Gestational age-specific hematological patterns in preterm infants following necrotizing enterocolitis

Abstract

Objective:

To investigate gestational age (GA) specific hematological and transfusion response patterns in preterm infants following necrotizing enterocolitis (NEC)

Design:

A retrospective study comparing hematological/transfusion information in three GA groups’ infants: Group A ≤28 weeks. Group B 28–32 weeks, Group C >32 weeks following necrotizing enterocolitis.

Results:

Group A infants responded with significantly higher WBC count, thrombocytopenia, higher absolute neutrophil, and higher absolute monocyte and lower absolute lymphocyte counts following NEC onset, received more blood transfusions before NEC onset (59.8% vs. 30.0%; p=0.007), and had higher odds of surgical NEC (OR 3.39 [95% CI 1.19–10.38]; p=0.02) than group C. One unit increase in absolute lymphocyte count on the day, and 24 hours following NEC was significantly associated with lower surgical NEC odds than groups C.

Conclusion:

The infant’s in-group A had significantly different hematological response patterns following NEC than infants with higher gestational age (group B and C).

Introduction

Necrotizing enterocolitis (NEC) affects about 5–10% of preterm infants with a birth weight ≤1500 grams [1,2]. NEC remains a leading cause of surgical intervention, morbidity, mortality, and resource utilization in premature infants [3–9]. NEC histopathology is characterized by extensive bowel necrosis, hemorrhage, and inflammation [5,10–13]. The developmental limitations in both the innate and adaptive immune systems in preterm infants increase the risk of NEC and inflammatory injury risk. In addition, NEC has been associated with various hematological changes [14] with altered counts of platelets, monocytes, neutrophils, and lymphocytes, and coagulopathy with diagnostic and prognostic significance [14–17]. The studies have shown thrombocytopenia [18], neutropenia [14,16,17,19], and lymphopenia [20] has been associated with poor outcomes.

Neonates with necrotizing enterocolitis commonly receive packed cell and platelet transfusions to treat severe anemia and thrombocytopenia [21] which can also play a role in exacerbating mucosal inflammation and poor clinical outcomes [22–24]. In addition, the clinical and animal reports have shown the effect of anemia and subsequent RBC transfusion on intestinal inflammatory response and NEC development [25–28].

We have recently reported nadir hematological and transfusion patterns following disease onset in neonates with medical/surgical necrotizing enterocolitis to forecast disease progression and outcome in the first 24 hours following the NEC onset [29]. However, the gestational age-specific hematological and transfusion patterns in preterm infants with NEC have not been well studied. Therefore, we speculate that preterm infants with lower gestational age (less than 28 weeks) would have different hematological responses and receive transfusion differently than those with higher gestational age ( more than 28 weeks).To address this knowledge gap, we aimed to determine the gestational age-specific hematological and transfusion patterns in infants with medical and surgical NEC.

Methods:

This retrospective study was done at the University of Mississippi Medical Center at Jackson, Mississippi, after approval by the Institutional Review Board (2017–0127), a level 4 neonatal intensive care unit, and a referral center for neonates with surgical NEC. A detailed review of the medical records between January 2000 and 2018 identified 336 patients with medical and surgical NEC (NEC Bell stage 2 and above). Infants with an intestinal anomaly, gastroschisis, and spontaneous intestinal perforation were excluded.

Clinical information:

We recorded demographic factors including gestational age, gender, birth weight, race (African-American, Caucasian, or Latino), mode of delivery (C-section / Vaginal delivery), APGAR scores at 5 minutes, outborn status, and small for gestational age status. We collected maternal data, including pregnancy-induced hypertension, chorioamnionitis, and antenatal steroids. We noted the NEC features such as pneumatosis seen on an abdominal radiograph, clinical presentation (abdominal distension, feeding intolerance, and bloody stools), and the age of onset.

We collected information on clinical variables such as patent ductus arteriosus, PDA treatment medically or surgically, C reactive protein after NEC onset, cholestasis (direct bilirubin >2 mg/dL), blood cultures collected at the time of NEC onset, use of inotropes at 24 hours, assisted ventilation, and antibiotic duration after NEC onset. In addition, we recorded information on the length of stay and mortality. Mortality was defined as death due to any cause prior to hospital discharge.

