Serum chemokine IL-8 acts as a biomarker for identifying COVID-19-associated persistent severe acute kidney injury

Abstract

Objectives

Persistent severe acute kidney injury (PS-AKI) is associated with poor clinical outcomes. Our study attempted to evaluate the diagnostic value of chemokines for early-stage PS-AKI prediction.

Methods

According to the KDIGO criteria, 115 COVID-19 patients diagnosed with stage 2/3 AKI were recruited from the intensive care unit between December 2022 and February 2023. Primary clinical outcomes included detecting PS-AKI in the first week (≥ KDIGO stage 2 ≥ 72 h). Cytometric Bead Array was used to detect patient plasma levels (interleukin-8 (IL-8), C-C chemokine ligand 5 (CCL5), chemokine (C-X-C Motif) ligand 9 (CXCL9), and interferon-inducible protein 10 (IP-10)) of chemokines within 24 h of enrollment.

Results

Of the 115 COVID-19 patients with stage 2/3 AKI, 27 were diagnosed with PS-AKI. Among the four measured chemokines, only the IL-8 level was significantly elevated in the PS-AKI group than in the Non-PS-AKI group. IL-8 was more effective as a biomarker while predicting PS-AKI with an area under the curve of 0.769 (0.675-0.863). This was superior to other biomarkers related to AKI, including serum creatinine. Moreover, plasma IL-8 levels of >32.2 pg/ml on admission could predict PS-AKI risk (sensitivity = 92.6%, specificity = 51.1%). Additionally, the IL-8 level was associated with total protein and IL-6 levels.

Conclusion

Plasma IL-8 is a promising marker for the early identification of PS-AKI among COVID-19 patients. These findings should be validated in further studies with a larger sample size.

Keywords:

• covid-19
• acute kidney injury
• IL-8
• IP-10
• biomarker

Introduction

Early in the COVID-19 pandemic, an autopsy study of 26 patients who died of COVID-19 in China strongly suggested that SARS-CoV-2 may directly infect kidney tissue and produce diffuse proximal tubular injury [1]. A substantial body of evidence has since accumulated further suggesting that SARS-CoV-2 may directly infect the kidney, but the role of direct viral effects in COVID-19-associated acute kidney injury (AKI) remains controversial [2,3]. Hospitalized COVID-19 patients had a higher incidence of severe AKI (stage 2 or 3 AKI) than those without [4]. Worse outcomes and higher mortalities were associated with patients having persistent severe AKI (PS-AKI) (defined as AKI lasting ≥ three days) than those with no AKI or nonpersistent all-stage AKI [5–7]. COVID-19 patients with PS-AKI experience difficulties associated with kidney function recovery and often require renal replacement therapy (RRT) [4]. A study involving stage 2/3 AKI ICU patients indicated that the 1-year mortality rate of patients with late or no AKI reversal (within seven days) was higher than those with early sustained complete reversal [8]. Therefore, a novel biomarker can help distinguish patients experiencing difficulties related to kidney function recovery and PS-AKI progression.

AKI diagnosis is based on serum creatinine (sCr) levels and urine output. The sCr biomarker possesses a low sensitivity due to a 30% decline in the glomerular filtration rate for sCr to become detectable. A decrease in the urine output cannot specifically be used to detect AKI, as it could result from physiological factors and other diseases that do not cause direct damage to the kidney [9]. New markers with high levels of specificity and sensitivity are urgently needed to compensate for the disadvantages associated with traditional biomarkers. AKI results in the upregulation of chemokines (e.g. CXCL1) and adhesion molecules (e.g., P-selectin) in the endothelium of blood vessels in the kidney, which results in the infiltration of inflammatory cells from the blood vessels into the interstitium of the kidney. An injury induces the generation of inflammatory mediators, such as cytokines and chemokines, from tubular and endothelial cells. These mediators contribute to the recruitment of leukocytes into the kidneys [10]. High IL-8 levels correlated with the severity [11] and unfavorable COVID-19 outcome [12]. Furthermore, IL-8 is related to other AKI diseases, including liver transplantation [13] and ureteroscopic lithotripsy-related urosepsis [14]. Moreover, specific chemokines (e.g. CCL5) could effectively predict COVID-19 severity and correlate with patient mortality [15]. Additionally, CCL5 attenuates angiotensin II-dependent kidney injury by limiting renal macrophage infiltration [16]. Children with AKI after hematopoietic cell transplantation suffer from elevated urine IP-10 and CXCL9 [17], with urine IP-10 levels independently correlated with septic AKI risk [18].

