Acute kidney injury in ventilated patients with coronavirus disease-2019 pneumonia: A single-center retrospective study.

BACKGROUND: Acute kidney injury (AKI) is repeatedly observed in ventilated critically ill patients with coronavirus disease-2019 (COVID-19) pneumonia. This study aimed to determine the incidence, risk factors, and consequences of AKI in the ventilated critically ill adult patients with COVID-19 pneumonia.
METHODS: This retrospective study included all the ventilated critically ill adult patients with COVID-19 pneumonia from March 1, 2020, to June 1, 2020. Data were collected from the electronic medical system. AKI was diagnosed using the Kidney Disease: Improving Global Outcomes 2012 Clinical Practice definition. Patients were followed 90 days from the intensive care unit (ICU) admission time or to the date when they were discharged from the hospital.
RESULTS: AKI occurred in 65.1% of patients, with 26.6% of these started on continuous renal replacement therapy (CRRT). Patients with AKI had higher comorbidity and illness severity scores (P < 0.001). Age and the vasopressor requirements were predictors of AKI (P= 0.016 and P = 0.041) and hypertension predicted AKI (P = 0.099) and its progression (P = 0.05). The renal recovery rate was 86.7% and was associated with the mean arterial pressure on ICU admission in the no-CRRT group (P = 0.014) and the hypoxic index in the CRRT group (P = 0.019). AKI was associated with higher mortality (P = 0.017) and significantly longer ICU length-of-stay (P = 0.001). Additionally, AKI patients were more often discharged to a long-term skilled nursing facility (P = 0.005).
CONCLUSION: COVID-19-associated AKI was common and associated with poor outcome, with the specific mechanisms being the main driving factors. Copyright:

INTRODUCTION

Acute kidney injury (AKI) is repeatedly observed in the ventilated adult patients with coronavirus disease-2019 (COVID-19) pneumonia. Moreover, AKI may have detrimental effects on the outcome of these patients, and it can increase the burden on health facilities.[1234]However, there are significant variations in the reported incidence rates of AKI and the severity of the renal disease in severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection among different cohorts.[5]Several proposed mechanisms may explain how AKI may develop in these patients, such as the genetic factors and host immune responses.[67891011]It is also postulated that management in intensive care plays a significant role in the development and progression of AKI and could be the basis to prevent or limit AKI in these vulnerable patients.[1112]

This study aimed to (1) evaluate the incidence of AKI in the ventilated critically ill patients with COVID-19 pneumonia; (2) identify the risk factors for developing AKI in these patients; and (3) define the clinical impact of acute kidney.

METHODS

In this single-center study, we collected data retrospectively from the hospital Electronic Medical Record System, Epic Hyperspace® version 2019 (Epic Systems Corp., Headquarters City, Verona, USA). Patients were followed up for 90 days from the hospital admission or to discharge time from the hospital.

We included all the ventilated adults, aged above 18 years, admitted to the Rashid hospital intensive care department from March 1, 2020, to June 1, 2020, with COVID-19 pneumonia.

COVID-19 pneumonia was diagnosed according to the National Guidelines for Clinical Management and Treatment of COVID-19.[13]Specifically, the diagnostic criteria for COVID-19 pneumonia were as follows: (1) fever or respiratory symptoms, (2) chest X-ray or CT showing radiographic abnormalities in the lung, and (3) a SARS-CoV-2 polymerase chain reaction test by an approved laboratory (mandatory). Sample analyses were performed with a real-time RT-PCR test using the Xpert® Omni SARS-CoV-2 kit.

Exclusion criteria were - (1) patients aged younger than 18 years; (2) patients whose entire stay lasted ≤ 48 h; and (3) patients with Stage 5 according to the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) definition and classification of Chronic Kidney Disease (CKD).[14]

The primary outcome was to evaluate AKI incidence in the ventilated critically ill patients with COVID-19 pneumonia. The secondary outcomes were to determine the risk factors related to the clinical course of AKI (development, progression, and recovery), the impact of AKI on mortality, and intensive care unit (ICU) length of stay (LOS).

AKI was defined according to the KDIGO 2012 clinical practice definition.[15]The baseline creatinine level was considered as the latest level in the last 90 days before the development of COVID19 pneumonia. If that was unavailable, the lowest value of creatinine from the time of presentation to the emergency department to the first creatinine level on ICU admission was considered the baseline level. Serum creatinine was measured at least once daily from admission and during the stay in the ICU based on our unit protocols and hospital protocol.

Patients who met the criteria for AKI were classified as “AKI group,” whereas those who did not were classified as “no AKI group.” Subsequently, patients in the AKI were stratified into two subgroups according to the commencement of continuous renal replacement therapy (CRRT): CRRT group and no-CRRT group.

CRRT was the only extracorporeal kidney support in the ICU, and the decision to start the dialysis was based on the modern criteria of initiation of CRRT.[16]

To quantify the severity of comorbidity, we used the Charlson Comorbidity Index (CCI). Furthermore, we used three ICU scoring systems: Sequential Organ Failure score in different groups (SOFA score), Acute Physiology and Chronic Health Evaluation (APACHE) II, and Simplified Acute Physiology Score I. Likewise, we used the 2012 Berlin Definition of ARDS to quantify the severity of pulmonary involvement.[17]

Renal recovery was assessed by monitoring the serum creatinine and RRT requirement in the AKI group for 90 days. Patients with AKI were categorized into three grades based on the consensus report of the Acute Disease Quality Initiative 16 Workgroup.[18]These groups specifically are (1) complete renal recovery: when the serum creatinine returned to its baseline level or a lower value within 7 days from the onset of AKI; (2) partial renal recovery: when the serum creatinine level remained above the baseline for after 7 days from the onset of the AKI but no ongoing RRT; and (3) no-renal recovery: when there was an ongoing need for renal replacement therapy or the serum creatinine remained ≥4.0 mg/dl from its baseline.