Hematological/Transfusion information:

We recorded complete blood cell count results from the patient chart before the NEC onset (last available CBC within 3–7 days before NEC onset), on the day of NEC onset, 24 hours,48 hours, 72 hours, 96 hours and at one week after onset. We collected data on absolute and relative (presented as percentages) values. If we had multiple CBC on the same day, we recorded values from what we assessed to be the most abnormal ones. We also collected RBC and platelet transfusion data before and after the NEC onset. In our unit, the on-call attending makes decisions to give RBC or platelet transfusions. In our unit, blood transfusions are considered if the hematocrit is ≤30 for neonates requiring respiratory support, defined as mean airway pressure >8 cm H2O and FiO2 >40% on a ventilator. In our practice, infants are considered for transfusion of platelet count is less than 30 × 10³/cm. Platelet transfusion is considered at counts less than 50 × 10³ /cm if the neonate is less than 1000 grams( less than seven days old) and has h/o previous intraventricular hemorrhage and lung bleed.

Statistical Methods:

Patient demographics, clinical and hematological information were compared between the different gestational age groups. Differences in continuous data were compared using the Kruskal-Wallis (Mann-Whitney U) test and summarized with medians and interquartile ranges (Q3-Q1). Categorical variables were compared using Chi-squared tests and reported with frequencies and percentages. To investigate the association between patients surgical NEC status and gestational age group, three multiple logistic regression was used on before the NEC onset, the day of NEC onset and 24 hours after NEC onset. WBC, Platelet count, ANC, ALC, AMC, gestational age group, hematocrit before NEC and PDA presence were used as predictors in the model while patient’s surgical NEC status was used as a binary outcome variable. Adjusted odds ratios and confidence intervals (aOR [95% CI]) were calculated and reported along with p-value. All the statistical analyses were performed in R statistical software (version 4.0.2; The R Foundation for Statistical Computing). A two sided p-value of less than 0.05 was considered significant.

Results:

Surgical NEC:

One hundred and eighty-four (N=184) infants with surgical NEC included for the analysis had a median birth weight of 780 gm [IQR 630; 1010]. The cohort had predominantly male (115/184, 62.5%) infants, belonged to African- American (139/184, 76.4%) race, born by Cesarean section (122/184, 66.3%) and were outborn (106/184, 59.9%). The median length of stay was 118 days [IQR 72.8; 178]. Sixty-one infants (33.3%) had died. The other clinical and demographic information is summarized in Table 1.

Table 1:

