Lipoprotein concentrations over time in the intensive care unit COVID-19 patients: Results from the ApoCOVID study.

INTRODUCTION: Severe acute respiratory syndrome coronavirus2 has caused a global pandemic of coronavirus disease 2019 (COVID-19). High-density lipoproteins (HDLs), particles chiefly known for their reverse cholesterol transport function, also display pleiotropic properties, including anti-inflammatory or antioxidant functions. HDLs and low-density lipoproteins (LDLs) can neutralize lipopolysaccharides and increase bacterial clearance. HDL cholesterol (HDL-C) and LDL cholesterol (LDL-C) decrease during bacterial sepsis, and an association has been reported between low lipoprotein levels and poor patient outcomes. The goal of this study was to characterize the lipoprotein profiles of severe ICU patients hospitalized for COVID-19 pneumonia and to assess their changes during bacterial ventilator-associated pneumonia (VAP) superinfection.
METHODS: A prospective study was conducted in a university hospital ICU. All consecutive patients admitted for COVID-19 pneumonia were included. Lipoprotein levels were assessed at admission and daily thereafter. The assessed outcomes were survival at 28 days and the incidence of VAP.
RESULTS: A total of 48 patients were included. Upon admission, lipoprotein concentrations were low, typically under the reference values ([HDL-C] = 0.7[0.5-0.9] mmol/L; [LDL-C] = 1.8[1.3-2.3] mmol/L). A statistically significant increase in HDL-C and LDL-C over time during the ICU stay was found. There was no relationship between HDL-C and LDL-C concentrations and mortality on day 28 (log-rank p = 0.554 and p = 0.083, respectively). A comparison of alive and dead patients on day 28 did not reveal any differences in HDL-C and LDL-C concentrations over time. Bacterial VAP was frequent (64%). An association was observed between HDL-C and LDL-C concentrations on the day of the first VAP diagnosis and mortality ([HDL-C] = 0.6[0.5-0.9] mmol/L in survivors vs. [HDL-C] = 0.5[0.3-0.6] mmol/L in nonsurvivors, p = 0.036; [LDL-C] = 2.2[1.9-3.0] mmol/L in survivors vs. [LDL-C] = 1.3[0.9-2.0] mmol/L in nonsurvivors, p = 0.006).
CONCLUSION: HDL-C and LDL-C concentrations upon ICU admission are low in severe COVID-19 pneumonia patients but are not associated with poor outcomes. However, low lipoprotein concentrations in the case of bacterial superinfection during ICU hospitalization are associated with mortality, which reinforces the potential role of these particles during bacterial sepsis.

Introduction

In late 2019, the novel coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was identified as the cause of COVID-19 in Hubei Province, China [1]. COVID-19 has since become a pandemic, and by May 2020 more than 4.1 million confirmed cases had been reported, with more than 280,000 attributable deaths worldwide [2]. Five to 20% of patients hospitalized with COVID-19 are admitted to the intensive care unit (ICU), with a mortality rate ranging from 25% to 60% [3-5]. At present, there is no effective specific treatment for COVID-19 described in the surviving sepsis campaign recommendations [6].

Lipoprotein particles are defined by their composition (lipids and proteins) and are classified according to their density (from very low to high density). They have the capacity to transport hydrophobic lipids (e.g., cholesterol) in a hydrophilic environment (plasma), but they also exert numerous pleiotropic properties [7]. High-density lipoproteins (HDLs), responsible for reverse cholesterol transport (RCT), display endothelioprotective functions, including anti-inflammatory, anti-apoptotic, and antioxidant effects; they can also bind and neutralize lipopolysaccharides (LPS), enhancing LPS clearance [8-11]. Numerous clinical studies have reported a marked decrease in the concentration of HDL cholesterol (HDL-C) during bacterial sepsis [12-19], and this decrease is currently associated with poor outcomes [17-20]. Experimental studies assessing both reconstituted HDL (rHDL) and ApoA-I mimetic peptide perfusion in animal models of septic shock have been performed, demonstrating protective effects against morbidity and mortality [21-23]. Low-density lipoproteins (LDLs) also neutralize LPS [24, 25], and observational studies have reported that LDL-cholesterol (LDL-C) concentrations could decrease by 30% in inflammatory states such as sepsis [19, 26]. Walley et al. also showed that low LDL-C concentrations were associated with a poor prognosis in septic patients [27].

The links between viral infection and lipoproteins are less clear than the association between bacterial infection and lipoproteins [7]. Lipid profiles are reportedly altered during some viral infections, such as dengue, in which Total cholesterol (TC) and lipoprotein levels are frequently low; specifically, a low LDL-C concentration is correlated with severity [28, 29]. In a recent study involving 21 mixed ICU and non-ICU COVID-19 patients, LDL-C concentrations were significantly decreased upon admission and were inversely correlated with disease severity [30]. Most patients with COVID-19 in the ICU develop bacterial ventilator-associated pneumonia (VAP) [31], suggesting that both COVID-19 and bacterial infection may influence the lipid profile.

Here, we aimed to characterize the lipid profiles of patients with severe COVID-19 pneumonia upon their ICU admission and during hospitalization, with a particular focus on changes in HDL-C and LDL-C concentrations. We also assessed lipid profile changes in these patients during bacterial VAP superinfection.

Materials and methods

This was a monocentric study conducted in the surgical ICU of Bichat Claude-Bernard University Hospital, Paris, France. Patients admitted between March 18, 2020 and April 13, 2020 for acute respiratory distress syndrome (ARDS) due to COVID-19 pneumonia were consecutively and prospectively included in a database, and their medical charts were reviewed retrospectively. The French Society of Anesthesiology and Critical Care Medicine Research Ethics Board approved this study and waived the need for consent (ApoCOVID study, IRB 00010254‐2020‐082).

Patient demographic information, Simplified Acute Physiology Score II (SAPSII), Sepsis-related Organ Failure Assessment (SOFA) severity scores and clinical data were collected. Inflammatory parameters (leukocytes, lymphocytes, C-reactive protein and procalcitonin) were also assessed. Mortality at 28 days, duration of mechanical ventilation, number of days alive without mechanical ventilation at day 28, length of stay in the ICU and in the hospital, renal replacement therapy, vasopressor use, need for extracorporeal membrane oxygenation (ECMO), and tracheostomy were collected. The number of prone positioning procedures and instances of ventilator-associated pneumonia were also collected.