Statistical analysis

Descriptive statistics were reported as medians (interquartile range: [IQR]) or means with standard error of mean for continuous variables according to the normality of distribution for continuous variables and number (%) for categorical variables. The normal distribution of data was analyzed using the Shapiro–Wilk test or Kolmogorov–Smirnov test.

A preliminary univariate analysis was performed to examine the association of the variables considered clinically relevant based on professional knowledge with incidence, progression, and recovery of AKI. We analyzed the relationship between the medications used to treat COVID-19 pneumonia individually (factorial) in the univariate analysis. To confirm the association between the drugs and the AKI, we created models representing the different combinations of antiviral and immunomodulators used to treat COVID-19 pneumonia in ICU patients. Specifically, variables with P < 0.05 in the univariate analysis were entered into a logistic regression analysis with a backward stepwise model to identify predictors of the development of AKI and its progression and recovery. The association between predictors and the dependent variable was quantified using the odds ratio (OR) with a 95% confidence interval (CI).

Receiver operating characteristic (ROC) curve analyses were applied to check the optimal cutoff values of the different factors associated with AKI development and evaluate the predictive power between them further, considering the differences between the areas under the empirical ROC curves.

Outcome analysis was performed using the Kaplan–Meier estimator for different stages on ICU LOS, and the proportional hazards Cox regression model was used to evaluate the effect of AKI and class groups on survival. Two-tailed P < 0.05 was considered statistically significant.

Statistical analyses were performed using IBM SPSS version 25.0 (IBM Corp., Armonk, NY, USA).

RESULTS

From a total of 224 ICU patients who tested positive for SARS-CoV-2, we found that 214 patients met the diagnosis of COVID-19 pneumonia. After applying the exclusion criteria, 198 ventilated patients were included in the study, as shown in Figure 1.

Figure 1

Consort flow diagram

AKI occurred in 65.1% of patients (129/198), and it started within 2 days from the ICU (77.5%, 100), and there were 53 patients (41.1%) who remained in Stage 1, 24 (18.6%) progressed to Stage 2, and 52 (40.3) developed Stage 3 AKI. RRT was commenced in 34 patients with AKI (26.4%), 33 of them were in Stage 3 when CRRT was commenced, and there was only one patient in AKI Stage 2 when the CRRT was started.

Factors associated with acute kidney injury development and progression

The demographic data, patients' clinical characteristics, laboratory findings, and treatment are shown in Tables 1-4. Patients with AKI were older in the AKI group (53.1 ± 2.75 vs. 46.8 ± 1.2; P = 0.001) and they had a higher rate of hypertension (33% vs. 14.5%; P = 0.007) and they had a significantly higher CCI (1.8 ± 0.14 vs. 1.0 ± 17; P < 0.001). They also had a significantly higher total sequential organ failure assessment (SOFA) (5.06 ± 0.103 vs. 4.492 ± 0.158; P = 0.002) and cardiovascular SOFA (1.77 ± 0.052 vs. 0.884 ± 0.06 P = 0.024). Similarly, the SAPS II and APACHE II scores were higher in the AKI (27 ± 0.54 vs. 24 ± 0.57; P < 0.001 and 11.89±0.217 vs. 10.2 ± 0.249; P < 0.001, respectively). Furthermore, our analysis showed higher maximal norepinephrine doses in patients with AKI (0.048 μg/kg/min [IQR: 0.01–0.1] vs. (0.03 μg/kg/min [IQR: 0.01–0.04]; P = 0.037). Patients in the no-AKI group had a higher mean arterial pressure (MAP) compared those in AKI group (79 mmHg [IQR: 73–88]; P = 0.042, versus 75 mmHg [IQR 70–83]). Similarly, laboratory findings were similar in both the groups such as the blood urea nitrogen on ICU admission (16 [IQR: 11–21] mg/dl vs. 18 mg/dl [IQR: 13–26] mg/dl, P = 0.101) and the baseline serum creatinine (0.9 [IQR: 0.7–1.23] mg/dl vs. 0.8 [IQR: 0.7–1.00] mg/dl; P = 0.281). In addition, there were no significant differences in the rate of proteinuria (71% [n = 49] vs. 77.5% [n = 100]; P = 0.52). Mild proteinuria (+1 by dipstick) was found in 121 patients, while heavy proteinuria (≥ +2 by dipstick) was present in 13 patients. Hematuria was present in 67.7% of the patients (134/198). Leukocyturia was present in 5.1% (10) of patients. Glycosuria was found in 18 patients (9.2%) and there were no significant differences in the rate of these findings in both the groups.