Elevated inflammatory chemokine levels may be related to persistent AKI in COVID-19 patients. Therefore, we intended to determine the plasma chemokine level at enrollment in COVID-19 patients suffering from stage 2/3 AKI to evaluate its diagnostic potential while predicting persistent AKI.

Materials and methods

Participants

The present study was approved by the Ethics committee of Zhongda Hospital, Southeast University (Approval NO: 2020ZDSYLL287-P01), and written informed consent was obtained from all participants.

We recruited 115 patients diagnosed with COVID-19 pneumonia caused by SARS-CoV-2 Omicron variants from the intensive care unit between December 2022 and February 2023. SARS-CoV-2 pneumonia was defined by clinical data indicating respiratory distress, bilateral alveolar opacities in two or more lobes, a ratio of partial arterial oxygen pressure/inspired oxygen fraction (PaO₂/FiO₂) < 300 mm Hg, and a positive result for SARS-CoV-2-real-time reverse transcription–polymerase chain reaction assay using a pharyngeal swab. The severity of COVID-19 patients was classified into four clinical subtypes: mild, moderate, severe, and critical, according to the Guidelines for COVID-19 Diagnosis and Treatment. All the septic patients were diagnosed based on the third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) [19]. The sepsis group included patients suffering from a deadly organ disorder characterized by an increase in the Sequential Organ Failure Assessment (SOFA) score of ≥2 points post-infection. The septic shock patients were clinically identified using a vasopressor to maintain a mean arterial pressure of 65 mm Hg or greater and serum lactate level > 2 mmol/L (>18 mg/dL) without hypovolemia. Diagnosis and staging of AKI were based on the Kidney Disease Improving Global Outcomes (KDIGO) criteria using sCr levels and urine output [20]. Stage 2 AKI is characterized by increased serum creatinine to two-three times baseline; or a urine output < 0.5 mL/kg/h for over 12 h. Stage 3 AKI involves an increased serum creatinine > three times baseline or 4 mg/dL (>353.6 µmol/L); initiating renal replacement therapy; urine output < 0.3 mL/kg/h for over 24 h; or anuria for over 12 h. PS-AKI is depicted as stage 3 AKI lasting ≥ 3 days, with death in ≤ 3 days or any need for dialysis in ≤ 3 days. Non-PS-AKI is characterized as stage 3 AKI lasting for ≤ 2 days or stage 2 AKI for ≤ 3 days. The patient inclusion criteria were: (1) age between 18–90 years; (2) complete clinical records; (3) valid test results; and (4) no history of chronic kidney disease. Individuals with leukemia, tumors, mental disorders, or autoimmune diseases were excluded from the study.

Sample and data collection

We collected 2 mL of peripheral blood on enrollment using an ethylenediaminetetraacetic acid (EDTA) tube. Plasma was separated from peripheral blood after centrifugation and frozen at −80 °C for further use.

All the data, including demographic profiles and risk factors, were obtained using a pretested interviewer-administered questionnaire. The clinical data and basic laboratory findings (white blood count, electrolyte and biochemistry profile, and blood gas analysis) were collected from the information-management system at the hospital. The medical records provided the details regarding the AKI stage and complications in each patient.

Assessment of laboratory parameters and clinical symptoms

The plasma chemokine level was measured with a BD FACSCanto II cytometer and Cytometric Bead Array technology. The data were analyzed using the FACP software (BD bioscience) to determine chemokine concentrations. Chemokine selection was based on a literature review and available commercialized kits. The BD™ Cytometric Bead Array Human Chemokine Kit (BD bioscience) helped measure the levels of four chemokines, such as interleukin-8 (IL-8), C-C chemokine ligand 5 (CCL5), chemokine (C-X-C Motif) ligand 9 (CXCL9), and interferon-inducible protein 10 (IP-10).