Demographics of Neonates with Surgical NEC by Gestational Age

	Total Cohort	GA <28 Weeks	GA 28–32 Weeks	GA >32 Weeks	p-value (Overall)	p1	p2	p3
	n = 184	n = 130	n = 33	n = 21
Prenatal Information
Pregnancy-Induced Hypertension, n (%)	39 (22.4%)	25 (20.5%)	10 (31.2%)	4 (20.0%)	0.462	0.291	1	0.57
Chorioamnionitis, n (%)	14 (8.64%)	14 (12.4%)	0 (0.00%)	0 (0.00%)	0.04	0.042	0.219	NA
Antenatal Steroids, n (%)	110 (62.9%)	84 (67.2%)	18 (58.1%)	8 (42.1%)	0.09	0.456	0.062	0.421
Infant Demographics
Birth Weight (grams; mean ± SD)	780 [630;1010]	710 [620;881]	1140 [898;1420]	2043 [1621;2530]	<0.001	<0.001	<0.001	<0.001
Small for Gestational Age, n (%)	59 (32.4%)	43 (33.3%)	12 (37.5%)	4 (19.0%)	0.343	0.813	0.291	0.26
Male Gender, n (%)	115 (62.5%)	76 (58.5%)	22 (66.7%)	17 (81.0%)	0.122	0.509	0.085	0.406
Ethnicity, n (%)					0.706	0.523	0.731	0.725
African American	139 (76.4%)	96 (74.4%)	26 (78.8%)	17 (85.0%)
Caucasian	36 (19.8%)	26 (20.2%)	7 (21.2%)	3 (15.0%)
Other	7 (3.85%)	7 (5.43%)	0 (0.00%)	0 (0.00%)
Mode of Delivery, n (%)					0.151	0.941	0.09	0.269
C-section	122 (66.3%)	90 (69.2%)	22 (66.7%)	10 (47.6%)
Vaginal	62 (33.7%)	40 (30.8%)	11 (33.3%)	11 (52.4%)
Apgar Score <6 at 5 Minutes, n (%)	39 (21.9%)	35 (27.6%)	2 (6.45%)	2 (10.0%)	0.013	0.024	0.16	0.64
Outborn, n (%)	106 (59.9%)	79 (62.7%)	15 (50.0%)	12 (57.1%)	0.427	0.285	0.808	0.827
Infant Medical Information Prior to NEC
Patent Ductus Arteriosus, n (%)	101 (55.8%)	82 (64.1%)	11 (34.4%)	8 (38.1%)	0.002	0.004	0.044	1
Patent Ductus Arteriosus, Indomethacin Treated, n (%)	30 (18.9%)	27 (24.8%)	2 (6.67%)	1 (5.00%)	0.017	0.056	0.073	1
Patent Ductus Arteriosus, Surgically Ligated, n (%)	9 (5.84%)	8 (7.41%)	1 (3.85%)	0 (0.00%)	0.656	1	0.356	1
Hematocrit Before NEC Onset (mean ± SD)	33.9 [30.2;40.1]	32.8 [29.4;38.1]	38.4 [33.6;49.3]	40.5 [34.8;44.9]	0.001	0.001	0.01	0.855
NEC Disease Features
Clinical Presentation, n (%)					0.1	0.728	0.033	0.314
Abdominal Distension	167 (90.8%)	121 (93.1%)	30 (90.9%)	16 (76.2%)
Bloody Stools	12 (6.52%)	6 (4.62%)	2 (6.06%)	4 (19.0%)
Feeding Intolerance	5 (2.72%)	3 (2.31%)	1 (3.03%)	1 (4.76%)
Radiological Findings, n (%)
Pneumatosis	57 (42.2%)	33 (34.0%)	11 (55.0%)	13 (72.2%)	0.005	0.131	0.005	0.446
Pneumoperitoneum	85 (49.1%)	60 (48.8%)	14 (48.3%)	11 (52.4%)	0.95	1	0.945	1
Portal Venous Gas	23 (13.3%)	10 (8.13%)	7 (24.1%)	6 (28.6%)	0.005	0.022	0.014	0.979
Age of NEC Onset (days; mean ± SD)	20.0 (20.5)	22.0 (21.5)	17.5 (17.1)	11.4 (16.4)	0.063	0.199	0.013	0.199
Feeding before the NEC onset
Feedings with Breast Milk, n (%)	84 (45.7%)	65 (50.0%)	13 (39.4%)	6 (28.6%)	0.137	0.371	0.112	0.603
Feedings with Donor Milk, n (%)	35 (19.0%)	29 (22.3%)	5 (15.2%)	1 (4.76%)	0.132	0.507	0.077	0.386
Feedings with Formula, n (%)	16 (8.70%)	6 (4.62%)	3 (9.09%)	7 (33.3%)	0.001	0.388	<0.001	0.035
Feedings with Both Formula and Breast Milk, n (%)	16 (8.70%)	11 (8.46%)	3 (9.09%)	2 (9.52%)	1	1	1	1
Post-Operative Systemic Course
Assisted Ventilation (intubated), n (%)	153 (91.6%)	113 (95.0%)	25 (86.2%)	15 (78.9%)	0.009	0.016	0.03	0.416
24h Pressor Support, n (%)	120 (70.6%)	90 (75.0%)	18 (60.0%)	12 (60.0%)	0.148	0.159	0.261	1
Cholestasis at NEC Onset, n (%)	80 (65.0%)	62 (68.9%)	9 (56.2%)	9 (52.9%)	0.329	0.483	0.319	1
Positive Blood Culture Sepsis, n (%)	47 (27.3%)	38 (30.9%)	5 (17.9%)	4 (19.0%)	0.249	0.251	0.399	1
Duration of Antibiotics (days; mean ± SD)	9.00 [6.00;12.0]	10.0 [7.00;12.8]	7.00 [3.00;10.0]	8.00 [4.25;10.0]	0.214	0.137	0.242	0.791
CRP on Day of NEC Onset (mean ± SD)	4.65 [1.40;8.38]	5.60 [1.70;10.8]	2.20 [0.75;4.00]	4.75 [2.28;6.32]	0.028	0.01	0.376	0.089
CRP at 24 Hours after NEC Onset (mean ± SD)	10.6 [3.02;19.8]	8.60 [2.90;19.7]	7.70 [4.70;19.1]	12.6 [4.40;19.7]	0.937	0.977	0.7	0.858
CRP at 48 Hours after NEC Onset (mean ± SD)	14.7 [3.75;21.6]	9.00 [3.30;21.2]	27.0 [8.80;33.8]	19.4 [3.20;21.1]	0.032	0.01	0.865	0.03
Discharge
Length of Stay (days; mean ± SD)	118 [72.8;178]	133 [87.5;180]	80.0 [26.0;159]	76.0 [53.0;117]	<0.001	0.005	0.001	0.653
Death, n (%)	61 (33.3%)	45 (34.9%)	10 (30.3%)	6 (28.6%)	0.783	0.772	0.751	1