Table 1

Patients’ demographic and premorbid clinical characteristics stratified according to the clinical course of coronavirus disease-2019-associated acute kidney injury

Variables	All patients (n=198; 100%)	AKI (n=129; 65%)	No-AKI (n=69; 35%)	P	No-CRRT (n=95; 74.6%)	CRRT (n=34; 26.4%)	P	No-RRT (n=80)		P	RRT (n=33)		P
								Complete (n=42; 52.5%)	Partial (n=38; 47.5%)		Partial (n=18; 54.5%)	No recovery (n=15; 45.5%)
Age years±(SEM)	50.9±0.08	53.1±2.75	46.8±1.2	0.001	53.1±2.75	52.8±1.5	0.446	52.5±1.55	51.8±1.95	0.782	52.5±2.98	56.5±2.73	0.376
Sex (male)	188 (94.4)	122 (94.6)	65 (94.2)	0.422	92 (96.8)	30 (88.2)	0.136	41 (97.6)	36 (94.7)	0.435	16 (88.9)	13 (86.7)	0.882
BMI	28.05±1.78	28.3±1.78	27.8±1.74	0.876	28.3±1.1	28.1±0.7	0.762	28.3±0.8	28.2±0.6	0.895	28.1±0.01	28.0±0.04	0.962
Medical history
Chronic kidney disease**	4 (2)	3 (2)	1 (1)	0.774	3 (3.2)	0	0.236	1 (2.4)	1 (2.6)	0.292	0	0	N/A
Hypertension	53 (26.8)	43 (33)	10 (14.5)	0.007	27 (28.4)	16 (47.1)	0.048	11 (26.2)	16 (42.1)	0.107	9 (50.00)	6 (40.00%)	0.79
Diabetes	70 (35.4)	48 (37.2)	22 (31.9)	0.769	32 (33.7)	16 (47.1)	0.702	12 (26.6)	12 (31.6)	0.712	9 (50)	8 (53.3)	0.765
Cardiovascular*** diseases	15 (7.6)	12 (9)	3 (4)	0.863	8 (8.4)	2 (5.9)		4 (10.5)	2 (4.8)	0.514	2 (11.1)	0	0.633
Other medical illness†	11	7	4	N/A	5	2	N/A	2	1	N/A	0	0	N/A
Charlson index‡	1.5±0.11	1.8±0.14	1.0±0.17	0.001	1.72±0.17	2.15±0.25	0.201	2.14±0.22	2.42±0.14	0.304	2.11±0.37	2.20±0.38	0.869

**CKD; excluding stage 5 CKD, renal transplant, ***Cardiovascular diseases (n=15; past ischemic heart disease IHD [n=7]; acute myocardial infarction 2, old PVD n=2, old CVS [n=4]), †Other medical illness included (COPD=6; (COPD [n=1], Asthma [n=3], pulmonary fibrosis [n=1]), Cirrhosis [n=1], peptic ulcer disease [n=1], rheumatoid arthritis [n=1], chronic anemia [n=1], lymphoma [n=1]), ‡CCI, All variables are presented as, categorical variables: n (%), Continuous variable with normal of distribution: Means with Standard Error of Mean ([Mean±[SEM]), Continuous variable with skewed distribution: Median and IQR, P value, from univariate analysis. bold if significance (P<0.05). N/A: Not applicable, PVD: Peripheral vascular disease, CCI: Charlson Comorbidity Index, IQR: Interquartile range, SEM: Standard error of mean, CKD: Chronic kidney disease, COPD: Chronic pulmonary disease, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, RRT: Renal replacement therapy, BMI: Body mass index

Table 4

Treatment stratified according to the clinical course of coronavirus disease-2019-associated acute kidney injury

Variables	All patients (n=198; 100%)	AKI (n=129; 65%)	No-AKI (n=69; 35%)	P	No-CRRT (n=95; 74.6%)	CRRT (n=34; 26.4%)	P	No-RRT (n=80)		P	RRT (n=33)		P
								Complete (n=42; 52.5%)	Partial (n=38; 47.5%)		Partial (n=18; 54.5%)	No recovery (n=15; 45.5%)
Antiviral therapy
Lopinavir-ritonavir	183 (92.4)	121 (94)	62 (90)	0.201	88 (92.6)	33 (97.1)	0.305	34 (89.5)	40 (95.2)	0.58	18 (100)	14 (93.3)	0.64
Chloroquine	147 (74.2)	97 (75)	50 (72)	0.893	72 (75.8)	25 (73.5)	0.753	31 (73.8)	28 (73.7)	0.47	14 (77.8)	11 (77.3)	0.09
Hydroxychloroquine	40 (20.2)	25 (19)	15 (22)	0.782	19 (20.2)	6 (17.6)	0.49	8 (19.0)	8 (21.1)	0.43	4 (22.2)	2 (13.3)	0.122
Favipiravir	128 (64.6)	81 (69)	37 (54)	0.576	31 (58.5)	17 (70.8)	0.566	26 (61.9)	38 (57.9)	0.38	12 (66.7)	10 (66.70)	0.66
Immunomodulating
Tocilizumab	53 (26.8)	34 (26.5)	19 (27.5)	0.851	40 (42.1)	11 (32.4)	0.238	15 (35.7)	19 (50)	0.055	6 (38.9)	4 (26.7)	0.86
Peginterferon α 2b	51 (25.8)	27 (21)	24 (35)	0.065	16 (16.8)	11 (32.4)	0.101	4 (9.5)	8 (21.1)	0.11	7 (38.9)	4 (26.7)	0.53
Corticosteroids	119 (60.1)	79 (61.2)	37 (54)	0.302	61 (64.2)	23 (52.9)	0.526	23 (54.7)	26 (68.4)	0.324	6 (38.9)	6 (40)	0.42
Other supportive therapy
Diuretics	63 (31.8)	51 (41)	23 (33)	0.247	40 (42.1)	11 (32.4)	0.404	23 (62.9)	20 (73.7)	0.408	8 (44.4)	4 (26.7)	0.66
ARBs/ACEi	4 (2.0)	3 (2)	1 (1)	0.094	1 (1.9)	1 (4.2)	0.568	1 (2.4)	1 (2.6)	0.52	1 (5.6)	0	N/A
NSAIDS	2 (1.0)	2	(1)	0.208	1 (1.9)	0	0.793	2 (4.8)	0	N/A	0	0	N/A
Contrast	9 (4.5)	7 (5)	2 (3)	0.2	4 (4.2)	3 (8.8)	0.286	1 (2.4)	0	N/A	2 (11.1)	1 (6.7)	0.428