Plasma concentrations of TC, HDL-C, LDL-C, and triglycerides (TG) were measured upon ICU admission and then daily in the Biochemistry Laboratory of Bichat Claude-Bernard Hospital. Owing to problems in the supply of anesthetic drugs during the COVID-19 epidemic, we had to administer propofol to all patients over a long period of time, which is why we measured daily lipid levels. TC, HDL-C, LDL-C and triglyceride concentrations were determined by routine enzymatic assays (CHOL, HDL-C, LDL-C and TRIG methods, Dimension VISTA® System, Siemens Healthineers™). The reference values for these assays were as follows: HDL-C: > 1.40 mmol/L; total cholesterol (TC): 4.40 < N < 5.2 mmol/L; and triglycerides: 0.50 < N < 1.7 mmol/L. According to the French National Authority for Health 2017 and the European Society of Cardiology 2016 recommendations, LDL-C concentration targets have been established depending on vascular risk factors [32]. According to these recommendations, LDL-C <3.0 mmol/L is advised in low- to moderate-risk patients, and LDL-C <2.6 mmol/L is recommended in high-cardiovascular risk patients.

Ventilator-associated pneumonia (VAP) diagnosis was based on the Infectious Diseases Society of America and the American Thoracic Society guidelines [33].

Statistical analysis: Continuous variables were expressed as medians with interquartile ranges (IQRs) and were compared using the Mann-Whitney U test. Categorical variables were expressed as counts and percentages and were compared using Fisher’s exact test or the chi-square test, as appropriate. The threshold defining the lower quartile was determined to have 25% of the overall population in that quartile. Survival was estimated by Kaplan-Meier analysis and compared by the log-rank test. Correlations were assessed by Spearman’s rank-order correlation. A mixed model for repeated measures was performed to compare the lipoproteins’ evolution over time in the overall population and according to 28-day mortality. All statistical analyses were performed using SPSS software, version 21 (IBM, Armonk, NY, USA). P-values < 0.05 were considered statistically significant.

Results

a. Population

From March 18, 2020 to April 13, 2020, 67 patients were hospitalized for COVID-19 pneumonia in our ICU. Because of a lack of lipid profile data at admission for 19 patients, 48 patients were finally included in the study. Twenty-nine patients were directly admitted into the ICU, while 19 patients had a short stay (1 [0-3] days) in another ward before ICU admission. The general characteristics and outcomes of the patients are presented in Table 1. Comparisons between survivors (n = 32, 67%) and nonsurvivors (n = 16, 33%) at day 28 are presented in Table 1. At ICU admission, no patient had a bacterial coinfection associated with COVID-19 pneumonia.

Table 1

General characteristics and outcome of the patients.

Characteristics	Overall population (n = 48)	Alive at day 28 (n = 32; 67%)	Dead at day 28 (n = 16; 33%)	p Value
Age, years, median [IQR]	57 [46–64]	55 [45–62]	59 [50–67]	0.283
Male sex, n (%)	31 (65)	21 (65)	10 (63)	0.831
BMI, kg/m2, median [IQR]	27.9 [25–31]	27 [24–29.6]	29.7 [26–35.5]	0.135
Presence of comorbidities / medications
High blood pressure, n (%)	24 (50)	14 (44)	10 (63)	0.221
ACEI or ARB use, n (%)	16 (33)	9 (28)	7 (44)	0.279
Diabetes mellitus, n (%)	17 (35)	8 (25)	9 (56)	0.033
Statin use, n (%)	13 (27)	7 (22)	6 (37)	0.310
Timing of hospitalization
Between first symptoms and hospitalization (days)	6 [3–7]	6 [4–8]	5 [2–7]	0.235
Between hospitalization and ICU admission (days)	1 [0–3]	2 [0–3]	0 [0–4]	0.530
Severity scores at admission
SAPSII, median [IQR]	43 [33–53]	40 [31–51]	50 [44–28]	0.006
SOFA, median [IQR]	5 [4–7]	5 [4–7]	6 [4–7]	0.241
Inflammatory parameters at admission
Leukocyte count (G/L)	8.7 [6.5–12.4]	9.7 [7–14]	7.6 [6.4–11.7]	0.155
Lymphocyte count (G/L)	0.8 [0.6–1.3]	0.9 [0.6–1.3]	0.7 [0.4–1.3]	0.323
Procalcitonin (μg/L)	0.8 [0.3–3.3]	0.8 [0.4–2.7]	0.6 [0.3–3.6]	0.850
C-reactive protein (mg/L)	136 [97–219]	136 [102–209]	145 [86–246]	0.786
Treatments during ICU stay
Norepinephrine, n (%)	32 (67)	19 (60)	13 (81)	0.130
Mechanical ventilation, n (%)	46 (96)	30 (94)	16 (100)	0.546
Length of mechanical ventilation, median [IQR]	12 [7–25]	20 [7–28]	7 [6–9]	0.007
Prone positioning, n (%)	33 (69)	21 (66)	12 (75)	0.509
Tracheostomy, n (%)	11 (23)	10 (31)	1 (6)	0.073
ECMO, n (%)	10 (20)	3 (9)	7 (44)	0.010
RRT, n (%)	13 (27)	9 (28)	4 (25)	1
COVID-specific treatments
Lopinavir/ritonavir, n (%)	6 (12)	5 (16)	1 (6)	0.648
Hydroxychloroquine, n (%)	3 (6)	2 (6)	1 (6)	1
Corticosteroids, n (%)	12 (25)	10 (31)	2 (12)	0.289
Outcome
ICU LOS, median [IQR]	12 [7–27]	22 [10–33]	7 [6–9]	0.002
Hospital LOS, median [IQR]	20 [7–31]	28 [19–38]	7 [6–09]	<0.001
Mortality at day 28, n (%)	16 (33)	-	-	-

BMI, body mass index; ECMO, extracorporeal membrane oxygenation; LOS, length of stay; RRT, renal replacement therapy; SAPS II, simplified acute physiology score II; ACEI, angiotensin-converting-enzyme inhibitor; ARB, angiotensin II receptor blocker.

b. Lipid concentrations at admission and kinetics over time

Upon ICU admission, the TC, TG, HDL-C and LDL-C concentrations were 3.2 [2.5–4.0] mmol/L, 2.0 [1.6–2.9] mmol/L, 0.7 [0.5–0.9] mmol/L and 1.8 [1.3–2.3] mmol/L, respectively. All results except the TG concentrations were below the abovementioned reference values (see Materials and Methods section).