Abbreviations: NEC = necrotizing enterocolitis; GA = gestational age; CRP = c-reactive protein, p1 = GA <28 weeks vs 28 to 32 weeks, p2 = GA <28 vs >32 weeks, p3 = GA 28 to 32 weeks vs >32 weeks, NA = Not available.

Continuous variables are presented as median with interquartile range and the differences were tested using Mann-Whitney U test for comparing two groups and Kruskal-Wallis test for more than two groups.

Categorical variables are presented as count (column percentage). Differences in categorical measures were tested using the Chi-square test when cell counts were adequate, otherwise Fisher’s Exact test with cell counts below 5. The presence of bold values signifies p < 0.05.

Group A vs. Group B &C:

As shown in Table 1, the infant’s in-group A had significantly lower evidence of portal venous gas (8.13% vs. Group B 24.1%; p=0.022) on an abdominal radiograph, later mean age of NEC onset (22 days ± 21.5 vs. Group C 11.4 ±16.4, p=0.013). Those with group A had significantly higher evidence of PDA (64.1% vs. Group B 34.4%; p=0.004) and received indomethacin (24.8% vs. Group B 6.6% vs. Group C 5%; p=0.017) more frequently compared to those with group B and C. Those with group A had significantly higher median CRP levels on the day of NEC onset (5.60 [IQR 1.70; 10.8] vs. 2.20 [IQR 0.75; 4.00]; p=0.01) and lower CRP levels at 48 hours B (9.00 [ IQR 3.3; 21.2] vs. 27.0 [ IQR 8.8; 33.8]; p= 0.01) than those with group B. Compared to group B, infants in group A had a longer length of hospitalization (133 days [IQR 87.5;180] vs. 80 days [IQR 26.0;159]; p=0.005).

The neonates with group A had significantly higher WBC count on the day and at 24 hours, 48 hours, and at day four following the NEC onset than in neonates with groups B and C. The group, A neonates, had significantly lower platelet count before and on the day, 24 hours 48 hours, and day seven after NEC onset. There was no significant statistical difference in platelet volumes in the three groups. Those with group A had significantly higher absolute neutrophil count and higher absolute monocyte count before and on the day, at 24 hours, 48 hours, and 72 hours compared to those with group B. Those with group A had significantly lower absolute lymphocyte counts at day three and day seven following NEC onset compared to those with gestational age between 28–32 weeks. In addition, the group A infants significantly had lower hematocrit (32.8 [29.4; 38.1] vs. 38.4 [33.6; 49.3]; p=0.001) compared to those with Group B before the NEC onset. The data is shown in Figure 1 and Supplemental Table 1 and Supplemental Figure 1.