Statistical analysis

Statistical analyses were performed using the SPSS 22.0 software (IBM, Armonk, NY, USA). Descriptive variables, such as categorical (n, %), normally distributed continuous (mean ± standard deviations), and nonparametrically distributed (median [Q1, Q3]) variables were measured. A comparison of COVID-19 patients with/without PS-AKI during hospitalization was performed with a chi-squared test for categorical variables and a parametric (unpaired t-test or ANOVA) or a nonparametric (Mann–Whitney U-test or Kruskal–Walis test) test for continuous variables. The correlation between chemokine-associated biomarkers and kidney function was determined using Spearman’s test. The logistic regression analyses helped identify the independent risk factors for PS-AKI through odds ratios and 95% confidence intervals (95% CIs) after adjusting sex, age, and septic shock. The receiver operating characteristic (ROC) curve helped analyze chemokine efficacy in predicting PS-AKI occurrence. The cutoffs for each biomarker were selected according to the highest area under the curve (AUC), specificity, and sensitivity values to predict AKI. The values were considered to be statistically significant if p < 0.05.

Results

Participant characteristics

Of the 156 COVID-19 patients with severe AKI recruited in our study, 115 (73.7%) were analyzed to determine if they had persistent AKI. Forty-one patients were excluded; 20 had insufficient medical records; 18 suffered from tumors, active immune diseases, and leukemia; and three possessed invalid test results. Twenty-seven COVID-19 patients with stage 2/3 AKI developed PS-AKI based on KDIGO criteria (Figure 1). Of the 27 PS-AKI patients, 7 (25.9%) received RRT, 5 (18.5%) were anuric, and 2 (7.4%) died. In the PS-AKI group, the median age was 73.63 years, and 48.1% were male. The baseline biochemistry, WBC, D-dimer, and blood gas analyses during enrollment did not differ significantly between patients with and without PS-AKI (p > 0.05). We observed significant differences in C reactive protein (CRP) and interleukin-6 (IL-6) between the two groups. No significant differences were identified in clinical symptoms (PaO₂, SpO₂, P/F Ratio, patient distribution in different COVID-19 severity groups, and its diagnosis onset) except for SOFA scores. At enrollment, PS-AKI patients were more likely to possess stage 3 AKI, high sCr levels, and septic shock (Table 1). The median time between sample sampling and persistent AKI diagnosis was two days.

Figure 1. Selection of patients included in this study.

Table 1. Baseline characteristics of the COVID-19 patients.

Variables	Not persistent severe AKI (N = 88)	Persistent severe AKI (N = 27)	P-value
Male (%)	48 (54.5%)	13 (48.1%)	0.562
Age (years)	73.02 ± 15.35	73.63 ± 12.57	0.852
Septic shock (%)	30 (34.1%)	19 (70.4%)	<0.001
Laboratory parameters
WBC (10^9/L) ^#	8.97(6.64,11.92)	10.48(6.04,13.24)	0.682
CRP (mg/L) ^#	102.5(48.4,194.2)	167.5(108.7,250.0)	0.002
IL-6 (pg/mL) ^#	62.7(22.9,261.5)	215.6(94.2,894.6)	0.001
Urea (mmol/L) ^#	13.7 (8.4,23.0)	15.8 (9.7,34.5)	0.088
Scr (μmoI/L) ^#	123.5 (95.5, 200.3)	196.0 (122.0,288.0)	0.008
Total protein (g/L) ^#	56.9 (51.2,61.3)	58.4 (50.8,60.6)	0.963
Albumin (g/L) ^#	30.6 (27.7,33.2)	32.1 (29.2,36.6)	0.164
D-Dimer (μg/L) ^#	2199.5 (1293.0,4596.0)	2193.0 (1568.0,4217.0)	0.707
Clinical symptom
PaO2, mmHg^#	78.9 (53.1,93.9)	71.9 (41.5, 82.0)	0.096
SpO2, %^#	95.8 (87.8, 97.8)	91.3 (80.0,97.1)	0.080
P/F Ratio^#	211.6 (149.3,290.6)	256.8 (190.5, 319.2)	0.139
SOFA score^#	6.5(5.0,10.0)	10.0(6.0,13.0)	0.007
The onset of COVID-19, days^#	3.5(2.0,7.0)	6.0(2.0,8.0)	0.348
Enrollment COVID-19 severity			0.988
Mild	25 (28.4%)	7 (25.9%)
Moderate	6 (6.8%)	2 (7.4%)
Severe	11 (12.5%)	3 (11.1%)
Critical	46 (52.3%)	15 (55.5%)
Enrollment KDIGO stage			<0.001
Stage 2	58 (65.1%)	6 (22.2 %)
Stage 3	30 (34.9%)	21 (77.8%)
The duration of disease, days^#		2.0 (1.0,3.0)

