Indicators of stress hematopoiesis in the blood predict COVID-19 progression in patients over 65 years old

ABSTRACT

Objectives

Advanced age is a well-established risk factor for severe coronavirus disease 2019 (COVID-19). Exacerbated inflammation affects multiple organs, among which hematopoiesis responds by increased output of various cells. We aimed to determine the association between COVID-19 progression and large immature cell (LIC) counts, changes in erythrocyte and platelet distribution widths (RDW, PDW) with reference to patients’ age.

Methods

A total of 755 patients with complete blood cell (CBC) analysis in the first 24 h of hospitalization were enrolled. Patients were divided into two groups: under and above 65 years of age.

Results

The LIC counts were different in both groups (p < 0.003). However, only the senior patients had markedly different values of RDW and PDW (p < 0.001). The receiver operating characteristic (ROC) curve analysis provided increased LIC (AUC = 0.600), RDW (AUC = 0.609), PDW (AUC = 0.556), and platelet to LIC ratio (AUC = 0.634) as significant in discriminating outcome in the older group. Importantly, these results were not repeated in the younger patients. In the elderly, the progression was predicted with LIC cut-off at ≥ 0.305 × 109/L (OR = 3.166) and RDW over 12.15% (OR = 2.081).

Discussion

Aging is characterized by a decline in immunological competence with a compromised control of inflammation leading to a proinflammatory state. This background together with the actions of pathogens may lead to emergency myelopoiesis.

Conclusion

Our results point to the important differences between age groups regarding CBC-related parameters of stress hematopoiesis during severe infection. Higher LIC, RDW and PDW levels were reliable in the early identification of COVID-19 progression only in the elderly.

KEYWORDS:

• Large immature cells
• red cell distribution width
• platelet distribution width
• elderly
• severe acute respiratory syndrome coronavirus 2
• platelets
• neutrophils

Plain Language Summary

Coronavirus disease 2019 (COVID-19) affects multiple organs, among which is the bone marrow, an organ that produces blood cells. Because of the infection, the bone marrow function changes and may prematurely release cells before they achieve complete maturation. The goal of our study was to assess whether there is a difference in the numbers of these premature cells between patients according to their age. It is well-known that aging is associated with multiple alterations in the immune response and control of the inflammation process. Therefore, there are differences in numbers and responsiveness of blood cells to infections depending on the patient’s age. We discovered that there are significant differences in the numbers of premature cells in the blood between the elderly, over 65 years of age, and younger COVID-19 patients. Moreover, these changes can be used as a prediction tool regarding the COVID-19 progression and adverse outcomes. There are also changes in the numbers and function of platelets and red blood cells (RBC) that are partly related to the changes in the bone marrow and that are seen in a blood test as an increase in a range of platelets and RBC size (parameters called: platelet distribution width – PDW, and RBC distribution width – RDW). These two parameters were also significantly associated with the COVID-19 progression only in the elderly group. A RDW value of over 12.15% carried a 2-times higher risk of COVID-19 progression.

Introduction

Exacerbated inflammatory reaction and inadequate immune response play decisive roles in the pathogenesis of severe coronavirus disease 2019 (COVID-19). Patients with COVID-19 are a representative population of severe acute coronavirus syndrome (SARS), a respiratory disease with a potential to progress to an acute respiratory distress syndrome (ARDS), septic shock, peculiar coagulopathy, and multi-organ failure. It is suggested that SARS-CoV2 affects hematopoiesis by directly invading the hematopoietic progenitor cells as they express angiotensin-converting enzyme-2 (ACE2) surface protein or indirectly through the induction of systemic inflammatory response with excessive release of cytokines. ACE2 expression peaks during erythropoiesis, making the early erythroid progenitors vulnerable to infection. The virus was detected in the progenitors, while an increase in circulating nucleated red cells was reported in intensive care COVID-19 patients [1–6].