Figure 1:

Hematological patterns (absolute counts) in preterm infants with surgical NEC in three different gestational age groups.

Group B and Group C comparison

As summarized in Table 1, infants in Group B had lower median birth weight (1140gm [IQR 898; 1420] vs 2043gm [IQR 1621; 2530]; p=<0.001) compared to infants in Group C. Those in Group B had lower rates of formula feeding prior to NEC onset (9.09% vs 33.3%; p=0.035) than infants in Group C. Group B infants had higher CRP levels 48 hours after NEC onset (27 [IQR 8.80; 33.8] vs 19.4 [IQR 3.20; 21.1]; p=0.03) when compared to Group C. Infants in Group B had higher levels of absolute lymphocytes count before NEC onset (3.43 [IQR 2.73; 4.92] vs 2.58 [ IQR 1.54; 3.3]; p=0.041) and 1 day after NEC onset (2.50 [IQR 1.53; 3.60] vs 1.25 [IQR 0.80; 2.08]; p=0.034) compared to Group C.

Medical vs. Surgical NEC

Group A:

In Group A, the infants with surgical NEC had significantly lower median lymphocyte counts and platelet count on the day of NEC onset and 1, 2, 3, and 4th day following NEC compared to neonates with medical NEC. The significant thrombocytopenia continued until the 7^th day following NEC in surgical NEC infants compared to medical NEC.

Group B:

In Group B, the infants with surgical NEC had significantly lower absolute neutrophil count (2.02 [IQR1.15; 4.08] vs. 4.70 [IQR 2.66; 6.70]; p=0.034), lower absolute lymphocyte count (3.25 [IQR 1.50; 4.13] vs. 4.37 [IQR 2.50; 6.26]; p=0.03), lower platelet counts on the day of NEC onset, lower absolute lymphocyte count at 24 hours (2.55 [IQR1.68;3.63] vs. 4.17 [IQR 2.38;5.90];p=0.006) and 3 days (1.70 [IQR 1.30;2.95] vs 5.00 [IQR 2.31;5.48];p=0.03) following the NEC onset than those with medical NEC. Those with surgical NEC also had significant thrombocytopenia until day 1, 2, 3 and day 4 following the NEC onset compared to infants with medical NEC.

Group C:

In Group C, the infants with surgical NEC had significantly lower absolute lymphocyte counts before, on the day of NEC onset and days 1,2,3 and 4 following NEC onset ( all P values <0.05). The absolute monocyte count was significantly lower on the day of NEC onset and one and the 7^th day following NEC onset in infants with surgical NEC compared to infants with medical NEC. The data is shown in Figure 2 and supplemental Table 2.

Figure 2:

Hematological patterns in preterm infants with surgical and medical NEC in three different gestational age groups.

Predictive modeling for surgical NEC vs. medical NEC Odds at different time points:

On multiple logistic regression analysis, Group A infants were significantly at higher odds of surgical NEC (aOR 3.39 [95% CI 1.19–10.38]; p=0.02) compared to Group C before the NEC onset. In addition, for every unit increase in absolute lymphocyte count on the day, and 24 hours following NEC onset was significantly associated with lower surgical NEC odds than medical NEC. The data has been summarized in Table 2.

Table2:

A multiple logistic regression to investigate the association between surgical NEC babies and potential predictors of surgery at different time points.