*Bold values indicate significant difference.

^#nonparametric distributions variables were expressed as median [Q1, Q3].

Abbreviations: AKI: acute kidney injury; Scr: serum creatinine; PaO₂: partial pressure of oxygen; SpO₂: arterial blood oxygen saturation; P/F: the ratio of pressure of oxygen and the fraction of inspired oxygen; WBC: white blood count; CRP: C reactive protein; IL-6: interleukin-6; SOFA: Sequential Organ Failure Assessment; KDIGO: The Kidney Disease: Improving Global Outcomes.

Plasma chemokine levels in COVID patients with severe AKI

We identified a reduction in CCL5 and an increase in CXCL9 and IP-10 levels in the plasma of PS-AKI patients without significance. The exception was plasma IL-8, with significantly higher levels than that in the Non-PS-AKI group (Figure 2). We conducted a subgroup analysis of septic shock patients since IL-8 is correlated with infection occurrence, and PS-AKI is primarily seen in such patients. Interestingly, the CCL5 level was significantly lower in the group of Non-PS-AKI patients with septic shock than those without septic shock (Figure 3). Non-PS-AKI patients in the group with septic shock had significantly lower CCL5 and significantly higher IL-8 levels than those without septic shock. However, the median IL-8 level was higher in PS-AKI patients with septic shock than those without septic shock, but the difference was insignificant (Figure 3). IL-8 can identify sepsis patients with higher mortality risk [21], which can be predicted using a two-biomarker model [22]. Due to the limited subgroup analysis power, the different IL-8 levels were not statistically significant among PS-AKI patients with and without septic shock.

Figure 2. Chemokine levels in the non-PS-AKI and PS-AKI groups. IL-8: interleukin-8; CCL5: C-C chemokine ligand 5; CXCL9: chemokine (C-X-C Motif) ligand 9; IP-10: interferon-inducible protein 10; PS-AKI: persistent severe AKI. *p < 0.05.

Figure 3. The impact of septic shock on chemokine levels in the non-PS-AKI and PS-AKI groups. IL-8: interleukin-8; CCL5: C-C chemokine ligand 5; CXCL9: chemokine (C-X-C Motif) ligand 9; IP-10: interferon-inducible protein 10; PS-AKI: persistent severe AKI. *p < 0.05. **p < 0.01.

Next, we investigated IL-8 correlation with other laboratory parameters in PS-AKI patients. Baseline IL-8 levels at admission were positively associated with IL-6 and total protein levels (p < 0.05, Figure 4). There were no significant associations between IL-8 and other indicators (including urea and albumin).

Figure 4. Correlation between plasma IL-8 levels and other clinical parameters. sCr: serum creatinine; IL-6: interleukin-6; IL-8: interleukin-8.