All variables are presented as Categorical variables: n (%), Continuous variable with normal of distribution: Means with (SEM±SEM), Continuous variable with skewed distribution: Median and IQR, P value, from univariate analysis. bold if significance (P<0.05). N/A: Not applicable, ARBs: Angiotensin II receptor blockers, ACEi: Angiotensin-converting enzyme inhibitors, NSAIDs: Nonsteroidal anti-inflammatory drugs, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, RRT: Renal replacement therapy, IQR: Interquartile range, SEM: Standard error of mean

The multivariate analysis of factors associated with AKI revealed that the patient age and the maximal requirements of norepinephrine were the main predictors of AKI. The predictive power of the model was 71% at cutoff 51.5 years for patient's age and 0.032 mcg per body weight in kg per minute norepinephrine dose [Table 5- Model 1].

Table 5

Predictors of the clinical course of the coronavirus disease-2019-associated acute kidney injury

Variable	Multivariate logistic regression		ROC analysis		Sensitivity and specificity analysis at the chosen
	OR (95%CI)	P	AUC*±SE	Cut-off levels	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Model 1: Factors associated with development of AKI
Age years	1.04 (1.021-1.083)	0.016	0.656±0.039	52	54.26	73.91	76.57	50.71
Vasopressor dose (mcg/kg/min)	1.075 (1.002-1.153)	0.041	0.651±0.039	0.032	63.57	68.12	61.10	75.80
Model 2: Logistic regression with progression to CRRT
Hypertension	2.293 (0.998-5.020)	0.05	0.593±0.027	0.5	47.06	71.58	37.20	79.08

Vasopressor dose; the norepinephrine dose; maximal dose in mcg/kg/min within the first 72 h of the ICU admission. ROC: Receiver operating characteristic curve, AUC: Area under the curve, PPV: Positive predictive value, NPV: Negative predictive value, OR: Odds ratio, CI: Confidence interval, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, ICU: Intensive care unit, SE: Standard error

Hypertension and the peak creatinine level, which were significantly high in the RRT group when compared to the No-CRRT group (47.1% vs. 28.4% P = 0.048) and 5.35 mg/dl (IQR 4.2–6.6 versus 1.8 mg/dl [IQR 1.3–2.9] P < 0.001) respectively.

Patients in the RRT and the no-RRT had similar clinical characteristics [Tables 1 and 2]. Hypertension increased the odds of AKI progression (OR: 2.293, 95% CI: 0.998–5.020, P = 0.05) [Table 5- Model 2].

Table 2

Patients’ clinical characteristics stratified according to the clinical course of coronavirus disease-2019-associated acute kidney injury