	Before NEC Onset			On the day of NEC			24 hours after NEC
Predictors	aOR	95% CI	p	aOR	95% CI	p	aOR	95% CI	p
WBC	1.21	0.99 – 1.67	0.146	1.07	0.96 – 1.20	0.221	0.90	0.80 – 1.00	0.057
Platelet Count	1.00	1.00 – 1.00	0.412	1.00	0.99 – 1.00	0.014	0.99	0.99 – 1.00	<0.001
ANC	0.82	0.59 – 1.03	0.174	0.93	0.80 – 1.07	0.320	1.17	1.02 – 1.37	0.035
ALC	0.74	0.52 – 0.97	0.053	0.58	0.45 – 0.72	<0.001	0.77	0.61 – 0.96	0.025
AMC	0.89	0.59 – 1.29	0.558	1.08	0.97 – 1.23	0.173	1.11	0.92 – 1.36	0.289
GA > 32 weeks	Ref.			Ref.			Ref.
GA 28–32 weeks	1.79	0.65 – 5.27	0.270	0.92	0.26 – 3.21	0.891	1.04	0.29 – 3.73	0.950
GA ≤ 28 weeks	3.39	1.19 – 10.38	0.026	1.32	0.37 – 4.62	0.666	0.98	0.28 – 3.39	0.978
Hematocrit Before NEC	1.00	0.96 – 1.05	0.981	1.02	0.97 – 1.08	0.463	0.98	0.93 – 1.04	0.499
PDA [No]	Ref.			Ref.			Ref.
PDA [Yes]	1.22	0.62 – 2.40	0.555	1.87	0.86 – 4.14	0.115	1.29	0.60 – 2.78	0.513

Abbreviations: aOR = adjusted odds ratio, NEC = necrotizing enterocolitis; GA = gestational age; WBC = white blood cell; ANC = absolute neutrophil count; AMC = absolute monocytes count; ALC = absolute lymphocytes count.

Model includes WBC, Platelet Count, ANC, ALC, AMC, GA, Hematocrit Before NEC and PDA

Survivor vs. non-survivors:

In-group A 57 infants (57/194, 29.3%), in group B 13 infants (13/68, 19.1%) and group C nine infants (9/49, 18.3%) had died. The non-survivors in-group A had significant thrombocytopenia following NEC onset until seven days and higher absolute monocyte counts (P=0.027) at day 4. The group B infants had significant thrombocytopenia (24 hours, 72 hours, and at day 7), leukopenia, lower absolute lymphocyte counts (at 72 and 96 hours), and significantly higher absolute monocyte counts before and at one week following NEC onset. The data has been summarized in Figure 3 and Supplemental Table 3.

Figure 3:

Hematological patterns in survivors and non-survivors in preterm infants with NEC in three different gestational age groups.

Transfusion Patterns:

Neonates in Group A received significantly more blood transfusions before NEC onset (59.8% vs. 30.0%; p=0.007) compared to Group B. Infants in Group A received an increased number of blood transfusions within the 48 hours following NEC diagnosis with those receiving two transfusions (25.2% vs. 4.76%) and three or more transfusions (20.6% vs. 9.52%; p=0.042) compared to Group C. The data is summarized in Table 3. The gestational age-specific red blood cell transfusions and platelet transfusions in survivors and non-survivors have been summarized in Supplemental Table 4. Infants in Group B who died received significantly more blood transfusion 48 hours following NEC onset ( 84.6% vs. 39.2 %, p=0.009).

Table 3 :

