cytokines released from placental tissue precipitates macrophage deposition in the placenta leading to TNF-α-induced apoptosis of trophoblasts [23]. Moreover, it has been demonstrated that lymphocyte proliferation is inhibited in preeclampsia [51], which may explain the increased MLR in these patients. Recently, elevated MLR was found to predict unfavorable outcomes in pregnant women with HG [52].
MLR has also been investigated in the context of cardiovascular disease by several groups. In CAD, it has been shown that MLR could be useful as an independent risk factor and as a predictor of CAD severity with superior performance than NLR [22]. This was also confirmed in another study which found that MLR was an independent predictor of long-term major adverse cardiac events in myocardial infarction patients [53]. A significant correlation between MLR and the length of hospital stay was also reported in myocarditis patients [54]. Cheng et al. have recently shown that elevated MLR was associated with the development of stroke-associated pneumonia and therefore may be exploited for earlier intervention [24]. Moreover, MLR and NLR were independent predictors of saphenous vein graft disease in patients undergoing coronary artery bypass grafting [55].
In arthritis, MLR was significantly elevated during gout attacks and may thus serve as a predictive marker for the attack [56]. However, although increased MLR was associated with severe tuberculosis, it failed to predict response to treatment or mortality [57]. Congruently, MLR was not significantly altered in COVID-19 patients [58]. In dialysis patients with increased MLR, the risk of cardiovascular events and length of infectious disease hospitalization was significantly increased compared to those with normal MLR [59]. Notably, in a retrospective study of a very large cohort of critically ill patients, increased MLR was associated with an increased risk of mortality, the need for continuous renal replacement therapy, mechanical ventilation, and hospitalization [60].
MLR has also gained interest as a clinical aid in oncology as it was found to be significantly elevated in lymphoma [61] and ovarian cancer [62]. In the latter, calculating MLR prior to surgical intervention could be of clinical value in predicting stage, grade, and metastasis. Additionally, MLR was associated with serum CA-125 levels in a retrospective, cross-sectional study of Nigerian ovarian cancer patients [63]. Furthermore, MLR performed better than NLR in predicting the progression and overall survival of bladder cancer patients [64][65][66][67], and was invaluable for the prognosis and survival of cervical cancer and non-small-cell lung cancer patients [68,69]. Likewise, MLR was better than NLR in diagnosing colorectal cancer [70] and complemented prostate-specific antigen in predicting prostate cancer and in reducing false positive results [71].
The current report also demonstrates that IFG and HG are more prevalent in H-MLR compared to N-MLR subjects (Table 1) and that elevated MLR may be a risk factor for IFG and HG (Table 2), highlighting the potential role of MLR to predict and monitor the glycemic status in Saudis. In congruence, MLR was associated with the prevalence of proliferative diabetic retinopathy in an American population [42], and reduced monocyte counts were associated with retinopathy in diabetic Chinese subjects [72]. Indeed, causation remains to be interrogated in longitudinal studies.
Our study has numerous advantages, most notably the very large sample size and the streamlined generation, acquisition, and recordkeeping of laboratory data. Limitations, however, include the inability to determine the causality and temporal relation between FBG and MLR given the cross-sectional design of the study, and the unavailability of sufficient potential confounding variables including lifestyle habits, comorbidities, and medication intake.

Conclusions
In conclusion, our study identifies MLR as a novel, inexpensive, and reproducible biomarker in the pre-diabetic and diabetic Saudi population. Prospective studies must examine why FBG, but not HbA1c, is influenced by MLR fluctuations in addition to investigating other glycemic and inflammatory markers, such as insulin, glycated albumin, fructosamine, and gasdermins. Longitudinal studies are also needed to assess the potential impact of anthropometric variables on MLR and glycemic control.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Biomedical Ethics Unit at Al-Borg Medical Laboratories (protocol code #07/21 on 27 December 2021).
Informed Consent Statement: Patient consent was waived due to the retrospective design of the study and lack of access to personal identifiable information.
Data Availability Statement: Data is available from the corresponding author upon reasonable request, and with permission of Al-Borg Medical Laboratories.