Fig 1 exhibits the kinetics of the TC, TG, HDL-C and LDL-C concentrations throughout the ICU stay. Except for the TG concentration, which remained within the normal reference value range, we observed a statistically significant increase in TC, HDL-C and LDL-C during the study period (p<0.001, p = 0.024 and p<0.001, respectively); these values returned to normal in the survivors.

Fig 1

Total cholesterol, triglycerides, HDL-C and LDL-C kinetics during the ICU stay.

c. Relationship between lipid concentrations and COVID-19-specific therapies

We found no differences in the lipid concentrations upon admission or over time between patients with and without lopinavir/ritonavir treatment (see S1 Fig). In addition, corticosteroid therapy did not significantly alter the lipid concentrations upon admission or over time (see S2 Fig).

d. Relationship between lipid concentrations and mortality

Upon ICU admission, there were no difference in lipid concentrations between survivors and nonsurvivors (TC: 3.3 [2.5–4.0] mmol/L in survivors vs. 3.0 [2.2–4.1] mmol/L in nonsurvivors, p = 0.43; TG: 2.0 [1.5–2.9] mmol/L in survivors vs. 2.0 [1.7–3.0] mmol/L in nonsurvivors; p = 0.827; HDL-C: 0.7 [0.5–1.1] mmol/L in survivors vs. 0.7 [0.4–0.8] mmol/L in nonsurvivors, p = 0.459; LDL-C: 1.9 [1.4–2.4] in survivors vs. 1.5 [1.1–2.2] mmol/L in nonsurvivors, p = 0.262). In addition, changes in the TC, TG, HDL-C and LDL-C concentrations over time during the first six days following ICU admission did not allow us to separate patients as alive or dead at day 28 (Fig 2).

Fig 2

Kinetics of total cholesterol, triglycerides, HDL-C and LDL-C concentrations over the first six days according to their status outcome (dead or alive at day 28).

Fig 3 shows the mortality at day 28 according to the lipid profile upon ICU admission as estimated by the Kaplan-Meier analysis and compared by the log-rank test. No relationship was found between patients with TC, TG, HDL-C and LDL-C concentrations in their respective lower quartile and mortality at day 28 (log-rank p = 0.092, p = 0.611, p = 0.554 and p = 0.083, respectively).

Fig 3

Mortality at day 28 according to the lipid profile upon ICU admission.

Survival analyses were estimated by Kaplan-Meier analysis and compared using the log-rank test.

e. Relationship between the lipid concentrations and patient outcomes

The data in Table 2 show the relationships between TC, TG, HDL-C and LDL-C concentrations upon ICU admission and the ICU outcomes (ventilator-associated pneumonia, renal replacement therapy, need for norepinephrine). No statistically significant links were found between TC, TG, HDL-C or LDL-C at admission and these ICU outcomes.

Table 2

Relationship between lipid concentrations at ICU admission and outcome variables.

Lipid concentrations at ICU admission	Overall population (n = 48)	VAP (n = 29; 64%)	No VAP (n = 19; 36%)	p Value
Total cholesterol, mmol/L, median [IQR]	3.2 [2.5–4.0]	3.1 [2.5–3.9]	3.5 [2.5–4.2]	0.349
Triglycerides, mmol/L, median [IQR]	2.0 [1.6–2.9]	2.1 [1.8–3.0]	1.9 [1.4–2.6]	0.260
HDL-C, mmol/L, median [IQR]	0.7 [0.5–0.9]	0.7 [0.4–0.8]	0.8 [0.6–1.1]	0.077
LDL-C, mmol/L, median [IQR]	1.8 [1.3–2.3]	1.7 [1.1–2.3]	2.0 [1.4–2.6]	0.266
Lipid concentrations at ICU admission	Overall population (n = 48)	RRT (n = 13; 27%)	No RRT (n = 35; 73%)	p Value
Total cholesterol, mmol/L, median [IQR]	3.2 [2.5–4.0]	3.4 [2.5–4.0]	2.8 [2.4–4.0]	0.609
Triglycerides, mmol/L, median [IQR]	2.0 [1.6–2.9]	1.9 [1.5–2.8]	2.1 [1.8–2.9]	0.372
HDL-C, mmol/L, median [IQR]	0.7 [0.5–0.9]	0.8 [0.5–1.0]	0.6 [0.4–0.8]	0.164
LDL-C, mmol/L, median [IQR]	1.8 [1.3–2.3]	1.9 [1.4–2.4]	1.5 [1.1–2.2]	0.439
Lipid concentrations at ICU admission	Overall population (n = 48)	NOR (n = 32; 67%)	No NOR (n = 16; 33%)	p Value
Total cholesterol, mmol/L, median [IQR]	3.2 [2.5–4.0]	3.4 [2.5–3.9]	3.1 [2.5–4.1]	0.864
Triglycerides, mmol/L, median [IQR]	2.0 [1.6–2.9]	2.2 [1.7–2.8]	1.9 [1.6–2.9]	0.743
HDL-C, mmol/L, median [IQR]	0.7 [0.5–0.9]	0.7 [0.6–1.1]	0.7 [0.5–0.9]	0.458
LDL-C, mmol/L, median [IQR]	1.8 [1.3–2.3]	1.8 [1.4–2.3]	1.7 [1.2–2.4]	0.927

Variables: ventilator-associated pneumonia, renal replacement therapy, need for norepinephrine and need for extracorporeal membrane oxygenation. Continuous variables are expressed as the median and interquartile range (IQR). VAP: ventilator-associated pneumonia; RRT: renal replacement therapy; NOR: norepinephrine.