Plasma IL-8 levels in PS-AKI prediction

Elevated IL-8 levels could be seen in COVID-19 patients with PS-AKI, correlating with IL-6 levels. ROC analysis helped examine the values of IL-6, IL-8, and sCr while predicting PS-AKI (Table 2). The AUC of IL-8 was 0.769 (0.675–0.863), superior to the AUC of IL-6 (0.714) or sCr (0.669) (Figure 5). The sensitivity, specificity, and Youden indices were 92.6%, 51.1%, and 0.517, respectively, at an optimal cutoff value of 32.2 pg/mL (Table 2).

Figure 5. Receiver operating characteristic curve of clinical indicators for the diagnostic prediction of PS-AKI. IL-8: interleukin-8; sCr: serum creatinine; PS-AKI: persistent severe AKI.

Table 2. The diagnostic efficacy of various indicators for PS-AKI.

Indicators	AUC	95%CI	P value	Optimal threshold	Sensitivity(%)	Specificity(%)	Youden Index
SCr(μmoI/L)	0.669	0.546–0.791	0.008	140	70.4	61.4	0.317
IL-8(pg/mL)	0.769	0.675–0.863	<0.001	32.2	92.6	51.1	0.517
IL-6(pg/mL)	0.714	0.609–0.819	0.001	83.365	85.2	56.8	0.420

Abbreviations: PS-AKI: ersistence severe acute kidney injury; sCr: serum creatinine; IL-8: interleukin-8; IL-6: interleukin-6.

The reference range for IL-8 in clinical applications is 0-20.6 pg/mL. Depending on the optimal cutoff values, we categorized serum IL-8 expression levels as low (≤ 20.6 pg/mL), medium (> 20.6 to ≤ 32.2 pg/mL), or high (> 32.2 pg/mL). Our results indicated that an elevated IL-8 expression level enhanced PS-AKI risk compared to a low IL-8 expression level (Table 3).

Table 3. The values of IL-8 for PS-AKI by multivariate logistic regression analysis.

Variables	Non-PS-AKI	PS-AKI	Crude		Adjusted
	(N = 88)	(N = 27)	OR (95%CI)	P	OR (95%CI)	P
IL-8 < 20.6 pg/mL	36(40.9%)	1(3.7%)	1.0	–	1.0	–
20.6 ≤ IL-8 ≤ 32.2 pg/mL	16(18.2%)	1(3.7%)	2.25(0.13,38.27)	0.575	2.51(0.15,43.24)	0.526
IL-8 > 32.2 pg/mL	36(40.9%)	25(92.6%)	25.0(3.21,194.5)	0.002	26.7(3.4,210.7)	0.002

*Bold values indicate significant difference.

^#Adjusted for age, sex and septic shock.

Abbreviations: PS-AKI: ersistence severe acute kidney injury; IL-8: interleukin-8.

Discussion

Our study used chemokine levels during enrollment to predict persistent severe AKI. Only the IL-8 level was significantly increased among the four chemokines, becoming a more effective biomarker than creatinine and IL-6 for predicting PS-AKI occurrence.

The explanation that IL-8 predicted the highest AUC for PS-AKI among the four cytokines is as follows. First, low CCL5 mRNA levels were related to intensive care unit admission or death during the COVID-19 disease [15]. Our study had significantly lower serum CCL5 levels in non-PS-AKI patients with septic shock than those without. Higher CCL5 serum levels were related to better kidney recovery in dialysis-dependent AKI patients [23]. Thus, CCL5 expression could be low in COVID-19-related AKI patients. The serum CCL5 levels were 2529.3(947.1,5446.8) and 2790.7(691.0,3638.9) in non-PS-AKI and PS-AKI patients, respectively (supplementary data). Due to the small sample size, no significant difference could be observed between non-PS-AKI and PS-AKI patients in serum CCL5 levels. Second, Singh et al. observed that urinary CXCL9 correlated best to tissue CXCL9 expression (r = 0.75, p < 0.001) while distinguishing interstitial nephritis from non-interstitial nephritis (AUC 0.781) [24]. Urine CXCL9 and IP-10 became potential AK1 biomarker [18]. However, unlike previous studies, we collected plasma specimen [18].