A number of cytokines affect the proliferation and differentiation of the hematopoietic stem and progenitor cells (HSPCs), some of the well-known are interleukin (IL) 6, tumor necrosis factor alpha (TNF-α), interferon (IFN) -α/β, IFN-γ, transforming growth factor (TGF)-β, granulocyte colony stimulating factor (G-CSF), etc [7–10]. A set of HSPCs may participate directly by sensing these cytokines and together with toll-like receptor stimulation by the pathogen they conjoin to produce diverse cytokines through the NF-κB pathway signaling, upon which granulopoiesis and monopoiesis are induced [4,7].

Tumor necrosis factor α appears to be a major pro-survival factor for HSPCs [10]. It activates the canonical NF-κB pathway, which is associated with myeloid cell proliferation in so called ‘stress’ or ‘emergency’ hematopoiesis following a hematopoietic insult, such as the systemic inflammation resulting from an infection [4,7,8]. Interleukin-6 is one of the key mediators regulating inflammation, with markedly elevated levels in COVID-19 and correlating with the severity of the disease [3,11]. Interleukin-6 is also an important regulator of myeloid differentiation and proliferation, with a noted role in myeloid recovery during neutropenia [7].

Inflammatory-driven activation of hematopoiesis increases the output of mature immune cells and sometimes systemic inflammation may induce the recruitment of immature cells as well, which is observed in critically ill COVID-19 patients [7,8,12–14]. The presence of early stages of myeloid and lymphoid lineages raises the count of large immature cells (LIC) within a complete blood count (CBC). Besides, there is enhanced generation of platelets (megakaryocytes fragmentation) and suppression of erythropoiesis, seen as alterations of their counts and derangements in size [3,8,11–16]. Accordingly, higher values of erythrocytes (red blood cells) distribution width (RDW) and platelet distribution width (PDW) are observed in patients with severe COVID-19 [14,17–21].

So far, changes in the lymphocyte, neutrophil and platelet counts are reported as very useful indicators of COVID-19 severity and prognosis in many studies [2,14,20,21]. In addition, immature populations in circulation corroborate distinctive profiles of potential prognostic value. However, fewer investigations assessed the prognostic relevance of these cells in the elderly [2,11,14,17,18,22–24].

Well established risk factors for severe COVID-19 are advanced age, over 65 years, and co-morbid diseases, such as hypertension, diabetes and chronic kidney disease, with associated higher risk for death [14,25–28]. It is well known that functions of the immune system and its regulations are progressively debilitated with aging, which predisposes to chronic low-grade inflammation and counteracting compensatory immunosuppression. Consequences of these functional alterations are compromised response to various threats and inadequate hematopoietic homeostasis [13,29].

Importantly, early and accurate assessment of the severity of coronavirus infection is decisive for appropriate management of patients. Therefore, we aimed to determine the association of parameters of stress hematopoiesis within CBC, hematopoietic immature cell counts and changes in erythrocyte and platelet distribution widths, with the progression of COVID-19. We evaluated these alterations according to the patients’ age, by dividing them into groups older or younger than 65 years of age.

Methods

A total of 755 adult patients hospitalized at the University Clinical Center of Nis (Serbia) were enrolled in this observational study. Eligible participants were those with an acute respiratory tract infection and confirmed diagnosis of COVID-19 (positive PCR testing or characteristic clinical picture with clear X-ray viral pneumonia) and CBC analysis performed within the first 24 h of hospitalization. COVID-19 progression was defined as a worsening of disease that required admission to an intensive care unit or a lethal outcome. Patients who needed immediate invasive mechanical ventilation were excluded from the enrollment, as well as those with an acute organ failure other than lungs, arterial hemoglobin O2 saturation (SaO2) below 85%, body temperature < 35 or > 39 °C, hypotension (< 60 mmHg) or other signs of shock.

At the time of data collection (April – May 2021), the Alpha variant (B.1.1.7) of the SARS-CoV-2 was prevailing in Serbia [30]. All patients were treated per the guidelines and current protocols for COVID-19.

Patients were divided into two groups according to their years of age: a group under and above 65 years.

Routine laboratory examination consisted of CBC and biochemical parameters analyses, performed using the CELL-DYN Ruby Hematology Analyzer - Abbott Diagnostics automated hematology analyzer (Illinois, U.S.A.).