There were no correlations between TC, TG, HDL-C or LDL-C concentrations upon ICU admission and length of ICU stay, length of hospital stay, or the number of days alive without mechanical ventilation (S1 Table).

f. Relationship between lipid concentrations and mortality in the subgroup of patients with bacterial ventilated-associated pneumonia (VAP)

To better characterize the relationship between the TC, TG, HDL-C and LDL-C concentrations and the outcomes, we focused on COVID-19 patients with bacterial VAP during their stay in the ICU (n = 29, 64%). The median period of time before the onset of VAP was 7 [3-9] days, and the median number of VAP episodes for each patient was 2 [1-3]. Concentrations of TC, TG, HDL-C, and LDL-C were measured on the day of VAP diagnosis. A statistically significant association was found between HDL-C and LDL-C concentrations on the day of the first VAP diagnosis and mortality on day 28 ([HDL-C] = 0.6 [0.5–0.9] mmol/L in patients alive on day 28 vs. [HDL-C] = 0.5 [0.3–0.6] mmol/L in patients dead on day 28, p = 0.036; [LDL-C] = 2.2 [1.9–3.0] mmol/L in patients alive on day 28 vs. [LDL-C] = 1.3 [0.9–2.0] mmol/L in patients dead on day 28, p = 0.006). No statistically significant association was found between the TC or TG concentrations and mortality at day 28. Box plots are presented in Fig 4.

Fig 4

Relationship between total cholesterol, triglycerides, HDL-C and LDL-C concentrations on the day of the first VAP diagnosis and mortality on day 28.

Discussion

In this cohort of patients who were hospitalized in the ICU for COVID-19 pneumonia, we observed that lipoprotein concentrations upon ICU admission were low but were not significantly related to the patient’s outcomes. In contrast, in the subgroup of patients who developed bacterial VAP, a strong association was found between mortality and the concentrations of HDL-C and LDL-C at the time of VAP diagnosis.

The low concentrations of HDL-C and LDL-C upon admission reported here are consistent with previous results obtained in two studies conducted by the same Chinese team in a cohort of heterogeneous patients hospitalized for COVID-19 [30, 34]. In these studies, the authors compared the lipid profiles of COVID-19 patients and healthy subjects and reported differences between the two groups. However, it should be noted that the severity scores of these two populations (COVID-19 and healthy volunteers) were not compared. The study involving 597 COVID-19 patients stratified into mild (n = 394), severe (n = 171) and critical (n = 32) cases showed that LDL-C and TG concentrations were significantly different between the groups, with lower concentrations in the most severe patients [34]. Moreover, low levels of HDL-C and LDL-C in our population are in accordance with another study involving 114 COVID-19 patients [35]. In this study, TC, HDL-C and LDL-C concentrations in COVID-19 patients were low but were also significantly lower than the values of age-matched healthy control patients.

During bacterial sepsis, a decrease in lipoproteins during the acute phase is well documented (in particular for HDL-C); however, the mechanisms underlying this decrease are poorly described [11]. Several hypotheses have been proposed to explain this HDL-C concentration decrease such as the consumption of HDL particles, hemodilution, capillary leakage, decreased HDL synthesis by the liver (particularly in cases of associated liver dysfunction) or increased HDL clearance following the upregulation of scavenger receptor class B type 1 (SRB1) expression [20]. During COVID-19, similar potential mechanisms might be present. A pulmonary histological study involving COVID-19 patients documented increased vascular permeability [36]. The considerable inflammation described during COVID-19 ARDS could potentially participate in the lipoprotein concentration decrease [37, 38]. Additionally, the direct action of the virus on lipoproteins should not be ruled out, but this hypothesis requires further investigation.

Notably, we have shown a gradual increase over time in both the HDL-C and LDL-C concentrations. These increases over time could be related to patient improvement, the restoration of normovolemia, a reduced inflammatory state, decreased capillary leakage or increased hepatic HDL synthesis [7, 11]. The findings described herein of decreased lipoprotein levels at admission followed by increases over time are in accordance with the studies by Fan et al. and Hu et al. [30, 39].

In our cohort of 48 severe COVID-19 patients hospitalized in our ICU, we found no statistically significant link between HDL-C or LDL-C concentrations and patient outcomes. These results are not in accordance with the results of the study by Fan et al., which showed that a low LDL-C concentration was a potential predictor of poor prognosis in a cohort of 21 heterogeneous ICU and non-ICU COVID-19 patients [30]. Interestingly, the studies by Hu et al. and Wei et al. have specifically shown that low HDL-C levels were associated with severe COVID-19 disease [34, 39]. However, the heterogeneity reported in these two studies, which included mild, severe and critical patients and ICU and non-ICU patients, does not allow for direct comparisons with our specific severe COVID-19 ICU cohort [34, 39].

In contrast to studies involving septic patients due to bacteria, in which HDL-C and LDL-C concentrations are correlated with the outcome [11, 17, 19, 22, 27], data concerning viral infections appear to be more controversial. While altered lipid profiles have been described in some viral pathologies, such as HIV (frequently associated with increased TC and LDL-C and decreased HDL-C) and hepatitis B virus infections, direct links to patient outcomes remain uncertain [40-42]. We hypothesize that during sepsis in the context of bacterial infections, lipoprotein concentrations seem to be a determinant for the patient’s outcome. Pleiotropic properties of lipoproteins and, in particular, of HDL particles, such as anti-inflammatory, anti-apoptotic or antioxidant effects, could play an important role. The ability of HDL to bind and inactivate LPS and lipoteichoic acid, leading to a potential increase in bacterial clearance, appears to be an essential function that could support the strong association between the HDL concentrations and patient outcomes [11, 43, 44]. The lack of a statistically significant association in our study focusing on viral pathology might be related to the specific action of HDL particles on bacteria. The relationship between mortality and HDL-C or LDL-C concentrations in the subgroup of patients with bacterial respiratory coinfection supports this hypothesis. However, these preliminary results suggest that the viral inflammatory model has a major impact on lipoprotein metabolism and deserves additional investigation. In particular, the dysfunction of lipoproteins (and especially of HDL particles) may be interesting to explore in relation to COVID-19 pneumonia [45, 46]. Because no patients in our cohort had bacterial coinfection at admission and owing to the late delay before the onset of VAP (7 [3-9] days), COVID-19 appears to be an interesting model to 1) study the effects of viral infection and 2) the effect of bacterial VAP on the lipid profile.