Transfusion Frequency in Neonates with NEC by Gestational Age

	Total Cohort	GA ≤ 28 Weeks	GA 28–32 Weeks	GA >32 Weeks	Overall p-value	p1	p2	p3
	n = 173	n = 121	n = 33	n = 21
Platelet Transfusion Before NEC Onset, n (%)	26 (15.3%)	21 (17.6%)	5 (15.2%)	0 (0.00%)	0.159	0.94	0.074	0.148
Number of Platelet Transfusions Before NEC Onset, n (%)					0.514	0.437	0.601	0.284
0	144 (84.7%)	98 (82.4%)	28 (84.8%)	18 (100%)
1	12 (7.06%)	8 (6.72%)	4 (12.1%)	0 (0.00%)
2	7 (4.12%)	7 (5.88%)	0 (0.00%)	0 (0.00%)
3+	7 (4.12%)	6 (5.04%)	1 (3.03%)	0 (0.00%)
Platelet Transfusion within 48 Hours of NEC Onset, n (%)	72 (41.6%)	49 (40.5%)	17 (51.5%)	6 (31.6%)	0.336	0.35	0.626	0.27
Number of Platelet Transfusions within 48 Hours of NEC Onset, n (%)					0.845	0.544	1	0.57
0	101 (58.4%)	72 (59.5%)	16 (48.5%)	13 (68.4%)
1	43 (24.9%)	28 (23.1%)	11 (33.3%)	4 (21.1%)
2	16 (9.25%)	11 (9.09%)	4 (12.1%)	1 (5.26%)
3+	13 (7.51%)	10 (8.26%)	2 (6.06%)	1 (5.26%)
Blood Transfusion Before NEC Onset, n (%)	81 (49.7%)	67 (59.8%)	9 (30.0%)	5 (23.8%)	0.001	0.007	0.005	0.866
Number of Blood Transfusions Before NEC Onset, n (%)					0.003	0.009	0.015	0.27
0	82 (50.3%)	45 (40.2%)	21 (70.0%)	16 (76.2%)
1	15 (9.20%)	11 (9.82%)	4 (13.3%)	0 (0.00%)
2	24 (14.7%)	19 (17.0%)	2 (6.67%)	3 (14.3%)
3+	42 (25.8%)	37 (33.0%)	3 (10.0%)	2 (9.52%)
Blood Transfusion within 48 Hours of NEC Onset, n (%)	121 (77.1%)	86 (80.4%)	22 (75.9%)	13 (61.9%)	0.18	0.784	0.086	0.453
Number of Blood Transfusions within 48 Hours of NEC Onset, n (%)					0.139	0.737	0.042	0.113
0	36 (22.9%)	21 (19.6%)	7 (24.1%)	8 (38.1%)
1	55 (35.0%)	37 (34.6%)	8 (27.6%)	10 (47.6%)
2	34 (21.7%)	27 (25.2%)	6 (20.7%)	1 (4.76%)
3+	32 (20.4%)	22 (20.6%)	8 (27.6%)	2 (9.52%)

Abbreviations: NEC = necrotizing enterocolitis; GA = gestational age

p1= GA <28 weeks vs GA 28–32 weeks; p2= GA <28 weeks vs GA >32 weeks; p3= GA 28–32 weeks vs GA >32 weeks

Categorical variables are presented as count (percentage). Differences in categorical measures were tested using a χ2 test. The presence of bold values signifies p < 0.05.

Discussion:

Our study demonstrates that preterm infants have different hematological profiles and responses based on their gestational age. Our data show that neonates in group A respond with a significantly higher WBC count, more extended lower platelet counts, higher absolute neutrophil count and higher absolute monocyte count, and lower absolute lymphocyte counts following NEC onset compared to those with gestational age between 28–32 weeks and higher. In addition, our data demonstrate that infants with surgical NEC in-group A had significant two cell lines (mainly platelets and lymphocyte counts); group B had three cell lines (lymphocytes, neutrophils, and platelets), and group C also had two cell lines involvement( lymphocytes and monocytes) compared to infants with medical NEC. Our study also shows a higher pro-inflammatory response as demonstrated by higher C reactive protein levels in younger preterm infants compared to infants with gestational age higher than 32 weeks. This observation indicates that the initial immunological response is more pro-inflammatory and immature anti –inflammatory responses in infants less than 32 weeks.

In this cohort, infant’s in-group B had higher absolute lymphocyte count levels before and after NEC onset than Group C. However, group A infants had lower absolute lymphocyte counts following NEC onset. This inverse response points towards in preterm infants with less than 28 weeks, the hematological response is mainly played by lymphocytes. The lower level of lymphocyte can be explained due to sequestration to the intestine as shown in animal models [30,31] and other organs such as the brain and kidney following NEC. However, infant’s in-group B had lower absolute monocyte and neutrophil counts than younger infants less than 28 weeks.

In our cohort, we noticed infants with gestational age less than 28 weeks had 10% higher mortality (A=29% vs. B/C=18–19%) than infants with higher gestation. In addition, those who died in-group A had significant thrombocytopenia and higher absolute monocyte counts. However, group B infants, in addition, had significantly lower total white cell counts and lymphopenia compared to infants with a gestation age of more than 32 weeks. These different hematological trends point towards continued developmental maturation, evolution, and response of the hematological system and intestine in preterm infants with necrotizing enterocolitis.