Ethics

The study has been performed according to the Declaration of Helsinki, all patients have signed an informed consent and the study was approved by the Institution Ethics Review Board of the University Clinical Center of Nis, Serbia (No. 11170 from April 2, 2021).

Statistical analyses

The normality of distribution was tested using the Shapiro-Willks test. Continuous variables are presented as median ± interquartile range (IQR) or mean ± standard deviation (SD) and categorical variables like numbers (n) and percentages (%). Categorical variables were analyzed using the Chi-square test.

The CBC parameters were tested by analysis of variance, receiver operating characteristic curve analysis, and binomial logistic regressions. One-way analysis of variance (ANOVA) was applied to compare associations between COVID-19 progression and CBC parameters. Variables which differed significantly in ANOVA tests were included in an area under the receiver operating characteristic (ROC) curve analysis to assess their discriminative ability and cut-off points, using Youden’s index. Binomial logistic regression, univariable and multivariable, were applied to variables with statistically significant areas under the curve (AUC) in the ROC curve analysis. Variables were assessed for collinearity and outliers. Statistical significance was set at p < 0.05. Statistical tests were performed using the IBM SPSS 25.0 software (SPSS Inc, Chicago, USA).

Results

A total of 755 consecutive COVID-19 patients were included in the study; 418 (55.4%) males and 337 (44.6%) females, with a median age of 70.0 ± 18.0 years. The median duration of symptoms was eight days. At the hospital admission, the median SaO2 was 93.0 ± 8.75%. At least one comorbidity was recorded in 77.9% of the patients, the most prevalent being hypertension in 60.7% and diabetes mellitus type 2 in 20.7%. COVID-19 progressed in 287 patients (38.0%), 271 (35.9%) died. Older age (65+), lower SaO2 level, C-reactive protein (CRP) levels, as well as higher lactate dehydrogenase (LDH) levels were significantly associated with the disease progression in all patients (Table 1). In all further analyses performed, we did not determine significant differences between genders in our study, both in all patients compared to outcome and when divided by age groups (p < 0.05).

Table 1. Differences between baseline characteristics in all patients according to COVID-19 progression (One-way ANOVA).

Parameter (mean ± SD)	COVID-19 progression		p value
	Yes	No
Age (years)	70.95 ± 13.29	65.78 ± 13.56	0.000
SaO2 (mmHg)	84.83 ± 12.23	92.99 ± 5.14	0.000
CRP (mg/L)	120.17 ± 80.72	104.21 ± 98.41	0.028
LDH (mg/L)	1026.62 ± 792.63	749.13 ± 305.52	0.000
IL-6 (pg/ml)	146.69 ± 331.98	65.42 ± 116.88	0.041
The CBC parameters
LIC (x10^9/L)	0.240 ± 142	0.188 ± 91	0.000
RDW (%)	12.03 ± 0.97	11.66 ± 0.74	0.000
PDW (%)	17.58 ± 3.13	16.72 ± 2.72	0.000
RBC (x10^12/L)	4.442 ± 0.746	4.532 ± 0.767	0.111
HGB (g/L)	133.83 ± 23.17	135.48 ± 19.06	0.289
HCT (%)	40.5 ± 6.65	40.8 ± 5.35	0.538
MCV (fL)	91.40 ± 6.09	90.81 ± 5.99	0.192
WBC (x10^9/L)	10.868 ± 6.854	9.137 ± 11.883	0.025
NEU (x10^9/L)	8.812 ± 5.733	6.712 ± 3.406	0.000
LY (x10^9/L)	1.347 ± 2.336	1.221 ± 736	0.280
MO (x10^9/L)	0.544 ± 474	0.493 ± 318	0.076
EO (x10^9/L)	0.100 ± 85	0.081 ± 59	0.000
BA (x10^9/L)	0.094 ± 160	0.076 ± 302	0.320
PLT (x10^9/L)	220.23 ± 109.15	240.91 ± 114.29	0.014
PCT (%)	0.203 ± 0.09	0.217 ± 0.09	0.047
MPV (fL)	9.39 ± 1.23	9.17 ± 0.99	0.010
P/LIC (x10^9/L)	1.271,7 ± 1.483,9	1.549,8 ± 1.174,6	0.012