Interestingly, compared to the HDL-C, LDL-C and TC concentrations, the TG concentration at admission was higher than the normal range. Several hypotheses can explain this finding. First, one-third of the patients had diabetes, a condition in which lipid metabolism disorders are frequent (particularly increased TG concentrations). Second, the early use of propofol upon admission may explain the increased TG concentration. Third, renal dysfunction and especially nephrotic-like syndrome is observed in COVID-19 disease, which may have induced an increased level of TG. Finally, direct effects of the virus cannot be excluded.

Our study has several limitations. First, it is a small monocentric study of limited size. Second, we did not measure cytokine levels; these could have been used to stratify the patients according to their inflammatory state. Third, we did not compare lipid concentrations upon admission with basal concentrations prior to hospitalization. Fourth, although the lipid concentrations of patients treated with and without corticosteroids or protease inhibitors were similar over time, these treatments have the potential to induce lipids disorders (such as increasing the TC, TG, LDL-C concentrations and decreasing the HDL-C concentration), which can influence patient outcomes. Finally, patients with diabetes mellitus had a higher mortality, which may have introduced bias. Diabetes itself affects lipid levels (e.g., decreases HDL-C levels and increases TG levels) and could have played a much more important role in death than did the altered lipid levels.

Conclusion

HDL-C and LDL-C concentrations were low in patients upon ICU admission for severe COVID-19, but they were not associated with poor outcomes. The low lipoprotein concentrations observed in bacterial superinfections were associated with mortality. These findings reinforce the hypothesis that lipoproteins play a role in bacterial sepsis. Studies with higher statistical power are needed to better characterize the role of lipoproteins during severe COVID-19. Further experimental studies evaluating the functionality of these particles, in particular HDL, are crucial to understand the physiopathology of this disease.

Comparison of the total cholesterol, triglycerides, HDL-C and LDL-C concentration kinetics over the first six days in patients with and without lopinavir/ritonavir treatment.

(TIF)

Click here for additional data file.

Comparison of the total cholesterol, triglycerides, HDL-C and LDL-C concentration kinetics over the first six days in patients with and without corticosteroid treatment.

(TIF)

Click here for additional data file.

Correlation between TC, TG, HDL-C and LDL-C concentrations and the number of days alive without mechanical ventilation, ICU length of stay, and hospital length of stay.

(TIF)

Click here for additional data file.

14 Aug 2020

PONE-D-20-20848

Lipoprotein concentrations over time in the intensive care unit COVID-19 patients:

Results from the ApoCOVID study

PLOS ONE

Dear Dr. TANAKA,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Sep 27 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Wenbin Tan

Academic Editor

PLOS ONE

Comments from the Editor:

1. The authors need to specify that the lipid profiles were determined at the time of ICU admission in the abstract, figure legends and figures. For example in the conclusion of abstract, "HDL-c and LDL-c concentrations at the ICU admission are..". and many other places in the contents.

2. justification of using LDL <1.4 mM as the cut-off value needs to be elaborated?

3. how many patients died after 28 days? what is the justification of using 28 days at ICU? whether it can add the lipid profiles at time of discharge or death for analyses instead of 28 days?

4. Had the patients stayed in non-ICU ward before being admitted into ICU? if so, did the authors have the lipid profiles of these patients in this cohort on the non-ICU ward admission? Would the LDL or HDL levels be higher than on the time of ICU admission for those survival patients?

5. the authors need to cite references Clinica chimica acta; 2020:10.1016/j.cca.2020.07.015. and FASEB journal 2020:10.1096/fj.202001451.

6. the authors need to explain the possible reason why other reports (refs 28, 32, and 2020:10.1016/j.cca.2020.07.015)  showed the association with disease severity because that they used lipid profiles at patient hospitalization admission which levels were generally higher than at the time of ICU admission for those surviving patients.

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If the need for consent was waived by the ethics committee, please include this information.

3. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

[Note: HTML markup is below. Please do not edit.]

**********

Review Comments to the Author

Reviewer #1: The goal of the study was to characterize the lipoprotein profiles of severe ICU patients hospitalized for COVID-19 pneumonia and to assess their changes during bacterial ventilator-associated pneumonia (VAP) superinfection.

The authors found that HDL-C and LDL-C concentrations are low during severe COVID-19 pneumonia but are not associated with poor outcomes, and that low lipoproteins in the case of bacterial superinfection during ICU hospitalization are associated with mortality.

Comments

1. The presence of diabetes or hyperglycemia is associated with worse prognosis of severe Covid-19 pneumonia. How was a difference in glycemic control during hospitalization between alive patients and dead patients at 28 days?

2. Did you look at changes in inflammatory markers, including leukocytosis and C-reactive protein 6, during hospitalization between alive patients and dead patients at 28 days?

3. Cytokine storm contributes to poor outcome of patients with severe COVID-19. Did you look at changes in cytokine levels (IL-2R, IL-6, TNF-alpha) during hospitalization?

4. Did the use of statin affect prognosis of severe Covid-19 pneumonia?

5. Please add prevalence of use of ACEI or ARB to Table 1.

Reviewer #2:

The manuscript definitely needs English language editing.

The authors use SARS-CoV-2 interchangeably with COVID-19. However, being a clinical study (i.e., not basic science), it is advisable if they stick to using the name of the disease i.e., COVID-19, and not the name of the virus (SARS-CoV-2) which is to be reserved mostly for mechanistic and preclinical studies.

In the abstract they mention “the short-term prognosis outcome was assessed”. That is a very vague statement and requires clarification by saying that survival at 28 days post admission with incidence of VAP was assessed.

They mention that LDL-C levels have been previously correlated with some viral infections (eg, Dengue fever) but do not clarify the direction of the correlation (direct/positive or inverse/negative)?