Variables	All patients (n=198; 100%)		AKI (n=129; 65%)		No-AKI (n=69; 35%)	P		No-CRRT (n=95; 74.6%)
AKI stage
Stage 1	53 (26.7)		53 (41.1)		N/A	N/A		53 (55.8)
Stage 2	24 (12.1)		24 (18.6)		N/A	N/A		18 (24.2)
Stage 3	52 (26.2)		52 (40.6)		N/A	N/A		19 (20.0)
Respiratory variables
PF ratio**	0.895 (0.975-3.150)		0.85 (0.66-1.31)		1.02 (0.75-1.4)	0.103		0.90 (0.69-1.33)
ARDS severity***
Mild	13 (6.5)		9 (7)		4 (5.9)	0.169		6 (6.3)
Moderate	74 (37.6)		43 (33.3)		32 (46.3)			34 (35.8)
Severe	110 (55.8)		77 (59.7)		33 (48.5)			55 (57.9)
Invasive mechanical ventilation	195 (98.5)		128 (99.2)		67 (97.1)	0.755		128 (98.9)
High PEEP†	177 (89.4)		118 (91.5)		59 (85.5)	0.329		88 (92.6)
Hemodynamics parameters‡
Shock	163 (82.3)		109 (88.4)		54 (78)	0.25		82 (86.3)
Mean arterial pressure	76 (71-85)		79 (73-88)		75 (70-83)	0.042		75 (69-81)
Vasopressor dose (IQR)A,B	0.039 (0.017-0.08)		0.048 (0.021-0.1)		0.03 (0.01-0.04)	0.037		0.05 (0.023-0.100)
Fluid balance till ICU admission
Negative	35 (17.7)		24 (18.6)		11 (15.9)	0.898		19 (20.0)
Iso-balance/positive	163 (82.3)		105 (81.4)		58 (84.1)			76 (80.0)
Clinical severity score
APACHE II score	10.9 (9.9-12.9)		11.9 (10.9-13.9)		10.9 (8.9-10.9)	0.001		11.8±0.4
SAPS II	24 (22-29)		27±0.54		24±0.58	0.001		26.8±0.06
SOFA score	4.883±0.890		5.06±0.103		4.492±0.158	0.002		6.14±0.127
Cardiovascular	1010±0.0533		1077±0.052		0.884±0.067	0.024		0.263±0.06
Respiratory	3.46±0.05		3.52±0.057		3.35±0.094	0.102		3.51±0.06
Renal	0.292±0.045		0.174±0.082		0.356±0.054	0.056		0.368±0.536
Neurological	0.152±8		0.01±0.01		0.29±0.02	0.246		0.0316±0.01
Liver	0.0202±0.01		0.0233±0.015		0.145±0.0145	0.678		0.0842±286
Coagulation	0.657±0.176		0.077±0.0236		0.043±0.0247	0.359		0.0842±286
APACHE II score	11.32±0.174		11.89±0.217		10.2±0.249	0.001		11.8±0.4
SAPS II	24 (22-29)		27±0.54		24±0.57 0.58	0.001		26.8±0.06
Variables	CRRT (n=34; 26.4%)	P	No-RRT (n=80)		P	RRT (n=33)		P
			Complete (n=42; 52.5%)	Partial (n=38; 47.5%)		Partial (n=18; 54.5%)	No recovery (n=15; 45.5%)
AKI stage
Stage 1	0	0	33 (78.6)	15 (39.5)	00.001	0	0	0.37
Stage 2	1 (2.9)	0	9 (21.4)	8 (21.1)		1 (5.6)	0
Stage 3	33 (97.1)	0	0	15 (39.5)		17 (94.4)	15 (100)
Respiratory variables
PF ratio**	0.74 (0.58-1.20)	0.904	90.5 (70.0-145.0)	87.7 (65.2-112.0)	0.531	116 (73.4-180)	62.8 (56-77)	0.003
ARDS severity***
Mild	3 (8.8)	0.47	4 (9.5)	1 (2.6)	0.772	3 (16.7)	0	0.03
Moderate	9 (26.5)		15 (35.7)	14 (36.8)		8 (44.4)	1 (6.7)
Severe	22 (64.7)		23 (54.8)	23 (60.5)		6 (38.9)	14 (93.3)
Invasive mechanical ventilation	52 (100)	0.552	38 (90.5)	37 (97.4)	0.898	18 (100)	15 (100)	N/A
High PEEP†	30 (82.2)	0.453	38 (90.5)	37 (97.4)	0.678	17 (94.4)	12 (80)	0.266
Hemodynamics parameters‡
Shock	27 (79.4)	0.334	36 (85.7)	33 (86.8)	0.698	12 (66.7)	14 (93.3)	0.158
Mean arterial pressure	77 (73-85)	0.267	76 (71-88)	72 (69-76)	0.02	84 (78-86)	74 (70-77)	0.157
Vasopressor dose (IQR)A,B	0.04 (0.02-0.07)	0.567	0.04 (0.32-0.12)	0.05 (0.032-0.12)	0.143	0.034 (0-0.05)	0.04 (0-0.05)	0.998
Fluid balance till ICU admission
Negative	5 (14.7)	0.5	9 (14.8)	10 (26.3)	0.146	1 (5.6)	4 (26.7)	0.427
Iso-balance/positive	29 (85.3)		36 (85.7)	28 (73.7)		17 (94.4)	11 (73.3)
Clinical severity score
APACHE II score	11.7±0.25	0.796	11.6±0.314	11.8±0.46	0.678	10.97±0.62	12.5±0.45	0.285
SAPS II	27.4±1.16	0.652	26.8±0.85	26.4±0.98	0.788	25.4±1.8	29.6±1.3	0.086
SOFA score	6.73±0.208	0.150	4.95±0.144	5.263±0.205	0.213	4.222±0.262	5.333±0.211	0.154
Cardiovascular	0.147±0. 04	0.17	1.047±0.089	1.0132±0.0238	0.362	0.722±0.135	1.133±0.133	0.04
Respiratory	5.97±0.12062	0.971	3.4545±0.11652	3.578±0.089	0.090	3.1677±0.1165	3.9333±0.1033	0.02
Renal	0.325±0.061	0.741	0.357±0.890	0.421±0.111	0.651	0.222±0.1294	0.2667±0.1817	0.840
Neurological	0	0.298	0	0.0238±0.0238	0.520	0	0	N/A
Liver	0.0.588±40.9	0.58	0.238±0.238	0.263±0.263	0.783	0.994	0	N/A
Coagulation	0.0.588±40.9	0.638	0.0476±0.033	0.1053±0.0545	0.335	0	0.111±0.72	N/A
APACHE II score	11.7±0.25	0.796	11.6±0.314	11.8±0.46	0.678	10.97±0.62	12.5±0.45	0.285
SAPS II	27.4±1.16	0.652	26.8±0.85	26.4±0.98	0.788	25.4±1.8	29.6±1.3	0.086

**PF ratio (PaO2/FiO2) ratio; the highest PaO2 partial pressure of arterial oxygen; FiO2, percentage of inspired oxygen on the day 1 of ICU, ARDS severity, Mild (200 mmHg

Renal recovery

In the analysis of renal recovery, there were 113 patients who were included and the remaining 16 patients whose renal outcomes could not be followed up completely because they survived <7 days from the onset of AKI. Thirty-eight percent of the patients had complete renal recovery (43/113), partial renal recovery was observed in 48.6% (55/113), and the remaining 15 patients had no recovery from the AKI [Table 6 and Figure 1].

Table 6

Main follow-up outcomes

Variables	All patients, n (%)	AKI*	No-AKI	P	No-CRRT	CRRT	P
Renal recovery*
Complete	43 (21.7)	43 (38.1)	N/A	N/A	43 (53.8)	0	0
Partial	55 (27.8)	55 (48.7)	N/A	N/A	37 (46.2)	18 (54.5)	0
No recovery	15 (7.5)	15 (13.3)	N/A	N/A	0	15 (45.5)	0
CKD	11	11	N/A	N/A	2	9	0
Survival and mortality rates
Survived	79 (39.9)	44 (34.1)	35 (50.7)	0.017	31 (32.6)	13 (38.2)	0.349
Death (total)	119 (60.1)	85 (65.9)	34 (49.3)		64 (67.4)	21 (61.5)
Discharge destination
Home	45 (57.0)	19 (44.2)	26 (74.3)	0.005	16 (51.6)	3 (23.1)	0.078
Long-term skilled nursing facility	34 (43.0)	25 (56.8)	9 (25.7)		15 (48.4)	10 (76.9)

*Number=113. AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, CKD: Chronic kidney disease

The univariate analyses of the patient clinical characteristics are shown in Tables 1 and 4. Factors that were identified as significant in the univariate analysis were entered in the multivariate analysis [Table 7]. We found that MAP predicted complete renal recovery for patients in the no-CRRT group (OR: 1.084, 95% CI: 1.016–1.156, P = 0.014) and PF ratio was a predictor of the renal recovery in the CRRT group (OR: 0.015, 95% CI: 0.01–0.505, P = 0.019).