CBC – complete blood counts, SaO2 – Hemoglobin saturation with O2, CRP – C reactive protein, LDH – lactate dehydrogenase, IL-6 – interleukin 6. LIC - large immature cells, RDW - red blood cell distribution width, PDW - platelets distribution width, RBC – erythrocytes, HGB – hemoglobin, HCT – hematocrit, MCV – mean cell volume, WBC – leukocytes, NEU - neutrophils, LY - lymphocytes, MO - monocytes, EO - eosinophils, BA - basophils, PLT – platelets, PCT – plateletcrit, MPV – mean platelet volume.

Comparison of the patients according to age

There were 487 (64.5%) patients older than 65 years (males 51.5%, females 48.5%). Among them, 214 (43.9%) suffered COVID-19 progression (mean age 77.0 ± 7.1), compared to individuals younger than 65 years of age (27.1%) (mean age 52.8 ± 10.6), which was significantly different p = 0.000 (Table 2). Patients over 65 years who progressed had markedly elevated levels of IL-6 compared to others (p = 0.052), unlike younger patients (Table 3).

Table 2. Differences between baseline parameters according to the patients’ age.

	Age ≥ 65 year	Age < 65 year	p value
COVID-19 progression	43.9%	27.1%	0.000
Males, n (%)	51.5%	62.4%	0.005
Comorbidity (%)Hypertension T2DM	84.8% 66.5% 23.0%	65.0% 49.2% 16.2%	0.000 0.000 0.030
SaO2 < 90%	41.5%	27.9%	0.000
CRP (mg/L)	110.74 ± 73.02	108.94 ± 119.97	0.779
LDH (mg/L)	850.54 ± 642.87	859.23 ± 353.76	0.805
IL-6 (pg/ml)	139.66 ± 339.01	77.05 ± 130.49	0.116

SaO2 – Hemoglobin saturation with O2, T2DM – type 2 diabetes mellitus. CRP – C reactive protein, LDH – lactate dehydrogenase, IL-6 – interleukin 6.

Table 3. Differences between baseline characteristics according to the COVID-19 progression (One-way ANOVA).

	Age ≥ 65 year		p value	Age < 65 year		p value
Progression	Yes	No		Yes	No
SatO2 (mmHg)	84.34 ± 12.37	92.33 ± 5.67	0.000	86.07 ± 11.86	93.91 ± 4.13	0.000
CRP (mg/L)	119.98 ± 80.56	103.98 ± 66.32	0.022	121.54 ± 82.05	104.37 ± 130.92	0.316
LDH (mg/L)	1016.78 ± 911.56	725.17 ± 295.08	0.000	1054.28 ± 372.56	784.15 ± 316.81	0.000
IL-6 (pg/ml)	189.66 ± 409.72	44.04 ± 44.53	0.052	77.79 ± 111.50	76.46 ± 145.18	0.963
CBC parameters
LIC (x10^9/L)	241.47 ± 143.73	187.25 ± 92.26	0.000	236.72 ± 138.66	190.71 ± 91.98	0.003
RDW (%)	12.17 ± 0.98	11.78 ± 0.75	0.000	11.65 ± 0.84	11.49 ± 0.71	0.139
PDW (%)	17.88 ± 3.14	16.99 ± 2.71	0.001	16.68 ± 2.98	16.34 ± 2.70	0.379
RBC	4.388 ± 0.746	4.427 ± 0.866	0.596	4.567 ± 0.686	4.681 ± 0.537	0.259
HGB	132.42 ± 24.28	132.76 ± 19.98	0.862	137.57 ± 18.92	139.34 ± 17.06	0.467
HCT	40.14 ± 6.98	40.06 ± 5.69	0.891	41.48 ± 5.43	41.82 ± 5.67	0.619
MCV	91.64 ± 6.62	91.71 ± 6.40	0.906	90.67 ± 4.21	89.53 ± 5.13	0.092
WBC	10.904 ± 7.113	9.565 ± 15.240	0.236	10.587 ± 5.933	8.552 ± 3.694	0.001
NEU	8.796 ± 5.771	6.662 ± 3.469	0.000	8.727 ± 5.587	6.794 ± 3.329	0.001
LY	1.400 ± 2.619	1.262 ± 852	0.417	1.156 ± 1.131	1.164 ± 530	0.936
MO	550 ± 512	517 ± 353	0.404	520 ± 341	458 ± 257	0.116
EO	98.9 ± 73.3	85.2 ± 63.7	0.030	103 ± 113	76.5 ± 51.1	0.011
BA	99.4 ± 17.9	88.2 ± 39.3	0.699	77.6 ± 70.7	57.9 ± 36.7	0.003
PLT	211.691 ± 105.607	239.809 ± 117.779	0.007	243.639 ± 116.083	243.113 ± 109.401	0.973
PCT	19.74 ± 8.86	21.73 ± 9.52	0.019	21.76 ± 9.61	21.63 ± 8.48	0.912
MPV	9.46 ± 1.26	9.22 ± 1.07	0.027	9.15 ± 1.11	9.10 ± 0.89	0.704
P/LIC	1259.16 ± 1642.86	1576.92 ± 1348.57	0.028	1307.97 ± 889.97	1517.07 ± 894.63	0.108