Table 1 shows that diabetes was far more prevalent among the patients who died at the end of the study so that is a very major confounding factor that the authors fail to discuss. Diabetes itself affects lipid levels, and could have played a much more important role in death than did the altered lipid levels. This should be addressed.

Also, the use of corticosteroids an protease inhibitors differed between those who were alive vs dead at the end of the study (even if not significant; it was a marked difference in rates). Both drugs are associated with marked variation and have different effects on lipoprotein levels so these too are major confounders that should be noted and discussed as the differences in lipoprotein levels between the two groups could simply be a result of the different rates of using these drugs among the groups.

All patients had higher than normal TG levels at admission which also warrants discussion.

They also fail to discuss the direction of associations found in their studies or other similar studies they cite. They do not mention whether the associations were positive or negative, which is crucial to clarify throughout the manuscript, including when citing other studies. Also, refrain from saying “statistical association” and replace with “statistically significant association”.

Finally, since many COVID-19 patients develop a nephrotic-like syndrome, that can be the reason for alteration of the lipid levels and not the infection per se. This should be noted and clearly discussed in the manuscript.

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

2 Sep 2020

Dear Editor,

Thank you for reviewing our manuscript entitled “Lipoprotein concentrations over time in the intensive care unit COVID-19 patients: Results from the ApoCOVID study”. We appreciate your positive comments and relevant questions that will help to improve the manuscript. Please find attached a revised version of our manuscript and a point-by-point response to the editor and reviewer comments.

Comments from the Editor:

1. The authors need to specify that the lipid profiles were determined at the time of ICU admission in the abstract, figure legends and figures. For example in the conclusion of abstract, "HDL-c and LDL-c concentrations at the ICU admission are..". and many other places in the contents.

In accordance with this comment, we have added the term “at ICU admission” in the abstract, manuscript, figure legends and figures.

2. justification of using LDL <1.4 mM as the cut-off value needs to be elaborated?

To evaluate the impact of low LDL-C values on outcomes, the first step was to determine the threshold of a low LDL-C level. As no cutoff is consensual, our population was split into four quartiles according to LDL-C level, and the threshold to determine the lowest quartile was retained as the cutoff value. This value was 1.4 mM.

This point is mentioned in the statistics section: ”The threshold defining the lower quartile was determined to have 25% of the overall population in that quartile”.

3. how many patients died after 28 days? what is the justification of using 28 days at ICU? whether it can add the lipid profiles at time of discharge or death for analyses instead of 28 days?

Thank you for raising this important question. Only one patient died after 28 days. He died at day 67. A 28-day endpoint is often applied in studies involving ICU patients, especially for sepsis because of the high mortality rate of this condition, particularly in the first month. In this context, although this cutoff has become more controversial according to some authors, a 28-day mortality endpoint is relevant (Vincent et al. Crit Care Med 2004 PMID 15118519).

We collected all lipid concentrations from ICU admission to discharge or death. To perform the analyses, we had to choose a mortality endpoint. In most cases, the death of our ICU patients occurred during the first month (median delay = 7 days [7-13]). In this context, and regarding the literature in the field, we selected the 28-day endpoint.

4. Had the patients stayed in non-ICU ward before being admitted into ICU? if so, did the authors have the lipid profiles of these patients in this cohort on the non-ICU ward admission? Would the LDL or HDL levels be higher than on the time of ICU admission for those survival patients?

A total of 29 patients were directly admitted to the ICU, and 19 patients stayed in the non-ICU ward before being admitted to the ICU. The median delay between hospital and ICU admission was very short (1 [0-3] days), and unfortunately lipid profiles of non-ICU patients are not routinely performed.

We added the number of patients who were directly admitted to ICU in the Results section: “Twenty-nine patients were directly admitted to the ICU, while 19 patients had a short stay (1 [0-3] days) in another ward before ICU admission.”

5. the authors need to cite references Clinica chimica acta; 2020:10.1016/j.cca.2020.07.015. and FASEB journal 2020:10.1096/fj.202001451.

In accordance with this suggestion, we have added these interesting references to the Discussion section.

6. the authors need to explain the possible reason why other reports (refs 28, 32, and 2020:10.1016/j.cca.2020.07.015) showed the association with disease severity because that they used lipid profiles at patient hospitalization admission which levels were generally higher than at the time of ICU admission for those surviving patients.

Thank you for this relevant question.

Fan et al. (Fan et al. Metabolism 2020 PMID 32320740) collected lipid profiles of 21 COVID-19 patients. Compared to data collected before falling ill with COVID-19 pneumonia, the authors showed that TC and LDL-C levels were decreased at the time of admission and then returned to the level prior to infection. The HDL-C level also decreased but remained low over time. Interestingly, logistic regression analysis showed increasing odds of a lowered LDL level being associated with disease progression and in-hospital mortality.

In the Wei et al. study (Wei et al. J of Clinical Lipidology 2020 PMID 32430154) involving 597 COVID-19 patients and control patients, the authors showed that COVID-19 patients had lower LDL-C, HDL-C and TG levels at admission compared with control patients. When patients were stratified as mild/severe and critical, the HDL-C levels were lower in the critical patients than in the mild/severe patients.

Hu et al. (Hu et al. Clinica Chimica Acta PMID 32653486) analyzed 114 COVID-19 cases and 80 age-matched healthy controls. TC, HDL-C, LDL-C levels were significantly decreased in COVID-19 patients compared to healthy controls. Additionally, HDL-C levels were lower in the severe group than in the common group. The authors concluded that decreased serum HDL-C was associated with an increased severity of COVID-19.

These different studies demonstrate that hypolipidemia, and in particular a decreased HDL-C concentration, is associated with severe disease. Interestingly, high variability was observed in patient severity in these studies, with very few patients hospitalized in an ICU, while the majority was hospitalized in a non-ICU ward. This allows stratification of the patients according to severity. Since our patients were all severe ICU patients, stratification was not performed. In addition, the results of these studies cannot be directly compared with those of our work. It should also be noted that these studies compared data from COVID-19 patients with healthy volunteers. We did not perform this type of comparison.