Table 7

Predictors of the renal recovery of the coronavirus disease-2019-associated acute kidney injury

Variable	Multivariate logistic regression		ROC analysis		Sensitivity and specificity analysis at the chosen
	OR (95%CI)	P	AUC±SE	Cut-off levels	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Model 3: Factors associated complete renal recovery from AKI in the no-CRRT group
MAP (mmHg)	1.084 (1.016-1.156)	0.014	703±0.059	75	54.76	78.95	74.19	61.22
Model 4: Factor associated renal recovery from AKI for the CRRT group
P/F ratio	1.046 (1.0427-1.0799)	0.019	0.744±0.091	65	75.00	71.43	74.19	61.22

PF ratio (PaO2/FiO2) ratio; the highest PaO2 partial pressure of arterial oxygen; FiO2, percentage of inspired oxygen on the day 1 of ICU, MAP in mmHg on the ICU admission. ROC: Receiver operating characteristic curve, AUC: Area under the curve, PPV: Positive predictive value, NPV: Negative predictive value, MAP: Mean arterial pressure, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, SE: Standard error, OR: Odds ratio, CI: Confidence interval

Survival and mortality analysis

The overall death rate was 60.1% (119/189). Patients with AKI had a higher death rate of 65.2% versus 50.7 for patients without AKI and P = 0.035. The Cox regression analysis showed statistically significant differences in the proportional hazards for inpatient mortalities between the patients with AKI and both AKI groups. However, considering time as a covariate, we find that these differences were insignificant [Table 8].

Table 8

Analysis of relative risk for inpatient mortalities

	Total number of patients	Mortality rate, n (%)	Hazards ratio with 95% CI for in-patient mortality
			Proportional hazard	Nonproportional	P*
Model 1: Analysis of inpatient mortality with stratified for the included patients
No-AKI group	69	34 (49.3)	2.428 (1.311-4.496)	1.010 (0.966-1056)	0.653
No-CRRT	95	64 (67.4)	3.036 (1.748-5.274)	1.006 (0.988-1.2025)	0.497
CRRT group†	34	21 (61.8)
Mode 2 1: Analysis of inpatient mortality with stratified for patients in stage 3 AKI the included patients
Stage 3 No-CRRT†	19	15 (78.9)	0.848 (0.277-2.592)	0.988 (0.976-1.008)	0.242
Stage 3 CRRT†	33	21 (63.6)

*P value from the nonproportional hazards model, †Indicator. AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, CI: Confidence interval

Among the 79 patients who survived, 45 patients (57%) were discharged home, and the remaining 34 patients (43%) were transferred to skilled nursing facilities. AKI increased the odds for transfer to a post-COVID-19 advanced nursing care facility by 2.293, with a 95% CI of 0.998–5.020 [Table 6]. Moreover, survivors in the CRRT group were more likely to be discharged to similar facilities 76.9% (n = 10) versus 48.9% (n = 15). Nevertheless, this difference was not statistically significant, P = 0.078.

Impact on the length of stay

The median ICU LOS for the whole study population was 19 days (IQR: 10–41 days). The survival analysis showed that the time for discharge from the ICU was significantly longer when CRRT was started regardless of the survival status at discharge; the median ICU LOS for the CRRT group was 58 days (IQR: 22–90) versus 14 days (IQR: 8–26) for patients in the no-CRRT group and 16 days (IQR: 9–24) for the no-AKI group; P < 0.001 [Table 9 and Figure 2].

Table 9

Analysis of intensive care unit length of stay

Variables	All patients	No-AKI	No-CRRT group	CRRT group	P*
All patients	19 (10-41)	16 (9-24)	14 (8-26)	58 (22-87)	0.001
For survival	22.5 (13-43)	17 (12-36)	15 (9-34)	77 (66-87)	0.001
For dead	15 (8-25.5)	14.5 (8-19)	12 (9-21)	19.5 (11-84)	0.002

*P Significant if < 0.05 by log-rank test, †Median and (IQR). IQR: Interquartile range, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy

Figure 2

Kaplan–Meier survival plots for intensive care unit length of stay for survived patients (a) Kaplan–Meier survival plots for intensive care unit length of stay for the deceased patients (b)

DISCUSSION

Not only did this study confirm the high incidence of COVID-19-associated AKI but also it demonstrated that renal involvement was progressive with a high rate of CRRT. Consistently, previous reports of ICU patients with severe SARS-COV 2 infection revealed that the prevalence of AKI ranged from 50% to 80%, with an unusual surge in extracorporeal renal replacement therapy requirements, up to 40% in some studies exceeding early reports from China.[19202122232425]The raised incidence of AKI and CRRT in this study could be explained by the high rate of comorbidities and the severity of the hypoxemia, which necessitated mechanical ventilatory support and other illness severity indicators. These factors are assumed to be the basis for the lung–kidney cross-talk and AKI.[21126]Meanwhile, it may be debated that the center size was a confounding factor that inappropriately increased the observed rate of AKI in our cohort. However, the high prevalence of AKI in similar COVID-19 ICU patients makes this assumption improbable.[1920212223]