SatO2 – Hemoglobin saturation with O2, CRP – C reactive protein, LDH – lactate dehydrogenase, IL-6 – interleukin 6. CBC – complete blood counts, LIC - large immature cells, RDW - red blood cell distribution width, PDW - platelet distribution width, RBC – erythrocytes, HGB – hemoglobin, HCT – hematocrit, MCV – mean cell volume, WBC – leukocytes, NEU - neutrophils, LY - lymphocytes, MO - monocytes, EO - eosinophils, BA - basophils, PLT – platelets, PCT – plateletcrit, MPV – mean platelet volume.

Large immature cell counts were significantly different in both age groups considering COVID-19 progression compared to survivors (p < 0.003). However, only the senior patients had markedly different values of RDW (p = 0.000) and PDW (p = 0.001) in the progression (Table 3).

Lymphopenia was present in all patients but without significance according to the progression. Both groups had increased neutrophil and eosinophil counts (p < 0.030), and only the younger group showed markedly increased total leukocyte and basophil granulocytes counts (WBC) linked to the progression (p < 0.003) (Table 3).

Besides PDW, other platelet-related parameters (their count, plateletcrit and mean volume) were significantly different (p < 0.027) between outcomes only among older patients (Table 3). Moreover, significant correlations were determined between platelet parameters (p = 0.000). Decreased platelet counts positively correlated with plateletcrit (r = 0.970), inversely with mean platelet volume (MPV) (r = −0.438) and PDW levels (r = −0.478).

Receiver operating characteristic (ROC) curve analysis

Significantly changed CBC variables are further tested by ROC curve analysis to determine their diagnostic ability and optimal cut-off points for the prediction of an outcome.

The ROC curve analysis provided increased LIC, RDW and PDW values as significant (p < 0.020) in identifying COVID-19 progression (area under the curve (AUC) = 0.600, AUC = 0.609 and AUC = 0.556, respectively), as well as decreased ratio of platelets to LIC (P/LIC) with AUC = 0.622 (p = 0.000) (Figure 1). Other CBC parameters provided significant but rather poor performance, including WBC: AUC = 0.586, p = 0.000, neutrophils: AUC = 0.603, p = 0.000, eosinophils: AUC = 0.560, p = 0.003, lymphocytes: AUC = 0.583, p = 0.000, platelets: AUC = 0.560, p = 0.011, platelcrit: AUC = 0.547, p = 0.029, and MPV: AUC = 0.547, p = 0.029.

Figure 1. The ROC curve analyses presenting the complete blood counts parameters significantly associated with COVID-19 progression in all patients. Figure 1. Legend: ROC - receiver operating characteristic, AUC – area under the curve, LIC - large immature cells, RDW - red blood cell distribution width, PDW - platelets distribution width, WBC – leukocytes, NEU - neutrophils, PLT – platelets, PCT – plateletcrit, P/LIC – ratio of platelets and LIC.