We further addressed this comment in the Discussion section: “In our cohort of 48 severe COVID-19 patients hospitalized in our ICU, we found no statistically significant link between HDL-C or LDL-C concentrations and patient outcomes. These results are not in accordance with the results of the study by Fan et al., which showed that a low LDL-C concentration was a potential predictor of poor prognosis in a cohort of 21 heterogeneous ICU and non-ICU COVID-19 patients (30). Interestingly, the studies by Hu et al. and Wei et al. have specifically shown that low HDL-C levels were associated with severe COVID-19 disease (34,39). However, the heterogeneity reported in these two studies, which included mild, severe and critical patients and ICU and non-ICU patients, does not allow for direct comparisons with our specific severe COVID-19 ICU cohort (34,39)”.

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://clicktime.symantec.com/3187EePMb4iL9YFXPcRHTy26H2?u=https%3A%2F%2Fjournals.plos.org%2Fplosone%2Fs%2Ffile%3Fid%3DwjVg%2FPLOSOne_formatting_sample_main_body.pdf and

https://clicktime.symantec.com/3RE1YxDqS7h8kUXmYktK8Ap6H2?u=https%3A%2F%2Fjournals.plos.org%2Fplosone%2Fs%2Ffile%3Fid%3Dba62%2FPLOSOne_formatting_sample_title_authors_affiliations.pdf

The manuscript was modified according to PLOS One’s style requirements.

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If the need for consent was waived by the ethics committee, please include this information.

The need for consent was waived by the ethics committee. We have added this information in the Methods section and in the online submission information section.

3. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://clicktime.symantec.com/393EgirqgVZPnoGeJag8CU46H2?u=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D_xcclfuvtxQ

We have added the ORCID iD online, which is validated in the Editorial Manager.

Review Comments to the Author

Reviewer #1: The goal of the study was to characterize the lipoprotein profiles of severe ICU patients hospitalized for COVID-19 pneumonia and to assess their changes during bacterial ventilator-associated pneumonia (VAP) superinfection.

The authors found that HDL-C and LDL-C concentrations are low during severe COVID-19 pneumonia but are not associated with poor outcomes, and that low lipoproteins in the case of bacterial superinfection during ICU hospitalization are associated with mortality.

Comments

1. The presence of diabetes or hyperglycemia is associated with worse prognosis of severe Covid-19 pneumonia. How was a difference in glycemic control during hospitalization between alive patients and dead patients at 28 days?

As reported in numerous studies, we found that diabetes mellitus was associated with a worse prognosis (mortality at 28 days in diabetic patients = 56% vs. 25% in non-diabetic patients, p = 0.033).

In our ICU, glycemic control during hospitalization was standardized for all patients (diabetic and non-diabetic patients) according to the most recent recommendations (Surviving Sepsis Campaign International Guidelines for Management of Sepsis and Septic Shock, Critical Care Medicine 2016). We used a glycemic control protocol. If two consecutive blood glucose levels were >180 mg/dl, IV or subcutaneous insulin was administered. The target was an upper blood glucose level ≤180 mg/dl. Blood glucose values were monitored every 1 to 2 hours until the glucose values and insulin infusion rates were stable, then every 4 hours thereafter in patients receiving insulin infusion.

With this standardized glycemic control protocol, we found no difference in glycemic control during hospitalization between survivors and nonsurvivors at 28 days.

2. Did you look at changes in inflammatory markers, including leukocytosis and C-reactive protein 6, during hospitalization between alive patients and dead patients at 28 days?

We thank the reviewer for this interesting question.

At admission, no significant difference was noted between survivors and nonsurvivors at 28 days in the levels of leukocytes, lymphocytes, C-reactive protein and procalcitonin. Furthermore, no significant differences were noted during the course of ICU stay, but survivors tended to have higher lymphocyte counts. This observation is in accordance with several studies reporting that lymphopenia leads to poor outcomes in COVID-19 patients (Li tan et al. signal transduction and targeted therapy, 2020, lymphopenia predicts disease severity of COVID-19, PMID 32296069; Zhao et al. International journal of infectious diseases, lymphopenia is associated with severe coronavirus disease, a systematic review and meta-analysis, PMID 32376308).

We added this information (leukocytes, lymphocytes, C-reactive protein and procalcitonin values) to Table 1.

3. Cytokine storm contributes to poor outcome of patients with severe COVID-19. Did you look at changes in cytokine levels (IL-2R, IL-6, TNF-alpha) during hospitalization?

Unfortunately, we did not monitor cytokine levels in our study. We added this point to the limitations of the study.

4. Did the use of statin affect prognosis of severe Covid-19 pneumonia?

We thank the reviewer for this question. A recent retrospective study involving 1,219 patients receiving statins and using propensity score-matching showed that the risk for 28-day all-cause mortality was 5.2% and 9.4%, respectively, in the matched statin and non-statin groups, with an adjusted hazard ratio of 0.58 (Zhang et al. Cell Metab. 2020 PMID 32592657). Other studies did not find this association. In our cohort of severe COVID-19 patients, as indicated in Table 1, statin medication did not affect prognosis (22% in survivors vs. 37% in nonsurvivors, p = 0.310).

5. Please add prevalence of use of ACEI or ARB to Table 1.

We found no difference between survivors and nonsurvivors. We have added this information to Table 1.

Reviewer #2:

1. The manuscript definitely needs English language editing.

We used an English language editing service to improve the quality of the manuscript (AJE certificate number D83B-0D4E-75F4-9249-C616, September 1,2020).

2. The authors use SARS-CoV-2 interchangeably with COVID-19. However, being a clinical study (i.e., not basic science), it is advisable if they stick to using the name of the disease i.e., COVID-19, and not the name of the virus (SARS-CoV-2) which is to be reserved mostly for mechanistic and preclinical studies.

Thank you for this relevant comment. We changed the term SARS-CoV-2 to COVID-19 throughout the manuscript.

3. In the abstract they mention “the short-term prognosis outcome was assessed”. That is a very vague statement and requires clarification by saying that survival at 28 days post admission with incidence of VAP was assessed.

We changed “the short-term prognosis outcome was assessed” to “The assessed outcomes were survival at 28 days and the incidence of VAP”.