To correlate our findings with the pathophysiology, we divided the potential mechanisms for COVID-19-associated AKI into three main categories: the nonspecific mechanisms, COVID-19-specific mechanisms, and host-immune responses.[211]

First, the nonspecific mechanisms such as age and comorbidities are recognized risk factors of renal vulnerability in ICU patients.[2728]Hypovolemia in some of these patients can cause pre-renal AKI.[2]Other factors include severe hypoxemia and cardiorenal syndrome secondary to the effects of positive pressure ventilation.[293031323334]Furthermore, nephrotoxic exposure such as drugs might also be involved in the pathogenesis of AKI.[12]

Second, COVID-19-specific mechanisms include viral entry into the renal cell, which could provoke acute tubular injury (ATI) and collapsing glomerulopathy.[34353637]Other components of these mechanisms include the overactivation of the renin–angiotensin system (RAS) and the SARS-CoV-2-associated procoagulant state.[113839404142]

COVID-19-specific mechanisms start by cleaving of viral spike (S) protein by host cell proteases such as transmembrane protease, serine 2 in the proximal tubules following that the (S) proteins bind to the cellular receptor; Angiotensin conversion enzyme 2 (ACE2). In the human kidney, ACE2 is expressed on proximal tubules and podocytes.[36]The entry of the virus to tubular and podocytes cell can cause tubular dysfunction and ATI, collapsing glomerulopathy.[34353637]The attachment of SARS-CoV-2 to ACE2 promotes downregulation of membrane-bound ACE2 that leads to the accumulation of angiotensin II (AngII) responsible for the overactivation of RAS, resulting in vasoconstriction, ischemia, inflammation, and subsequently endothelial dysfunction and ATI.[383940]Another effect of reduced ACE2 levels may be the increase in bradykinin levels, promoting coagulopathy in at least two distinct ways: (1) enhancing complement activation and promoting neutrophil activation and NET formation. Inflammation generally activates the complement system, and subsequent depletion of plasma complement is associated with increasing disease severity and a prothrombotic state.[11404142]

Third, the host immune responses secondary to severe SARS-CoV-2 infection can contribute to AKI.[1211]The early notion of cytokine storm syndrome being the hallmark of the immune response has been replaced by current evidence supporting the presence of two different host immune responses: (1) the hyperinflammatory response, which involves the release of pro-inflammatory cytokines (tumor necrosis factor [TNF] α, interleukin [IL]-6, IL-1 β) and chemokines (C-X-C motif ligand [CXCL], CXCL10), (2) the hypoinflammatory response manifests by a disease similar to immunoparalysis in sepsis with more detrimental multiple organ failure, including the AKI.[11434445464748]

This study showed that the patient's age was associated with an increase in the risk of AKI, whereas other comorbidities did not have a similar significant association. Consistently, age has been demonstrated to be a risk factor to AKI in a similar population.[2021222324]Likewise, we showed that hemodynamic instability was mild and corrected by a minimal dose of vasopressor, which denotes that it was primarily attributed to the use of sedation.[4950]

The inflammatory markers such as C-reactive protein and the other ubiquitous risk factors were similar in patients with AKI and those without AKI. Accordingly, it can be inferred that the nonspecific mechanisms were not the major risk factor for AKI. We also showed that renal abnormalities associated with COVID-19 include proteinuria, hematuria, and glucosuria. Therefore, it could be assumed that the main pathophysiological factors were COVID-19-specific mechanisms.

Furthermore, we demonstrated that some degree of renal recovery was observed in the majority of the patients. In addition, we showed that early factors on ICU admission such as the PF and the MAP could predict renal outcomes. These data agree with renal recovery reports from COVID-19-associated AKI.[5051]

Consistent with previous studies, we demonstrated the association between AKI development and inpatient mortalities.[52535455]Interestingly, the death rate among the study population was high, particularly in the AKI group; possible explanations could be that our patient had a higher illness severity and comorbidities scores. Furthermore, the prolonged follow-up period in this study, particularly for the AKI group, may have had a role in that finding. We demonstrated that CRRT did not change the cumulative inpatient mortality risks. Nevertheless, there are significant discrepancies in the reports of the impact of CRRT on the death rate among similar patients with AKI.[20212223245356]These disagreements may be explained by the observation period's duration and the local clinical practice.[45]

This study showed the excessive burden of AKI on the health-care facility such as unreasonable requirement RRT, unduly prolonged ICU LOS, and the increased risk of discharge to a skilled nursing facility. These data agree with the previous report of COVID-19-associated AKI.[12345]

Our study has certain limitations. It was a retrospective, observational study limited to a single center, and female patients were underrepresented; therefore, the generalization of our findings could be limited. Among this study's strengths, our cohort was limited to ventilated critically ill patients with COVID-19 pneumonia. Our follow-up period was 3 months, and there was a final discharge place reached at the end of this period for our patients. Finally, it showed a high prevalence of COVID-19-associated AKI and its influence on the outcome. Therefore, it draws the attention of intensivists toward the kidneys in the context of severe COVID-19 pneumonia.

CONCLUSION

Among this cohort of ventilated critically ill adult patients with SARS-CoV-2 pneumonia, there was a high prevalence of COVID-19-associated AKI, which was progressive but frequently reversible. AKI was largely ATN and essentially attributable to COVID-19 specific mechanisms. It was associated with poor outcomes, and it imposed an extra burden on our health-care facility.