When divided according to the age groups, only the results of WBC, neutrophils, and lymphocyte counts stayed significant in both groups. The rest of the tested parameters remained significant in discriminating outcome only in the older group, including: LIC (AUC = 0.596, p = 0.001), RDW (AUC = 0.606, p = 0.000), and PDW (AUC = 0.577, p = 0.004). The ratio P/LIC demonstrated the best predictive ability with AUC = 0.634 (0.580-0.687) (p = 0.000). Besides, significant results remained for the platelet parameters, while eosinophil counts did not show predictive ability.

Immature cell count was not able to discriminate COVID-19 progression in patients younger than 65 years, nor were the measures for distribution widths and platelet parameters. Significant variables in this group comprised: WBC (AUC = 0.608, p = 0.007), neutrophils (AUC = 0.610, p = 0.006), and lymphocytes (AUC = 0.600, p = 0.012).

Logistic regression analysis

We further analyzed variables that provided significant discriminatory ability in the ROC curve analysis. The variables were tested as continuous and at the cut-off points according to Youden’s index. All regressions were adjusted to SaO2 level, age and gender.

Univariate binomial regression provided levels of LIC, RDW, PDW, neutrophil and platelet counts, PCT, MPV and P/LIC as significant in predicting COVID-19 progression in all patients and the older group. Total leukocytes and neutrophils were significant in the younger group.

Next, the variables were tested for collinearity and then in multivariate binomial logistic regression. In the analyses with continuous values in all patients and the older group, LIC, RDW, neutrophil and platelet counts turned out significant in predicting COVID-19 progression. However, WBC and neutrophils were not able to predict the outcome in the younger group.

Specifically, LIC was significant (p = 0.004) with the odds (OR) of 1.027 (95%CI: 1.009-1.045), similarly in the older group: OR = 1.035 (95%CI:1.012-1.057), p = 0.002. Values of RDW showed OR = 1.756 (95%CI: 1.385-2.226) for all participants and OR = 1.827 (95%CI: 1.375-2.429) in the elderly (p = 0.000). Neutrophils displayed OR = 1.084 (95%CI: 1.033-1.138, p = 0.001) and OR = 1.094 (95%CI: 1.033-1.158, p = 0.002), while platelet counts provided the lowest odds when decreased, OR = 0.997 (95%CI: 0.995-0.999, p = 0.002) and OR = 0.996 (95%CI: 0.993-0.999, p = 0.003), respectively. Moreover, the results did not change substantially when additionally adjusted to the values of LDH and CRP.

The LIC count cut-off at ≥ 0.305 x109/L (AUC sensitivity 30.8%, specificity 89.0%) significantly predicted COVID-19 progression (p = 0.000) with 2.511 (95%CI: 1.551-4.066), and 3.166 (95%CI: 1.747-5.736) times greater odds in the elderly. RDW higher than 12.15% provided 2.081-fold greater odds for disease progression in the elderly, while the cut-off at 9.495,0 x109/L for neutrophil counts also showed double chances for progression. Results for platelet counts were modest (Table 4).

Table 4. Cut-off points from ROC curve analysis and results of the binomial logistic regressions to COVID-19 progression.

	Cut-offs, Youden’s index	Sen / Spec. (%)	Wald	Odds	95%CI		p value
					Lower	Upper
LIC x10^9/L	0.305	30.8 / 89.0	8.458	2.511	1.551	4.066	0.000
RDW (%)	11.55	68.7 / 48.7	13.174	2.110	1.410	3.159	0.000
NEU x10^9/L	7.665,0	47.7 / 71.6	6.942	1.650	1.137	2.395	0.008
PLT x10^9/L	184.500	44.3 / 64.9	10.071	0.556	0.387	0.799	0.002
Aged ≥ 65
LIC x10^9/L	0.305	30.4 / 89.1	14.442	3.166	1.747	5.736	0.000
RDW (%)	12.15	40.8 / 74.3	8.314	2.081	1.265	3.425	0.004
NEU x10^9/L	9.495,0	35.5 / 83.9	8.317	2.137	1.275	3.581	0.004
PLT x10^9/L	178.500	46.3 / 67.3	9.799	0.494	0.317	0.768	0.002

Sen/Spec. – sensitivity / specificity, CI – confidence interval, LIC - large immature cells, RDW - red blood cell distribution width, NEU - neutrophils, PLT – platelets.