This study noticed significant and prolonged thrombocytopenia in infants with surgical NEC with gestational age less than 32 weeks (group A and B) compared to infants with gestational age more than 32 weeks (group C). The exact pathogenesis of low platelets in NEC is still unclear, although consumptive coagulopathy [32] and platelet aggregation secondary to activation by bacterial products [33] are few reported explanations. Animal data have reported the role of thrombin-mediated platelet activation [34,35] in NEC pathogenesis. Our findings point towards the gestation age-specific role of platelet in NEC etiopathogenesis, which needs further investigation and translational research. In this cohort, different gestational age group infants did not show any significant difference in mean platelet volume following NEC onset. However, few published reports have reported increased mean platelet volume in neonates with necrotizing enterocolitis [36,37].

In this study, infants in groups A and B (less than 32 weeks) received more platelet transfusions than group C (15% vs 0%). We did not observe any difference between platelet transfusions in three gestational age groups and mortality. A report by Kenton et al. did not find an association between platelet transfusion and mortality [23]. However, few reports recently have reported an association between platelet transfusions and mortality in preterm infants [22,24]; however, the outcomes in preterm infants by gestational age were not studied. We have previously reported infants who died received platelet transfusion significantly more before NEC onset. The neonates who died were more likely to have received a platelet transfusion before NEC onset (24.2% vs. 11.7%; p=0.03). However, platelet transfusions during the 48 hours after NEC onset did not discriminate between non-survivors vs. survivors (p=0.98). [29]. We did not observe those observations in this cohort due to splitting the mortality sample into three groups by gestational age. A further large study is needed to understand the association between gestation age, platelet transfusion, and clinical outcomes.

In our study, the infants with GA ≤ 28 weeks had significantly lower hematocrit and received more packed cell transfusion before and following the NEC onset compared to higher gestational age infants with NEC. However, on the multiple regression modeling the lower hematocrit level was not a significant risk factor in this cohort. However, non-survivors preterm infants with gestational age between 28–32 weeks received more packed cell transfusion within 48 hours after NEC onset compared to infants with less than 28 weeks and infants with more than 32 weeks gestational age. Wang et al. reported associations between red blood cell transfusions and clinical outcomes [38]. The number of RBC transfusions within seven days of birth was not different among those who went on to develop NEC (3.0±2.6) vs. those who did not develop NEC (2.8 ±1.7, p=0.521) [38]. Similarly, in our previous meta-analysis, we did not find an association between RBC transfusion and NEC onset [27].

In this cohort, the group A infants had significantly higher frequency of chorioamnionitis and lower Apgar scores compared to group B and C. Moore et al. [39] noted that infants with NEC (≥ Bells stage II) had an increased risk of placental infarcts and histological chorioamnionitis when compared to controls. Lower APGAR scores points towards hemodynamic compromise compared to other groups.

Our study’s strengths include our gestational age-specific evaluation of different hematological variables in preterm infants with NEC and provide insights about disease pathophysiology and guidance for further investigation and prognosis. However, we acknowledge study limitations, including the single-center retrospective design and predominantly African-American infants.

In conclusion, neonates with medical and surgical NEC have different gestational age-specific hematological patterns after the NEC onset. The neonates with gestation ages between 28–32 weeks were more likely to be associated with pancytopenia than group A and C infants following the NEC onset. Thrombocytopenia was more likely associated with NEC and death in preterm infants less than 32 weeks. The survivors and non-survivors had different hematological cell involvement depending on the gestational age. In addition, infants with gestational age between 28–32 weeks received a higher frequency of red blood cell transfusions following NEC onset. In the future, larger, prospective multi-center clinical and translational studies, which investigates gestation age-specific role of different hematological variables in the NEC etiopathogenesis, is much needed to improve the early diagnosis, prognosis, intervention, and clinical outcomes.