Research quality and ethics statement

This study was approved by the Dubai Scientific Research Committee (DSRE C), Dubai Health Authority (approval reference DSREC-06/2020_27; approval date July 2, 2020). Informed consent was waived due to the retrospective observational nature of our study. We followed the STROBE guidelines (the Strengthening the Reporting of Observational Studies in Epidemiology).

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 3

Patients’ laboratory finding stratified according to the clinical course of coronavirus disease-2019-associated acute kidney injury

Variables	All patients (n=198; 100%)	AKI (n=129; 65%)		No-AKI (n=69; 35%)	P	No-CRRT (n=95; 74.6%)	CRRT (n=34; 26.4%)
Laboratory finding
Urea	17 (12-24)	18 (13-26)		16 (11-21)	0.281	18 (14-26)	18 (13-25)
Creatinine
Baseline	0.9 (0.7-1.1)	0.9 (0.7-1.23)		0.8 (0.7-1.00)	0.101	0.9 (0.7-1.3)	0.9 (0.8-1.1)
Peak	1.5 (0.97-3.15)	2.45 (1.4-4.78)		0.9 (0.7-1.28)	N/A	1.8 (1.3-2.9)	5.35 (4.2-6.6)
Days till AKI start	N/A	N/A		2.16±0.095	N/A	2.08±0.114	2.3083±0.190
CRP on ICU admission*	168 (107-247)	175 (112-242)		158 (88-248)	0.734	177.5 (107-237)	177.5 (131-272)
Peak CPK**	1454 (522-3236)	1617 (573-3394)		1232 (489-2541)	0.075	1617 (573-3394)	1566 (594-3467)
Urine analysis‡
Proteinuria (total)	157 (79.3)	100 (77.5)		49 (71)	0.705	66 (75.3)	34 (100)
None	41 (20.7)	29 (22.5)		20 (29.0)	0.705	29 (30)	0
Mild (+)	150 (75.8)	96 (74.4)		46 (66.7)	0.799	62 (65.3)	34 (100)
Heavy (+2)	7 (3.5)	4 (3.1)		3 (4.3)	n/a	4 (4.2)	0
Hematuria	134 (67.7)	84 (65.1)		50 (85.2)	0.542	54 (56.8)	30 (70.6)
+1 dipstick	121 (61.1)	76 (58.9)		45 (65.2)	0.533	54 (56.8)	22 (64.7)
≥+2 dipstick	13 (6.6)	8 (6.2)		5 (7.2)	0.913	6 (6.3)	2 (5.9)
None	170 (85.9)	112 (86.8)		59 (86.8)	0.421	81 (85.3)	31 (91.2)
Leucocyturia	10 (5.1)	7 (5.4)		3 (4.3)	0.424	5 (5.3)	2 (5.9)
Glycosuria	18 (9.2)	10 (7.8)		8 (11.6)	0.511	9 (9.5)	1 (2.9)
Variables	P	No-RRT (n=80)		P	RRT (n=33)		P
		Complete (n=42; 52.5%)	Partial (n=38; 47.5%)		Partial (n=18; 54.5%)	No recovery (n=15; 45.5%)
Laboratory finding
Urea	0.856	19.5 (15-30)	18 (15-24)	0.265	21 (15-33)	17 (12.5-22)	0.128
Creatinine
Baseline	0.902	0.9 (0.7-0.13)	0.9 (0.9-0.13)	0.491	0.9 (0.8-1.1)	0.9 (0.8-1.05)	0.833
Peak	0	1.4 (1.2-1.9)	2.8 (1.8-4.5)	0.094	5 (3.7-6)	5.4 (4.7-7.4)	0.481
Days till AKI start	0.168	2.14±0.22	2.42±1.39		2.55±0.245	2.06±0.0118	0.112
CRP on ICU admission*	0.337	170 (103-233)	167 (112-237)	0.265	147 (121-207)	236 (168-318)	0.022
Peak CPK^	0.147	1474 (760-2687)	1904 (490-3399)	0.178	2045 (447-4548)	1317 (853-2734)	0.096
Urine analysis‡
Proteinuria (total)	0.729	33 (78.5)	25 (65.8)	0.872	18 (100)	15 (100)	N/A
None	0.669	9 (21.4)	13 (34.2)	0.883	18 (100)	15 (100)	N/A
Mild (+)	0.791	31 (73.8)	23 (60.5)	0.987	0	0	N/A
Heavy (+2)	0.88	2 (4.8)	2 (5.3)	0.304
Hematuria	0.669	29 (69.5)	25 (65.8)	0.903	13 (61.10)	10 (66.70)	0.752
+1 dipstick	0.791	25 (59.5)	24 (63.2)	0.752	11 (61.10)	10 (66.70)	0.752
≥+2 dipstick	0.88	4 (9.5)	1 (2.6)		2 (11.1)	0
None	0.174	32 (76.2)	36 (94.7)	0.808	15 (83.3)	12 (80)	0.897
Leucocyturia	0.062	4 (9.5)	0	0.611	2 (11.1)	0	0.171
Glycosuria	0.288	6 (14.3)	2 (5.3)	0.292	1 (5.6)	0	N/A

*CRP on ICU admission, **CPK, the maximal level within 7 days of ICU admission, ‡From urine analysis on ICU admission. All variables are presented as, Categorical variables: n (%), Continuous variable with normal of distribution: Means with (SEM±SEM), Continuous variable with skewed distribution: Median and IQR, P value, from univariate analysis. bold if significance (P<0.05). N/A: Not applicable, CRP: C-reactive protein, AKI: Acute kidney injury, CRRT: Continuous renal replacement therapy, RRT: Renal replacement therapy, ICU: Intensive care unit, IQR: Interquartile range, SEM: Standard error of mean, CPK: Creatine phosphokinase