Discussion

In systemic inflammatory response syndrome (SIRS), the HSCs and especially erythro-myeloid progenitors respond to cytokine signaling and pathogen associated molecular patterns directly to generate adequate support for the immune response [7,9]. Therefore, the presence of immature myeloid cells in CBC is a common finding of acute and severe viral infections. Dysregulated hematopoiesis with profound alterations in the myeloid compartment is described in critical COVID-19 and may even contribute to severity of disease and ARDS development [8,11].

All patients with in-hospital progression of COVID-19 in this study had significantly higher counts of LIC, neutrophils and eosinophils compared to survivors. Lymphopenia was present in all patients but without predictive ability for disease outcome. The senior patients with severe illness had markedly increased values of LIC, RDW and PDW, as well as lower measures of platelet parameters (count, plateletcrit and volume), which were not detected in the younger patients. Similarly, the P/LIC ratio showed marked ability for prediction of disease progression and in-hospital mortality when assessed to all patients and in the elderly, but not in those less than 65 years old. Interestingly, increased concentrations of IL-6 correlated with COVID-19 progression in the elderly.

The number of elderly patients with sepsis is rising worldwide [16]. A distinctive trait of aging is a decline in immunological competence with a compromised balance between immune response and inflammation control. The proinflammatory state of aging is associated with increased blood concentrations of inflammatory biomarkers, including TNF-α, IL-6, IL-1α, and IL-15 [29,31–33]. CD4 + cells are considered the main source of altered cytokine expression during aging. The numbers of naive CD4(+) recent thymic emigrants are low, the Th17/Treg balance is disturbed, there is heightened NF-κB induction in senescent CD8 + CD28− T cells, while diminished NF-κB response in immunocompetent cells, with a resultant state of inflammaging [29,31–34].

Emergency myelopoiesis in severe infections, including COVID-19, is proposed to be linked to immunosuppressive functions of B and T cells, and to be programed to anti-inflammatory phenotype. Severe COVID-19 was enriched with neutrophil precursors and dysfunctional mature neutrophils and monocytes [8,11–13]. Moreover, it is demonstrated that different SARS-CoV-2 variants can distinctively impact erythropoiesis. An abundance of CD71 + erythroid cells in the circulating blood was most pronounced in those infected with the original Wuhan strain, followed by the Delta and Omicron variants. Given the immunosuppressive properties of CD71 + cells, it can be speculated that besides hypoxia, the viral strain related clinical consequences are reflected in altered immunity. A strong negative correlation was observed between CD71 + and T and B cell proportions [35,36].

Results of the transcriptomic study of circulating immune cells in critically ill COVID-19 patients revealed profiles related to immature myeloid lineages with possible induction of trained immunity. Vadillo et al. [8] identified immature myeloid cell populations as predominate in patients with severe SARS-CoV-2 infection. And these cells comprise mostly metamyelocytes and immature monocytes (∼20%), promyelocytes-myelocytes and band neutrophils (∼10%). Lymphoid cell populations were present at low frequencies, in contrast to healthy individuals. Immature granulocytes index greater than 3% was shown as an early marker of sepsis in general, while a cutoff point of 2.0% was able to exclude sepsis with a very high specificity (90.9%) [12].

Large immature cell counts were one of the most reliable predictors of disease progression in this study, showing 2.5 and 3-fold higher odds with cell numbers ≥ 305,0 x109/L and high specificity (89%). This parameter of stressed hematopoiesis remained significant in patients with advanced age on the contrary to younger ones.