4. They mention that LDL-C levels have been previously correlated with some viral infections (eg, Dengue fever) but do not clarify the direction of the correlation (direct/positive or inverse/negative)?

We apologize for this inconsistency. We clarified the direction of the correlations in the manuscript.

- Marin-Palma et al. showed that dengue patients had significantly lower HDL-C, LDL-C and TC levels and higher triglyceride level compared to healthy controls (Marin-Palma et al. Plosone. 2019 PMID 30901375). Interestingly, as described by Biswa et al. in patients with dengue disease, lower LDL-C levels were associated with severe dengue outcomes (Biswa et al. plos neg trop dis 2015 PMID 26334914). We have added these modifications and references to the Introduction section.

- Interestingly, lipid profiles depend on the progression of hepatitis B disease. Cao et al. found that TC, HDL-C and LDL-C are mostly low in cases of hepatitis B-related cirrhosis (Cao et al. Clin lab 2020 PMID 31850701).

- HIV infection and its treatment are frequently associated with dyslipidemia, with elevated TC and LDL-C levels and decreased HDL-C levels (Rose et al. Atherosclerosis. 2008 PMID 18054941). We added this point to the Discussion section.

5. Table 1 shows that diabetes was far more prevalent among the patients who died at the end of the study so that is a very major confounding factor that the authors fail to discuss. Diabetes itself affects lipid levels, and could have played a much more important role in death than did the altered lipid levels. This should be addressed.

Thank you for this important point to discuss. As reported in numerous studies, we also observed a worse prognosis in diabetic patients (mortality at 28 days in diabetic patients = 56% vs. 25% in non-diabetic patients, p = 0.033). The cause of this excess mortality is not well-described for the population of diabetic COVID-19 patients. Diabetic patients frequently have dyslipidemia and, in particular, decreased HDL-C and increased triglycerides.

As proposed by the reviewer, diabetes itself could have played a much more important role in death than did the altered lipid levels. We agree with the reviewer all the more because we found no relationship between lipoprotein levels are mortality in our cohort. We remarked upon this in the limitations paragraph in the Discussion section.

6. Also, the use of corticosteroids an protease inhibitors differed between those who were alive vs dead at the end of the study (even if not significant; it was a marked difference in rates). Both drugs are associated with marked variation and have different effects on lipoprotein levels so these too are major confounders that should be noted and discussed as the differences in lipoprotein levels between the two groups could simply be a result of the different rates of using these drugs among the groups.

Thank you for raising this interesting point. Indeed, patients who received protease inhibitors had a lower mortality rate (17%) than the overall mortality rate of 33%.

Protease inhibitors can induce chronic dyslipidemia; this has primarily been described in HIV patients who have increased TC, TG, and LDL-C levels and a variable decrease in HDL-C levels (Berthold et al. Journal of Internal Medicine 1999 PMID 10620100, Roberts et al. CID 1999 PMID 10476757, Montes et al. JAC. 2005 PMID 15761071). Thanks to the reviewer, we compared lipid levels between patients with and without protease inhibitor treatment and found no differences over time (see new supplemental figure S1). This may be due to the fact that the altered lipid concentrations appear long time after the initiation of therapy (for example, in the study by Roberts et al, lipid profile alterations were observed 33 months after treatment initiation). In our study, the treatment duration was likely too short to see an impact of lopinavir/ritonavir on lipid metabolism.

However, undeniably, protease inhibitors could induce a bias; therefore, we included this point in the limitations.

Even more controversial than protease inhibitors, corticosteroids can also induce chronic changes in lipid levels, such as increased levels of TC, TG and LDL-C. Similarly, no differences were noted between patients with or without corticosteroid treatment (see new supplemental figure S2). This point was also added to the Discussion.

7. All patients had higher than normal TG levels at admission which also warrants discussion.

Thank you for this interesting point. We added this issue to the discussion: “Interestingly, compared to the HDL-C, LDL-C and TC concentrations, the TG concentration at admission was higher than the normal range. Several hypotheses can explain this finding. First, one-third of the patients had diabetes, a condition in which lipid metabolism disorders are frequent (particularly increased TG concentrations). Second, the early use of propofol upon admission may explain the increased TG concentration. Third, renal dysfunction and especially nephrotic-like syndrome is observed in COVID-19 disease, which may have induced an increased level of TG. Finally, direct effects of the virus cannot be excluded.”

8. They also fail to discuss the direction of associations found in their studies or other similar studies they cite. They do not mention whether the associations were positive or negative, which is crucial to clarify throughout the manuscript, including when citing other studies.

Thank you for this comment. We clarified the purpose in the Discussion section of the manuscript.

Also, refrain from saying “statistical association” and replace with “statistically significant association”.

We changed the term “statistical association” to “statistically significant association”.

9. Finally, since many COVID-19 patients develop a nephrotic-like syndrome, that can be the reason for alteration of the lipid levels and not the infection per se. This should be noted and clearly discussed in the manuscript.

Nephrotic syndrome induces lipid disorders with increased concentrations of TC and TG. Although we did not note nephrotic-like syndrome in our cohort of COVID-19 patients, this syndrome has been described in COVID-19 pneumonia and could explain part of the altered lipid levels, especially the increased TG concentrations, in our patient cohort. We added this point to the Discussion section of the manuscript: “Third, renal dysfunction and especially nephrotic-like syndrome is observed in COVID-19 disease, which may have induced an increased level of TG.”

10 Sep 2020

Lipoprotein concentrations over time in the intensive care unit COVID-19 patients:

Results from the ApoCOVID study

PONE-D-20-20848R1

Dear Dr. TANAKA,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Wenbin Tan

Academic Editor

PLOS ONE

Additional Editor Comments:

Please consider to add the following reference into the content during your proofreading:

https://doi.org/10.1038/s41392-020-00292-7

Reviewers' comments:

Reviewer #1: This manuscript has been considerably improved. I feel no significant concerns about this revised manucsript.

**********

17 Sep 2020

PONE-D-20-20848R1

Lipoprotein concentrations over time in the intensive care unit COVID-19 patients: Results from the ApoCOVID study

Dear Dr. TANAKA:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Wenbin Tan

Academic Editor

PLOS ONE