During a SARS-CoV-2 infection, an increased secretion or production of IL-6 and IL-8 is observed in COVID-19 patients, along with a reduction in CD4+ and CD8+ T cells [25]. Serum IL-6 and IL-8 levels were the highest within the severe COVID-19 patient group. Moreover, the mild-moderate group also had higher levels than healthy volunteers (p = 0.001) [11]. Postoperative IL-8 was significantly elevated in cardiac surgery-AKI patients than in non-cardiac surgery-AKI [26]. Fontnouvelle et al. observed that preoperative IL-8 was significantly related to higher AKI odds after cardiac surgery in children two years and older [27]. Moreover, IL-8 levels were significantly increased in stage 2/3 AKI patients than in stage 1 AKI [27]. The highest tertile of preoperative IL-8 showed higher adjusted odds of severe AKI (5.06 [95% CI: 1.36–18.74]) development than the first tertile [27]. Additionally, after cardiopulmonary bypass, IL-6 levels at 2 and 12 h and IL-8 at 2, 12, and 24 h correlated with AKI development [28].

Despite the role of elevated IL-8 levels in other disease-associated AKIs, little is known about their association with COVID-19-associated AKI. Critical SARS-CoV-2-associated AKI patients depict an increase in inflammatory cytokine levels (IL-1β, IL-8, IFN-γ, TNF-α), favoring renal dysfunction through intrarenal inflammation, elevated vascular permeability, volume depletion, microvasculature thromboembolic events, and persistent local inflammation [29]. Therefore, multiple chemokines were selected within 24 h of recruitment, such as IL-8, to predict persistent AKI. Our results indicate that IL-8 has moderate diagnostic value for PS-AKI prediction and outperforms traditional markers like initial sCr. A retrospective observational study involving 119 COVID-19 patients from China showed that high serum IL-8 was a risk factor for developing severe AKI [30], consistent with our results.

The possible explanations for persistent AKI development through IL-8 upregulation of IL-8 are: (1) AKI causes tubular injury and interstitial inflammation. CXCL8 induces a local innate immune response after NF-κB pathway activation. Sustained or overexpression of IL-8 can perpetuate immune cell recruitment with deleterious inflamed tissue effects [31]. In contrast, in a rabbit model of lung reperfusion injury, administering neutralizing anti–IL-8 mAb limited polymorphonuclear cell infiltration and tissue injury [32]. Therefore, IL-8 overexpression consistently regulates immune cell infiltration of inflamed tissues causing tissue damage; inhibiting IL-8 reverses the deleterious effect. (2) Neutrophil dysregulation and excessive neutrophil extracellular trap (NETs) formation under the local influence of cytokines could mediate tissue damage. The NET formation was related to local IL-8 mRNA levels (r = 0.708, p = 0.002) and CD8+ T-cell infiltration density [33]. Moreover, higher NETs were significant determinants of COVID-19 AKI [34]. (3) Kwon et al. described that enhanced urinary IP-10 and IL-8 could predict no renal functional recovery the following day [35], consistent with our findings.

Some limitations need to be noted to avoid result misinterpretation. First, the possibility of false-positive results could not be ruled out because it was a single center with a limited sample size. Second, the cause and type of AKI were not considered. IL-8 levels could differ in individuals with different AKI types or causes. Third, we did not include therapeutic variations strongly affecting the AKI time course. Finally, we did not detect dynamic changes in IL-8 levels at 24 h, 48 h, and 72 h or promptly monitor AKI occurrence. Despite these limitations, the present study is the first to examine the role of IL-8 in the prediction of PS-AKI. Additional larger multicenter studies are needed to validate the role of IL-8 in predicting PS-AKI in patients with and without COVID-19.

Conclusions

Our findings suggest that an elevated plasma IL-8 level facilitates predicting persistent AKI among critically ill patients. Consistent with previous studies, chemokine signaling and neutrophil migration can play an important role in renal injury and recovery.

Ethical approval

This study was approved by the Ethics committee of Zhongda Hospital, Southeast University (Approval NO: 2020ZDSYLL287-P01), and written informed consent was obtained from all participants.

Disclosure statement

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

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.