Unlike our results, in a recent study, distinctive profiles of hematological differential were reported, with marked emergency myeloid hematopoiesis at the expense of functional lymphoid innate cells, in patients between 21-64 years of age at 7 days follow-up. However, it was speculated that pre-existing metabolic conditions made these individuals prone to the development of an imbalanced myeloid phenotype [26].

Acute changes in hemoglobin levels with anemia are described in patients with sepsis and correlate to in-hospital mortality [16]. Proposed mechanisms include reduced production of erythrocytes due to SIRS, together with their increased peripheral destruction. Impaired oxygen transportation can add to the deterioration of patients with sepsis [16–18,36].

Erythrocyte related parameters in the blood are reported as prognostic indicators for COVID-19 severity [19]. There is a confirmed association between elevated RDW and COVID-19 severity and mortality [17,19], similarly to our results. RDW > 14.5% has been associated with increased risk of mortality 2.73 times [36]. In our work, RDW > 12.2% provided 2.08-fold greater odds for COVID-19 progression in the elderly.

Several studies reported an inverse association between platelet count and mortality in COVID-19, with decreasing platelet count signifying worse outcomes [1–3,14,15,20,37]. Low platelets were associated with an over 5-fold enhanced risk of severe COVID-19 [15]. The possible mechanisms of platelet decline include an inhibition of hematopoiesis, increased peripheral destruction of platelets, due to exaggerated inflammation or antibody related mechanisms, together with enhanced platelet consumption [1–3,21,27,37,38]. Apparent change in platelet counts occurs immediately and early, within a week after admission, and closely following the dynamics of the COVID-19 pathophysiology [2,17,18,22,38]. In the recovery phase, there is a rise in thrombopoietin levels and platelet counts in the blood [19,37]. Described protuberances in the platelet homeostasis can be observed through the presence of larger and younger platelets in the circulation and consecutive changes in their distribution width, MPV and ratio to LIC [1,2,14,25,38]. Therefore, PDW is considered a marker of platelet function and activation and was recently presented as a novel biomarker of the severity of acute illnesses in internal medicine with its increased values reported in different diseases [27]. In our study, PDW value showed significance in the mean comparison tests but failed to provide substantial information regarding prediction of COVID-19 progression in further analysis.

Assessment of the markers’ usefulness in prediction of treatment outcomes would perhaps be more challenging, due to the differences in treatments recommended according to the COVID-19 severity, together with therapies applied for patients’ comorbidities.

Besides alterations in hematopoiesis, current evidence describes an enhanced platelet activation in aging, but with poorly understood mechanisms that underlie this abnormality. Recent findings emphasize dysregulation of the proteasome system and autophagy, senescence-associated secretory phenotype rich in pro-inflammatory cytokines, mitochondrial reprogramming as well as general inflammaging as the main contributors of platelet hyperactivity in aging. Therefore, the evidence merely supports the impact of platelet dysfunction and related susceptibility to thrombosis on COVID-19 severity and complications, that are inevitably related to the changes in the CBC parameters, especially in the elderly [39–42].

Conclusion

Immature cell population counts are rising in critically ill COVID-19 patients. Our results point to the important differences between age groups regarding LIC counts, RDW and PDW measures. In patients with advanced age (≥ 65 years), an increased LIC, RDW and PDW were highly associated with worse outcomes. The LIC, RDW and decreased ratio of platelets to LIC (P/LIC) showed best discriminatory ability, and represented independent predictors of disease progression only in the elderly. Immature cell counts, although increased, were not able to discriminate COVID-19 progression in patients younger than 65 years in this study.

We can conclude that CBC parameters related to stress hematopoiesis are more characteristic finding in patients with advanced age, probably due to the age-dependent alteration in the immune system responsiveness and regulation, and can be used as reliable predictors of severe disease and worse outcome in this population. Our results may help in the arrangement of personalized medical approaches, especially in assessing the early initiation of therapy.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available on reasonable request from the corresponding author, J. M.

Additional information

Funding

This work was supported by the Ministry of Education, Science and Technological Development of Republic of Serbia under Grant number 451-03-47/2023-01/200113; the Faculty of Medicine University of Nis, Serbia under Grant number 46 and 56.